Chemical Process Control - Optimizing Manufacturing with Industrial IoT

Chemical Process Control -
Optimizing Manufacturing with Industrial IoT

In chemical manufacturing, we must look at the process both in detail and holistically, in order to identify inefficiencies. By analyzing production disturbances through the use of process-based machine learning, we have the opportunity to reach new levels of chemical process control.

 

What is a “production disturbance”?

The significance of the term “production disturbance” (PD) varies since every manufacturing facility has a unique operational structure, raw materials supply, machine configuration, and production environment.

For the sake of this discussion, a production disturbance is any unintentional event in the chemical production process that leads to process inefficiencies, unplanned stoppages, rework, or scrap, for example:

  • Valve leakages
  • Head pressure drops in pumps
  • Lubricant issues, e.g. frothing
  • Inconsistent bearing temperatures

It’s important that PDs are defined specifically on the basis of individual machines, processes, and manufacturing environments.

 

Reducing production disturbances – The paradox of preventive maintenance

For many years, one of the most championed best practices in asset maintenance was preventive maintenance. The idea of preventive maintenance is to preempt and avoid malfunction or production disturbances by performing scheduled asset maintenance regularly.

It has since been discovered that preventive maintenance can be inefficient in a number of ways, leading to:

  • Redundant planned downtime (up to 40% of preventive maintenance costs are spent on assets with negligible failure impact.)
  • Secondary damage to equipment – caused by invasive inspections
  • Premature/untimely equipment replacement
  • Materials waste – lubricants, oils, coolants etc.
  • An inflated inventory of spare parts

These cost factors  are part of what has led manufacturers to Industry 4.0, and more specifically, from preventive to predictive maintenance.

 

Is predictive maintenance the answer?

Predictive maintenance is without a doubt a game changer. It’s a much welcomed improvement over previous maintenance strategies, and is fast gaining recognition as the new best practice for leading maintenance operations.

Predictive maintenance is focused at preventing mechanical failure in specific assets. However, production disturbances are not necessarily asset failures.

In fact, disturbances are more often the result of process failures such as irregular cooling in a tank (the disturbing factor) that’s yielding high pressure in a pump (the disturbance), for example.

This calls for a broad-scope examination of the chemical production process and its production parameters. A narrow focus on individual asset behaviors leaves the production process context out of the equation.

Predictive maintenance is not a -one-size-fits-all solution. Not accounting for the context of the process will lead to too many false-positives and a flood of alerts that don’t provide insight and harm the credibility of the system.

 

And what about maximizing OEE?

Another well-known methodology for production optimization is by closely monitoring Overall Equipment Efficiency (OEE).

OEE is calculated using the formula:

Availability X Performance X Quality = OEE

The method was developed by Seiichi Nakajima in the 1960s as a means to maximize availability, performance, and quality – and in doing so, minimizing production disturbances.

OEE is a bottom-up approach that gives operators and technicians “ownership” of their assigned processes with the goal of minimizing the Six Big Losses:

six-big-losses-oee

The downside of the approach is that for complex facilities such as the ones in chemical processing, this formula can be too broad.

For example, the formula represents each of the components equally. This can be countered using weighted variables, but that can lead to overproduction and manipulations to the formula that don’t necessarily improve production throughput.

 

The solution: predict and prevent process disturbances

Focusing solely on deploying predictive maintenance or increasing the OEE percentage can lead to sub-optimization. The impact of individual sub-processes on the performance of the entire system needs to be evaluated at depth.

This leads us to the core concept behind “chemical process control”:

Using automated root cause analysis, predictive analytics, and what-if simulation to predict and prevent process disturbances that impact production throughput.

 

Shifting the focus to the process

While an individual pump, motor, or filter might malfunction, it is often an instability in the chemical production process – a process disturbance – which has led to the failure. In other words, the process disturbance is the root cause for the asset failure.

To tackle this complex problem, we need to account for relationships between production parameters across all stages of the manufacturing process.

The Seebo platform. A predictive alert is displayed within the context of the manufacturing process.
The Seebo platform. A predictive alert is displayed within the context of the manufacturing process.

Using process-based machine learning, we can uncover relationships that would otherwise be impossible to detect:

  1. The production plant is precisely modeled to include all the production lines, physical assets, manufacturing stages, and the product flows through the process.
  2. Production context is added through feature engineering – critical for closing the gap between the real-world manufacturing environment and data representations. Once the data is contextualized, machine learning algorithms such as Random Forest, XGBoost, Hidden Markov, and Directed Acyclic Graph can be used to form predictions. The data has been analysed by the ML algorithms within the context of the entire manufacturing process. This results in accurate predictions regarding quality levels, maintenance, and the supply chain.
  3. Personnel and management receive actionable predictions with supporting root causes in time to improve the performance of the production process.

 

Chemical Process Control – Optimization through AI

By using machine learning algorithms that take into account the process, we get focused and contextual predictive alerts. This is a huge opportunity for chemical process optimization since data is relatively well collected and stored in this sector already.

Leveraging this data with process-based AI means being able to pinpoint the root cause of process disturbances with extreme accuracy, and predict failures before they have the chance to affect production.

Design an Industry 4.0 solution in just two weeks

Get Demo


Preparing for Digital Transformation in Manufacturing - The Complete Guide -

Preparing for Digital Transformation in Manufacturing
- The Complete Guide -

At the risk of sounding a little melodramatic, we’ve passed the threshold and have entered a new age of manufacturing; a technological revolution that will change industry forever. The term “digital transformation” describes this next step, where the utilization of digital data, connectivity, and processing encompasses every aspect of manufacturing activity.

From rapid prototyping and R&D to production and performance analytics, digital transformation in manufacturing is poised to impact all aspects of business from the organizational structure of companies to how they generate revenue.

In this article, we’ll be highlighting some of the core principles behind the digitalization in manufacturing as well as some key trends taking manufacturing operations into the future.

 

What You Get – The Advantages of Digital Transformation in Manufacturing

The benefits of digitization in manufacturing are many, but can be summarized under these 5 categories:

Productivity – Design and development processes are faster and better informed using tools such as 3D printing and augmented reality, and by leveraging behavioral data from users in real time. Production is streamlined with minimal downtime due to connected machines sending vital maintenance data that can be leveraged to prevent malfunctions and optimize output.

Quality – High-definition sensors monitor production parameters of the product along the entire production line process. Sometimes referred to as Quality 4.0, machine learning algorithms are applied to production data to automatically decipher root causes of defects as well as to predict waste-related issues before they occur.

Cost – Capturing and analyzing data across all stages of the manufacturing process, including production line and machine data, logistics and transportation, makes it possible to identify new cost reduction opportunities. Inventory can be better managed to meet demands in a more accurate manner while machines offer a high level of flexibility that allows for quick changes between products.

With IIoT, inventory can be better managed to meet demands in a more accurate manner while machines offer a flexibility that allows for quick changes between products.

Digital transformation in manufacturing

Customization – Customization has become a key selling point for customers. Digitized manufacturing lines can offer customers attractive customization options while still operating on a mass scale and at a high level of efficiency so that prices remain competitive.

Safety – Work in dangerous environments can be handled by robots. Staff can be alerted about potential hazards well in advance thanks to dedicated sensors placed throughout the plant/factory.

 

Meet the 10 MVPs of Digital Transformation in Manufacturing

If you thought you were getting a list of CEOs here, sorry to disappoint you, but you’ll probably find this much more useful.

Here’s a list of the 10 “Most Valuable Players” who have already begun shaping the digital transformation in manufacturing…in alphabetical order for your reading pleasure!

1. Additive Manufacturing (3D printing)

Commonly referred to as “3D printing”, additive manufacturing incorporates a wide variety of processes and materials that share one common characteristic – the direct transformation of 3D data into the physical realm. This form of manufacturing allows for a freedom of design that was never before possible, and as this technology advances, we’re witnessing the emergence of applications in sectors including, but not limited to, aerospace, automotive, medical and consumer/lifestyle.

Prototyping for additive manufacturing

2. Asset Performance Management (APM)

With the emergence of Industry 4.0, the definition of Asset Performance Management has changed to include a broad set of functions. APM pools together a number of tools in order to improve the availability and overall reliability of physical assets within the manufacturing ecosystem. These tools collect, organize, visualize, and analyze data from the assets, and leverage it by performing predictive forecasting, condition monitoring and reliability maintenance.

3. Cloud Applications & Platforms

The older approach of client-server data management systems is rapidly being replaced by industrial cloud applications. This new method of developing and deploying software has many advantages over the older heavier and more complex server approach, and allows for easy updates and low-cost maintenance.

4. Connected Products and Services

Digitization of the supply chain in the form of connected products positions manufacturers as innovators within their markets. The precise needs of customers can be constantly met through enhancements made to a product’s features. This continuous connectivity strengthens the manufacturer’s stance as a service.

5. The Edge

As computing power improves, more tasks can be handled by devices “on-site”. This reduces latency by lightening the load placed on the IoT network and the cloud, mitigates data security risks, and reduces connectivity costs. This processing by devices in the field is known as “edge computing” and opens up many possibilities within the IoT realm such as obstacle avoidance, language processing, object detection, face recognition, and other machine learning applications.

6. The Fog

The fog, or fog computing, inhabits the space between IoT endpoints and the cloud. In other words, it’s a network that connects the points of data input and creation to where that data is stored. The fog is a midway processing zone that takes care of data from the edge, managing tasks that don’t require the cloud, but can’t be taken care of on-device.

7. Industrial IoT (IIoT) and Industry 4.0

Beginning with the term “Industrie 4.0” which was coined by the German government in 2011, the digital disruption in manufacturing is now well underway. Industrial IoT is advancing at a rapid pace as manufacturers understand the immense potential of this technological approach to completely change how their operations work. Leading use cases of Industry 4.0 are condition monitoring, predictive maintenance, digital twin, data-driven R&D, and fleet management.

8. Machine Learning & Advanced Analytics

The most significant premise of industrial IoT is the ability to receive and analyze constant updates from sensors and other data collection points in real time – and to be able to respond with immediate action. This makes machine learning a very powerful tool in leveraging IoT to benefit industrial production.

Through machine learning and predictive maintenance, the behavior and performance of machines in a manufacturing setup is learned and “understood” while algorithms adapt based upon new information. This allows for unusual behaviors to be identified and for errors and malfunctions to be predicted with high accuracy.

9. Open Process Automation

Many of the automated manufacturing systems in operation today are controlled by what is commonly known as a Distributed Control System (DCS) alongside programmable logic controllers (PLCs). This type of system isn’t very well suited to the current technological climate as the architecture is usually very proprietary in nature, difficult to change, and extremely customized to a specific production line. Open Process Automation aims to provide a new generation of automation infrastructure that can be easily implemented and adapted for use in industrial and consumer IoT scenarios.

10. Robotics

The utilization of robotics in industrial manufacturing is common today with close to 2 million industrial robots working around the world. The motivation behind robotics is clear. The efficiency they offer is unparalleled and they can do monotonous, unpleasant, and dangerous work instead of humans.

The Internet of Robotic Things (IoRT) takes robotic technology to the next level and will be a major part of the future of manufacturing. Production robots will be connected and fed with real-time data that will be used to make decisions with regards to synchronicity and performance on the factory floor. The IoRT will allow manufacturers to better meet the needs of their customers, and accurately respond to changes in the supply chain.

The Internet of Robotic Things (IoRT) takes robotic technology to the next level and will be a major part of the future of manufacturing.

Digital Transformation in Manufacturing

Digital Transformation in Manufacturing: The Complete Smart Factory

The list above describes individual technologies, but imagining a complete “smart factory” which applies Industry 4.0 technology, along with wearables, AR and VR, is where it starts to get really interesting.

In such a scenario, all aspects of manufacturing activity – human, machine, and human-machine interaction – are synchronised and coordinated to achieve optimal output, and ensure sustainability for the operation, and for the people that make it work.

Digital Transformation Strategy

A well-defined digital transformation strategy is critical for the overall success of IoT implementation in a manufacturing setting. The strategy should cover every aspect of business activity – from development and production, to advanced quality control, delivery, and analysis.

The state of the company’s legacy systems should be taken into account to identify potential challenges. As much data as possible should be collected from the machines in their current and past states before implementation of the new system begins.

A well-defined digital transformation strategy is critical for the overall success of IoT implementation in a manufacturing setting.

Digital transformation in manufacturing

IoT offers manufacturers so many potential directions for development that the myriad of options can be confusing. This is why a clear strategy is imperative to ensure focus. Part of this focus is of course the needs of the customer which should be a central goal of the digital transformation process.

Let’s take a look at a breakdown of a step-by-step digital transformation strategy:

  1. Create an Industry 4.0 road map. Take into account the current status of the company with regards to digitization, and then set targets for a 5-year period. Goals with the most significant ROI should take top priority, while measures should be taken to get leadership on board.
  2. Decide upon projects that establish POC. This somewhat experimental phase should use a variety of pilot projects to establish the performance of cross-functional teams, and gauge how agile the process is.
  3. Define target functionality. Based on knowledge gained from Step 2., decide which Industry 4.0 capabilities will drive the most value for the company. At this stage, you should be better informed about the abilities of your teams to implement the new technology, and whether additional recruiting is necessary.
  4. Learn to leverage data analytics. Progressing to Industry 4.0 means nothing without being able to analyze the collected data. This analysis should be immediately fed into the decision making process.
    Adopt digital transformation as a company. Implementing Industry 4.0 is more than a temporary adjustment phase. To reap the benefits, adoption of this new approach should be company-wide, led from the top with C-suite and financial stakeholders setting the tone.
  5. Develop as an integral part of your ecosystem. As you use IoT to create better solutions for your customers, keep a broad vision of your position within your business ecosystem. Share knowledge with partners and suppliers, and explore potential avenues for further collaboration to further the quality and scope of your products and services.

SEE ALSO: Top Industry 4.0 Use Cases

Digital Transformation Challenges

The roadmap for digital transformation does include a number of challenges. The good news is that there are already many tools and services in place to assist manufacturers in making the digital transformation process structured, predictable, and successful.

Here are some of the challenges to look out for when implementing digitization in manufacturing:

Budget limitations

Leading a manufacturing facility through the digital transformation journey requires a substantial investment. The rewards are numerous, both short and long-term, but it’s important to keep in mind that every business is different, especially when it comes to revenue and expense structures. This process requires planning and customization as no two digital transformation programs should ever look the same. The needs of every plant or factory are different, and so are the available resources.

The good news is that IoT is flexible and not a one-size-fits-all type of tool. Manufacturers with a more limited budget should think big initially since having a long-term vision is important for reaching a truly valuable goal down the line. Once this vision has been explored, a solution with a solid ROI should be sought out as a proof-of-concept. This means that as a first phase in the process, data collected by the network – and the analytics and resultant actions based on that data – should be the most important and influential information for that specific operation. Once those central parameters are being leveraged, decisions about how to further the capabilities of the network can be taken.

Lack of relevant knowledge

Introducing technology alone is not enough without the relevant knowledge to make it work. Investing in employees’ knowledge is an important part of integrating IoT into manufacturing. If the level of expertise within the company is insufficient, management will have to consider partnering with external consultants or hiring new employees. Even in such cases, the introduction of IoT shouldn’t be the sole responsibility of one employee or department, and instead should be a shared goal.

Rigid company structure

The introduction of IIoT to a manufacturing facility is more like a paradigm shift than a slight improvement. The organization itself will need to change in order for this new technology to be properly implemented. While this can be daunting, it can lead to a lot of positive outcomes as organizational structure is reset and re-tested, creating the opportunity for better employee placement and other improvements.

One approach is to form a multi-disciplinary team that includes engineers, product designers, data analysts, and service professionals to act as the primary agents for the digital transformation. This team will incubate new technologies, implement POCs, and then roll out successful iterations to the company after they’ve been approved.

Unsuitable development processes

Manufacturers need to understand that their technology stack and development processes will need to undergo numerous changes to suit the more agile nature of Industry 4.0. Release cycles based upon quarters, or other lengthy and rigid iteration schedules, will need to be replaced. The goal is to make use of the data in completely new ways, which naturally demands changes to business rules, the way that content is presented, and how data is leveraged

This represents a real sea-change in manufacturing. As product releases become continuous, the IoT development process will need to support this behavior, utilizing data from user feedback and using analytics in order to achieve a high level of digital performance.

To do this, updates will be needed to make the data read and write-accessible through secure and robust APIs. This is difficult to do with outdated tech, which unfortunately means more than 5 years old when it comes to core business systems.

Employee reluctance

Not everybody is open to change. In fact, most people don’t welcome changes to their work environment. The current digital disruption in manufacturing is experienced as a threat to many employees.

While no one can be certain of what the future holds, change is not something to fear. Commitment to the digital transformation process should start with executive management and be passed onto individual employees. Clear communication and transparency is key, and getting everyone excited about the potential of this new technology can’t hurt either.

Security

Cyber security should be taken into account at the start of any digital transformation project. Points of vulnerability should be identified and a number of defense layers and fail-safe mechanisms need to be in place to ensure that the system is completely secure.

 

Ok, I’m in. Now what?

A good approach to launching a digital transformation in manufacturing is to identify improvement opportunities in performance that will result directly or indirectly in significant benefits to the customer. This places the focus on areas such as the supply chain, operations, customer service, engineering and support as well as the business model itself.

Additional tips for beginning a digital transformation process are to:

  • Define your business objectives, and set out a clear strategy to reach them
  • Start with short-term projects that can deliver a measurable ROI
  • Begin moving applications and data to the cloud, and extracting machine data to a local gateway
  • Test a single production line or asset, and then scale up
  • Partner up with experts in digital technology so that you’re aware of the latest solutions and how to implement them

The digital transformation in the manufacturing sector is very evident in some companies and completely dormant in others. How do you view your operation at the moment? Where would you like to be in the future? These two very simple questions are important to answer as an initial step in laying out your digital transformation plan.

To learn more, pick up our free handbook on how to build a successful digital transformation strategy.

And when you’re ready, get in touch to start your free trial with the Seebo Industrial IoT platform.

Design an Industry 4.0 solution in just two weeks

Get Demo


Industrial IoT

IoT Design: Making Connected Products for Smart Manufacturing

What is IoT design?

IoT products and systems combine physical and digital components that collect data from physical devices and deliver actionable, operational insights. These components include: physical devices, sensors, data extraction and secured communication, gateways, cloud servers, analytics, and dashboards.

Not only do all these components need to be designed, but their inter-dependencies must also be fully accounted for.

 

IoT Design in Manufacturing

For product and engineering teams designing IoT systems, the core challenge lies in taking IoT use cases and turning them into a connected system – with full integration, the right IoT communication protocols, security, and a user-friendly look and feel. For industrial manufacturing, IoT product design is also known as Industry 4.0 design.

 

Industry 4.0 Design Principles

There are 4 universal design principles shaping IoT design today:

four universal design principles shaping IoT design for manufacturing

1. Interoperability

At the most fundamental level, a connected system requires sensors, machines, equipment, and sites, to communicate and exchange data. Interoperability is the underlying principle throughout all Industry 4.0 design processes.

2. Information transparency

The rapid growth of connected devices means continuous bridging between the physical and digital worlds. In this context, information transparency means that physical processes should be recorded and stored virtually, creating a Digital Twin.

3. Technical assistance

A driving benefit of IoT, technical assistance refers to the ability of connected systems to provide and display data that helps people to make better operational decisions and solve issues faster. In addition, IoT-enabled things should assist people in laborious tasks to improve productivity and safety.

4. Decentralized decisions

The final principle of Industry 4.0 design is for the connected system to go beyond assisting and exchanging data, to be able to make decisions and execute requirements according to its defined logic.

 

Designing with a Purpose

In order for the Industrial IoT system to effectively fill its purpose, it must be designed with the relevant solution in mind.

Industry 4.0 Solutions

Digital twins serve a variety of solutions across the product life cycle, and must be designed to match the specific solution in question.

To give an example, a digital twin might be designed primarily for accurate, visual tracing of the root causes of machine failure, and the ability to drill down to behaviors of individual assets.

In order for an Industry 4.0 system to correctly identify the source of disruptions, it must be designed to collect data not only from OT and IT systems, but from the entire plant environment – including recipes, batch history, and other operational processes.

Without including the relevant data sets in the logical design phase, the IoT solution will fail to provide the context necessary for accurate root cause analysis.

IoT solutions must also be designed for specific end users. A predictive quality system may have different end-users than a system created primarily to monitor asset health or one which aims at predicting asset failure before it occurs. Different end users and different end goals require dramatically different approaches on the design level, from dashboard user interfaces to the data sets chosen for relevant algorithms.

Each connected systems must be designed with the specific architectures that enables it to fulfill its purpose and generate business value.

 

Transitioning to IoT Product Design

New technology layers involved in IoT design require new skill sets as well.

Almost one third of companies today lack the full resources needed to design and deliver IoT products, such as data integration know-how, web and mobile app development, data analytics, and security.

Industrial IoT Design complexitiesFaced with the skills gap, many companies source additional partners. These new, multidisciplinary teams must efficiently collaborate throughout the design process.

 

IoT Design Complexity

IoT products are a lot more complex, and with the added complexity comes the risk of making errors in design.

Throughout the design process, the costs of mistakes escalate. The result is that the later mistakes are discovered during development, the more they cost for companies.

Industrial IoT Design - the later mistakes are discovered during development, the more they cost for companies

To validate the design and ensure there are no gaps in the data flows or use cases, teams can leverage an IoT simulator. With IoT simulation, a digital prototype of the designed system allows companies to visualize system behavior and mistake-proof the logic ahead of development.

 

IoT Design Tools

Best-in-class Industrial IoT platforms provide the needed tools to support rapid and accurate logical design of IoT:

Visual Modeling

A key tool for accelerating the design process, visual IoT modeling provides a canvas to define the mechanics, electronics, data connectivity, analytics, and dashboards of the entire system.

Visual IoT modeling enables teams to seamlessly extend their CAD models – which cover mechanics and electronics – to the full IoT model, with all its added technology layers and use cases.

The Seebo Platform - Industrial IoT modeling

IoT Simulation

To quickly and easily validate the IoT system design, teams should look for a platform with IoT simulation capabilities. Here, a digital prototype is used to visualize how connected devices, edge and cloud servers, web and mobile apps interact with each other when an event is triggered, and iteratively refine the model based on the simulation runs.

The Seebo Platform - Industrial IoT modeling

Another type of simulation – IoT analytics simulation – provides teams with a means to visualize the in-market insights their IoT system will deliver, including the dashboards and alerts, taken from their model. This ensures that all needed data sources have been accounted for in the design prior to development.

 

The Physical Design of IoT

As we enter a new age in which unprecedented data is exchanged between systems, the increasing demand for connectivity solutions brings with it new requirements for product functions and capabilities.

Once companies begin the physical design of their IoT system, whether by retrofitting existing products or developing new ones, there are a host of factors they need to account for at the outset. Here, physical design means more than creating connected products, but forming an overall intelligent system.

Industrial machines, for instance, must have sensors capable of generating significantly more data than ever before, and to send the information securely for analysis and action.

Exact placement of sensors on the device, and the ability of the sensors to function in extreme conditions must be factored.

Deciding which IoT communication protocol to use for data integration is another decision companies face at the beginning of IoT design.

To simplify the complexities that arise in the physical design process, companies turn to development platforms which focus on both IoT design and IoT delivery. These platforms ensure that the physical design attributes necessary for systems to properly function and communicate, are virtually represented in the system model.

 

What is the best IoT design methodology?

There are several approaches to IoT design aimed at overcoming the challenges that IoT presents. Here we propose a combination of tactics which together accelerate and help to ensure success for the design process.

Taking a lean, agile design approach

Several product development methodologies have been adapted for efficient IoT design and delivery. The leading approach today is Stage-Gate, in which teams carry out tasks based on a detailed plan, review their outcomes, reach a gate focused on analysis, and only then move onto the next stage.

To apply this methodology of innovation to IoT design, teams can leverage visual modeling and virtual prototyping to simulate the system design, present it internally, and iterate based on feedback.

Stage-Gate product development methodology for Industrial IoT

Incorporating “Design Thinking”

The core principle of design thinking is to factor people, technology, and business into all product design decisions. This approach is customer-centric and views the customer’s needs as a crucial consideration throughout the product development process. For IoT design, this is especially important as it reinforces the notion that an IoT system is not a goal in it of itself, but rather a business solution for specific user needs.

No matter which methodology teams choose for carrying out their IoT design, if it allows for continuous review and iteration based on business requirements and customer needs, it’s on the right track.

For more information about IoT design, visit our library of free IoT resources.

Designing a Factory 4.0 product/system? Contact us to find out how the Seebo platform can make your IoT design process faster and easier.

Design an Industry 4.0 solution in just two weeks

Get Demo


Supervised vs. Unsupervised Machine Learning

Supervised vs. Unsupervised Machine Learning

Machine learning is a category of artificial intelligence that includes a number of algorithmic approaches. In manufacturing, the 2 most common approaches are supervised and unsupervised learning.

There are endless application opportunities in industry for machine learning. Here are some examples:

  • Predicting process disturbance in chemical production
  • Predicting quality failures in a production line
  • Predicting production waste in food or PCB manufacturing
  • Predicting asset failure in a power plant
  • Determining the parameters of a “golden batch” for optimal production throughput

The topic of supervised vs. unsupervised machine learning is actually a somewhat contested one in the Industry 4.0 domain. The reality is that there is no one-size-fits-all machine learning technique that can meet the requirements of every type of manufacturing application.

So, what’s the plan of action?

Process engineers should aim to improve their understanding of machine learning techniques to support decisions regarding the use of AI in production optimization for their specific manufacturing challenges.

While this topic cannot possibly be covered completely in a single blog post, this article aims to touch upon the basics of common ML approaches as they pertain to manufacturing, and to describe when each might be appropriate.

Supervised Machine Learning

“Supervised Learning” describes a relatively didactic process by which predictive machine learning models are developed. For this type of machine learning, historical input and output data are made available to the model.

The method used to create an algorithm from a training dataset resembles a teacher guiding a student to reach a specific goal. The “student” algorithm progresses by making iterative predictions based on the training data, and is corrected by the “teacher”.

Supervised Learning problems can be split into 2 main types:

a. Classification – used when the output is categorical such as “normal” or “warning”.

An example of a classification algorithm is one that receives sensor information as input, e.g. pressure, flow rate, and vibration velocity, acceleration, and displacement, and determines the asset health of a machine.

b. Regression – used when the output is a continuous value such as temperature, voltage, or rpm.

An example of a regression algorithm could be one that receives a component’s code number and performance history as input, and predicts the component’s next malfunction. (An algorithm like this could be used to inform maintenance scheduling.)

Classification vs. regression in machine learning

Unsupervised Machine Learning

In unsupervised learning only input data is required. The goal is for the algorithm to do the work and discover the innate structure of the dataset – to model the distribution of the data and automatically provide insight into correlations.

Like supervised machine learning, unsupervised machine learning problems can be split into 2 main types:

a. Clustering – used to discover groupings found in the input data.

In manufacturing, clustering is used to detect behavior anomalies in the production process and equipment. Using measurements from sensors on a production line, clustering can detect and analyze anomalies/outliers, in turn identifying the root causes of process malfunctions or equipment failure.

Machine learning algorithms - Clustering

b. Association – used to discover rules that can describe relations in the distribution of the input data.

An example of association can be any instance of pattern/behavior detection such as the rise in a pump’s pressure as a result of a temperature increase in a cooling vessel earlier in the process.

Supervised Vs. Unsupervised Machine Learning

Semi-Supervised Machine Learning

In semi-supervised machine learning, labelled and unlabelled data are used together to train the algorithm.

Labelled data significantly improves the learning process of an algorithm. The problem is that large labelled datasets are labor-intensive to create.

This is why semi-supervised machine learning can be very advantageous. Data scientists have found that even when a small group of labelled data is used for training in conjunction with a large unlabelled group, learning accuracy is greatly improved.

 

Reinforcement Learning

Unlike cases where the input is formally fed to an algorithm, in reinforcement learning the algorithm receives input based upon experience.

For example, a robot (agent) can be given the task of learning how to connect two components together (reward). The robot can start off without any data about the task, but through experimentation (actions), will start to collect data about its movement, surroundings, and how the two components interact (observations).

When an action is taken that leads to the two components connecting, or coming close, the data related to that action is labelled accordingly and analyzed. As the robot continues to take more actions and record more data, it improves its knowledge about its task.

Human in the Loop

Unfortunately, the above scenario can be very challenging when it comes to real-world problems like the ones we see in manufacturing.

An algorithm can only perform at the level of our input definitions – how we define the reward, the methods of analysis, and other feature engineering attributes. In manufacturing, with an abundance of parameters affecting one another, it’s extremely difficult to account for everything when building this type of model.

Sometimes, it’s easy to see what an algorithm is doing wrong when you’re observing it objectively. This is the idea behind Human in the Loop (HITL).

With HITL, machine learning applications leverage human knowledge to rule out the obvious “bad ideas”. Instead of investing endless time in attempting to perfect a model, human-sourced experience can be consolidated with the algorithm’s process. This leads to improved results and a more efficient learning process.

 

Choosing the Right Machine Learning Algorithm

In manufacturing, a large number of factors affect which machine learning approach is best for any given task. And, since every machine learning problem is different, deciding on which technique to use is a complex process.

In general, a good strategy for honing in on the right machine learning approach is to:

Evaluate the data. Is it labeled/unlabelled? Is there available expert knowledge to support additional labelling? This will help to determine whether a supervised, unsupervised, semi-supervised or reinforced learning approach should be used.

Define the goal. Is the problem a recurring, defined one? Or, will the algorithm be expected to predict new problems?

Review available algorithms that may suit the problem with regards to dimensionality (number of features, attributes or characteristics). Candidate algorithms should be suited to the overall volume of data and its structure.

Study successful applications of the algorithm type on similar problems.

 

Interpretation is Key

How we interpret the algorithm’s output is crucial to how that algorithm helps us solve real-world manufacturing problems.

It’s important to keep in mind that the output is the result of how the algorithm was defined, how the data was collected and aggregated, and how the output is presented.

The interpretation stage holds a number of risks such as overfitting, which can distort our understanding of the results.

 

Machine Learning Approaches Used in Manufacturing

As in many cases of applied mathematical theory, the answer of which machine learning algorithm to use in manufacturing is the unsatisfactory “it depends”.

Every industry, facility, production line, and problem has its own characteristics. Accounting for as many of these factors as possible will improve the chances of building a system that can provide the desired results.

The decision is also affected by business factors, industry regulations, and the availability of expertise. Keeping sight of all of these parameters, and being able to come up with a machine learning solution that can meet the respective demands, will generate the most value.

 

Getting Started with Machine Learning in Manufacturing

For many manufacturers, the diversity of machine learning – the variety of theories, algorithms, methods and platforms – actually presents a barrier in the path to adoption.

It’s important to note that taking advantage of the benefits of machine learning doesn’t necessarily require a huge investment or major changes to the production floor.

The fact is that in many plants and factories, such as in the chemical processing industry, data is already being captured and stored in a well-structured way. Simply by gaining a better understanding of the type of problems machine learning can solve, manufacturers can begin to explore how their data can drive significant improvements.

 

Cut your production losses with machine learning built for manufacturing:

Get a 1-on-1 demo of the Seebo platform and see how accurate and timely alerts can significantly improve maintenance, product quality, and profitability in manufacturing.

 

Leverage your data to cut production losses.

Get Demo


Digital Twin in Manufacturing

How Digital Twins are Completely Transforming Manufacturing

With the rapid pace of technological growth, it’s not always easy to imagine where digital transformation is taking the manufacturing sector, but one good way of doing this is to take a closer look at the “Digital Twin” concept within the industrial Internet of Things (IoT).

As IoT connectivity provides the manufacturing sector with an increasing number of ways to access sensor-driven data locked in industrial machines and equipment, the need for data analysis, management, and control methods has also become more crucial.

The amount of data collected from monitoring a smart factory is enormous, but if that data isn’t aggregated and organized in a way that can support the decision-making process, then it’s of no use.

One method that’s proving to be invaluable to engineering and customer service teams that are looking to leverage collected data is that of the “Digital Twin”.

 

What is a Digital Twin?

Digital Twin is a virtual representation that matches the physical attributes of a “real world” factory, production line, product or component in real time, through the use of sensors, cameras, and other data collection techniques.

In other words, Digital Twin is a live model that is used to drive business outcomes, and can be implemented by manufacturing companies for multiple purposes:

  • Digital Twin of an entire facility
  • Digital Twin of a production line process within a facility
  • Digital Twin of a specific asset within a production line

 

The Emergence of Digital Twin in Industry

As with previous major turning points in the history of humankind, such as the birth of agriculture and the industrial revolution, Digital Twin software represents a sea change in manufacturing.

the four stages of the Digital Twin evolution

The diagram above shows the four stages of the Digital Twin evolution:

First stage – The entire manufacturing process exists only as a physical version.

Second stage – A digital version is added which augments the physical version with additional information.

Third stage – An interaction begins between the physical and digital versions.

Fourth stage – There is further interaction and convergence between the physical and digital versions.

 

The Digital Twin Concept

Since Digital Twin is essentially a tool, it can be applied in a large number of business use cases.

On a high level, industrial IoT Digital Twins allow us to:

  • Evaluate production decisions based upon analytics
  • Visualize products performing in their environments or being used by actual people in real-time
  • Commission machines from remote service centers, thereby reducing service costs
  • Connect separate systems/processes for improved tracking and monitoring
  • Troubleshoot equipment in remote locations to reduce incident resolution times
  • Gain control over complex processes and systems-of-systems

 

Impact Zones of Digital Twin on Manufacturing

What makes Digital Twin (and IoT in general) so valuable is that it can provide value to businesses throughout the entire manufacturing process, and beyond.

This impact can be categorized into 3 main “zones” for both factories/plants and OEMs:

The Impact on Factories/Plants

Digital Twin for factories

Impact Zone 1 – Maintenance

Digital twins are much more than graphical models. Machine learning algorithms applied to production data detect correlations and form predictions about the remaining useful life (RUL) of assets.

In this way, digital twin solutions take maintenance from a reactive to a predictive approach. Repairs are done “just-in-time”, not too late as to disrupt production, but not too early as to unnecessarily switch out parts or inflate planned stops.

Digital twins allow for technical teams to be extremely well informed, arriving on site with the right spare parts, tools, and instructions for the required maintenance.

Impact Zone 2 – Quality

Digital twins can help identify the source of quality issues and prevent them, mitigating defects. The quality control expertise of in-house professionals can be translated and introduced to the system.

Then, process-based machine learning is applied to the data, offering predictive quality alerts and their root causes. This mitigates defects and the harm to the bottom line associated with quality deviations.

Impact Zone 3 – Process Optimization

When testing new ideas for optimization, a digital twin is extremely useful in that production can continue uninterrupted.

Even fairly extreme changes can be made to the operation without the need of unplanned stops.

Most importantly, digital twins provide actionable insight into all phases of the product life cycle by:

  • Testing assumptions about the production process by using predictive analysis
  • Building a digital thread that connects systems to improve traceability
  • Visualizing the behavior of equipment in real time.

 

The Impact on OEMs

Digital Twin for factories

Impact Zone 1 – Production & Design

Industrial IoT Digital Twins optimize efficiency by predicting failures in production so that they can be fixed before they affect manufacturing targets. Improvements can be simulated by adjusting parameters along the production line in the twin without risking harm to production. Successful simulations can then be applied to the real-life system.

In addition, Digital Twins of products can be analyzed by engineering teams to compare the actual product behavior to its design. Behavior deviations can be assessed to influence future development iterations of the product.

Impact Zone 2 – Products in the Field

Industrial IoT Digital Twins enable remote commissioning and diagnostics of products that are already in the field – lowering service costs, and improving customer satisfaction.

In cases where a technician is required to physically engage with the product in order to troubleshoot a malfunction, the problem can first be diagnosed remotely via the twin so that necessary equipment and parts can be ordered.

Similarly, when new products are to be commissioned for clients, configuration can be performed by service personnel remotely.

Impact Zone 3 – Future Products

New products can be developed with insights based upon the behavior of existing products in the real world. Performance and customer usage are reflected in the twin, and then feed into the product development and manufacturing process to help boost product margins, and increase customer satisfaction and market share.

 

The 2 Major Advantages of Digital Twin Manufacturing

Visualization

Human learning and decision-making is enhanced through visualization. With the complexity of today's manufacturing processes, it’s not always easy to get an accurate grasp of the factory floor and individual machine status. And when live data is presented to managers in the form of sheets of figures or basic charts, it can sometimes seem too abstract to form the basis for action.

Industrial IoT's Digital Twin offers hybrid visualization, combining visual information with live and historical data. Managers can go under the hood of each and every machine to view physical parameters such as wear-and-tear and temperature abnormalities. Information that isn’t critical can be hidden in order to prevent visual clutter, but can be recalled at any moment.

This level of visualization wasn’t available before and significantly improves the ability to make informed decisions and to identify critical areas that need immediate attention.

Collaboration

The visual aspect of Industrial IoT Digital Twins also lends itself to an improved level of collaboration. Physical distance from the real-world product/system no longer impedes stakeholders from being able to monitor activity and weigh in. This connectivity means that alerts reach management immediately while human errors are mitigated by the removal of single points of failure.

Digital Twin gives access to a broader set of professionals than that provided by a shop floor. Data scientists, product managers and designers gain a much better understanding of the machines at work and the process as a whole. This leads to better design, and processes that are more efficient, saving time and resources, especially those involved in creating prototypes and testing them.

 

Product Line Engineering

When the two advantages above merge, we get a very powerful engineering tool. Specifically, the Digital Twin approach lends itself very naturally to Product Line Engineering (PLE). In complex manufacturing operations, multiple iterations of a product are worked on by multiple teams. This can lead to confusion and human errors. Materials are often wasted as are hours of work.

With an industrial IoT Digital Twin, versions of the product or process can exist side-by-side. These versions can also be divided into iterations per department so that ideas can be tested against specific requirements.

PLE comes into play on the Industry 4.0 field in that it allows for an integrated approach to utilizing Digital Twin and making sure that there’s a strong and fluid connection between the various phases of manufacturing.

 

Getting Started with Digital Twin Technology

A Digital Twin helps manufacturers build better products, prevent malfunctions and errors, and predict outcomes that affect business, but what’s the process for actually building one?

Let’s start off with a useful tip: the first version of any Digital Twin should be moderate in complexity.

Too many sensors and an overwhelming amount of data will be too much to consolidate and will only add confusion to the decision-making process. On the other hand, leaving out critical alerts and not having enough data to produce useful analytics will prevent the twin from reaching its potential as a powerful tool for engineers and managers.

When adopting new technology, it’s important to follow a path that will be challenging, but not overbearing. The ROI needs to be short-term to provide momentum to the project and to make sure stakeholders stay on board and motivated.

Here’s a general plan-of-attack for building a Digital Twin:

Digital Twin implementation
Digital Twin implementation

STEP 1 – Envision
What is the most crucial insight that you would gain? In this first step, imagine situations where a Digital Twin would unlock value by either boosting efficiency or improving customer satisfaction. Focus on processes that are already providing value for the company, and validate the Digital Twin use cases with a team that represents a variety of skill sets.

STEP 2 – Select
Pick your pilot. Your first project should have the highest potential to reap rewards while also having a very good chance of being successfully carried out. An intricate twin of a very specific machine that will allow for deep analysis and monitoring might be a valuable tool with a high ROI, but as a pilot it might be too complex. Broader projects have the potential to scale across the organization, applying to different equipment and processes.

STEP 3 – Implement
Make the pilot a reality and focus on your initial ROI objectives. An agile and iterative development strategy should be employed. Use a visual IoT modeling tool with simulation abilities to design the Digital Twin so that it replicates the behavior of the factory/asset/process.

The pilot you’ve selected should have a limited scope by focusing on a specific part of the business, and must be able to demonstrate its value to the enterprize. It’s important for management and team members to be focused, but open-minded enough to leverage new data collected during the process. As the first signs of success are measured, targets for more significant results should be set.

STEP 4 – Industrialize
The project’s identity should be shifted from pilot to established tool. This should occur through improvements in performance, and the leverage of new twin-derived resources such as a data lake. Insights from the development and deployment process should be shared with other departments.

STEP 5 – Scale
Can you identify opportunities to scale the Digital Twin to realize a more complete industrial IoT platform? Limits originally set to control the scope of the twin should now be removed so that it can add value across the enterprise. Lessons learned and methodologies formed while producing the pilot should be used to improve the process.

STEP 6 – Analyze
The Digital Twin should be assessed against tangible IoT benefits such as improvements in yield, quality and efficiency, as well as cost reduction and prevention of issues. This analysis should provide a springboard for further tweaks to the twin, and its relationship with the enterprise.

In other words, Digital Twin is a dynamic tool, not a set-and-forget project.

 

Model-Based Systems Engineering for a Better Digital Twin

For the initial development stages, and for continuous maintenance, Model-Based Systems Engineering (MBSE) is one method that has been growing in popularity as a way to successfully deliver and manage Digital Twin and IoT projects in general.

Model-Based Systems Engineering is an established engineering discipline, which means it offers a defined pool of professionals with a specific skill set.

As a framework with its own principles and solutions, and because of its data-driven approach, Model-Based Systems Engineering could be extremely useful in successfully delivering a Digital Twin project.

 

The Future of the Digital Twin Concept

The future of industrial IoT Digital Twinning lies in how it will be utilized alongside other emerging technologies such as machine learning, object recognition, acoustic analytics, advanced signal processing and Natural Language Processing (NLP).

Digital Twin is already helping companies better meet the needs of their customers and quickly adapt to new demands from the market, and there’s still a lot to discover.

Get in touch and see how best-in-class tools can help you generate a Digital Twin for your enterprise or asset in just 2 weeks.

New call-to-action


Feature engineering - machine learning for manufacturing

Feature Engineering for Machine Learning:
An Introduction for Manufacturers

What is Feature Engineering?

Feature engineering is recognized by data scientists as a critically important part of designing machine learning algorithms.

To better understand feature engineering in manufacturing, let’s start by defining what a feature is:

On the most basic level, if we take as an example a generic dataset of machinery sensor data like the one below, a feature is simply one of the columns.

Feature engineering: A column of features within a dataset
A column of features within a dataset.

In other words:

a feature is a measurable attribute of the asset or production process that we’re aiming to analyze.

Features can be as simple as temperature, vibration, and pressure readings, but in many cases we will want to add contextual information to the data. In manufacturing, for example, adding categorical information is one technique used by data scientists to make data more contextually meaningful.

 

Feature Engineering in Manufacturing

In Industry 4.0 applications, machine learning algorithms are used to solve specific sub-optimum issues in maintenance, quality, and overall performance. The features determine the focal point of the algorithm’s learning process.

For this reason, the definition of features greatly affects the performance of a machine learning model, and most importantly, how that model will help us solve a manufacturing problem.

While this explanation might seem simple, successfully engineering features is one of the most challenging and time-consuming parts of constructing a successful machine learning application.

Feature engineering is extremely challenging because:

When trying to represent the physical world as a virtual model, there’s always going to be a gap.

What makes feature engineering inherently complex is that it involves the task of representing real-life manufacturing characteristics as purely numerical data.

The goal in feature engineering is to build a bridge that closes this gap between the physical and virtual versions of a manufacturing environment in as accurate a way as possible.

 

Why feature engineering is critical to machine learning that works

Since data is the fuel of any AI system, there’s a tendency to emphasize the importance of screening, aggregating, and categorizing data so that it’s “clean” enough to allow an ML algorithm to operate successfully.

The problem is that these processes demand time and significant development.

When feature engineering is performed properly, a sub-optimal model can still produce excellent results, even when the data being fed into the algorithm is considered messy.

 

Using feature engineering to solve manufacturing problems

Applying feature engineering to manufacturing problems is an iterative process that progresses by cycling through a number of activities including:

Research - learning about how features have been used to transform data in other similar manufacturing problems.

Construction - defining the features to be included in the model. This can be done manually, through automation, or by using a hybrid of both approaches.

Focus - comparing the value of certain features and combinations of features by employing importance scorings and a variety of feature selection methods.

Evaluation - monitoring the accuracy of the model as it works on new data, each time using a different set and configuration of features.

An important step in successfully applying machine learning to manufacturing is to target a well-defined problem. This could be to reduce the number of times a pump’s pressure exceeds a given threshold in a production line. Or, to predict whether the number of quality rejects is likely to exceed 2% in the next 30 minutes.

Sticking to a well-defined problem will keep you on track when it comes to comparing the effectiveness of various models so that you don’t get locked onto trying to optimize one specific model.

Another critical factor is to have a test harness in place. This will allow for the objective testing of models, and will be the only way to gauge your feature engineering efforts.

 

Manual Vs. Automated Feature Engineering

When performed manually, the process of feature engineering can be prone to error. In addition, manual feature engineering is problem-specific - the algorithm cannot be applied to solve other manufacturing issues.

Another shortcoming of manual feature engineering is that it’s limited by the imagination of those working on the problem. You can only think up so many features, before feeling that you’ve done enough.

With automated feature engineering, none of these obstacles exist. A group of related tables can provide data for the automatic construction of hundreds, if not thousands, of useful and interpretable features that can be applied across a range of manufacturing problems.

The ability of automated feature engineering to produce more meaningful features that can also be applied to a large number of manufacturing problems makes it an extremely valuable tool for improving maintenance and overall plant/factory efficiency.

Cut your production losses with machine learning built for manufacturing:

Get a 1-on-1 demo of the Seebo platform and see how accurate and timely alerts can significantly improve maintenance, product quality, and profitability in manufacturing.

New call-to-action


Total Productive Maintenance & Industry 4.0

How Industry 4.0 Brings Total Productive Maintenance
into the Digital Age

Industry 4.0 is disrupting manufacturing on multiple fronts – from production throughput, predictive maintenance and quality, to supply chain and inventory management.

While this wave of innovation is being greeted with much enthusiasm by a traditionally conservative industry, a clear strategy for deployment and ongoing management is required to successfully adopt Industry 4.0 technologies.

Total Productive Maintenance (TPM) is a lean manufacturing approach developed in Japan in 1971. The approach includes a number of methodologies still widely used today such as the Six Big Losses, and is well suited to the smart factory and its IoT use cases.

In this article, we’ll cover:

  • The fundamentals of Total Productive Maintenance
  • How TPM is enhanced by Industry 4.0 technologies
  • Choosing a pilot for TPM implementation in the smart factory

What is Total Productive Maintenance?

In a nutshell, Total Productive Maintenance is a system for optimizing maintenance and reaching a state of perfect efficiency in production.

TPM focuses on driving efficiencies by organic means ie. by using existing company resources.

The main goals of Total Productive Maintenance are:

  • No short stoppages or sub-optimal production rates
  • No defects
  • No unplanned downtime
  • No accidents

 

The Origins of Total Productive Maintenance

The TPM approach is credited to Nippondenso (known today as Denso Corp.), a company that created parts for Toyota.

Dissatisfied with preventive maintenance methods carried over from the United States in the 1960s, Seiichi Nakajima, considered the founder of TPM, promoted the idea that factory workers should take on a wider range of responsibilities regarding the upkeep of machinery.

Instead of every machine/production line having separate employees for operation and maintenance, employees would be trained with the tools and knowledge to do both, giving them a more holistic approach to ensuring machine health.

The TPM methodology creates a shared responsibility amongst factory workers and an increase in morale and pride in the efficiency and condition of machinery and production.

 

5S – The Foundation of Lean TPM

TPM has as its foundation in another management methodology, also originating from Japan, known as 5S.

5S is focused on organizing the workplace environment to improve efficiency and effectiveness through 5 main activities:

Sort (Seiri)

Action – the equipment and materials in an area are sorted, with unnecessary items moved to another area or disposed of.

Result – reduced clutter makes inspection easier, frees up available space, and makes things easier to find.

Set in order (Seiton)

Action – Tools and equipment are placed in a way that suits the work. Tools that are most frequently used are easiest to reach, and storage is marked to make it easy to return items to their designated spots.

Result – Smoother workflow.

Shine/Sweep (Seiso)

Action – Work areas, tools, and machinery are cleaned and inspected regularly.

Result – Slower deterioration of equipment and infrastructure, and improved safety.

Standardize (Seiketsu)

Action – Employees are informed of the different procedures in detail and provided with an organized schedule, clear instructions and necessary on-site visual aides in the form of markings, photographs, and illustrations.

Result – Procedures covering the first 3 “S” practices are scheduled, performed regularly, and monitored.

Sustain/Self-discipline (Shitsuke)

Action – Training sessions are set up and regular monitoring is done to ensure compliance.

Result – The 5S methodology is followed not because workers are told to, but because they choose to, initiating additional improvements through experience.

 

Industry 4.0 & The 8 Pillars of Total Productive Maintenance

With 5S as a foundation, TPM proposes an 8-pillar approach that aims to cover every possible aspect of maintenance in the industrial manufacturing setting.

Industry 4.0 & The 8 Pillars of Total Productive Maintenance

Here is an outline of these 8 pillars along with how Industry 4.0 can take this approach even further:

1. Autonomous Maintenance

Probably the most unique characteristic of TPM – the idea here is that the people working with a machine on a day-to-day basis are the most “in-tune” with its behavior and performance.

Operators are trained to claim “ownership” over their machines, taking care of routine maintenance activities such as cleanliness, lubrication and inspection, and should be the first to attempt handling issues within the realm of their training, before calling upon expert technicians.

With Industry 4.0: As machines become more automated, monitoring improved, and dashboards easier to read, operation will become less complex making the “ownership” suggested by TPM much simpler and therefore more accessible to workers.

2. Planned Maintenance

Maintenance preempts malfunction while interventions by high-level technicians are carefully planned so that minimal downtime is required for any software updates or part replacements.

With Industry 4.0: Using predictive maintenance by means of machine learning, maintenance activities are only performed when necessary and can be timed to avoid downtime completely.

3. Quality Management

Workers are trained and encouraged to identify issues in production that ultimately lead to defects and quality issues.

With Industry 4.0: Enter “predictive quality” – sensor data and machine learning help identify anomalies in machine behavior, alerting operators, who can then perform focused root cause analysis. Problems can be corrected much earlier than what was previously possible, reducing the financial damage of quality deterioration and defects.

4. Focused Improvement

Cross-functional teams are formed and proactive involvement is encouraged. Problems affecting production are tackled by workers who start with the major hindrances/showstoppers, moving down to more minor inefficiencies.

With Industry 4.0: Through organized data collection and the application of artificial intelligence algorithms (eg. artificial neural networks), less obvious correlations between defects and root causes can be exposed. Inspection information and hypotheses can be shared company-wide, allowing for better synchronized and more successful collaboration.

5. New Equipment Management

The design and installation processes of new equipment should be planned based upon previous experiences to ensure that performance targets are reached quickly with minimal startup issues and for improved safety.

With Industry 4.0: Production data in historian systems can be analyzed to identify best practices from previous installations/designs while taking into account current plant/factory conditions.

6. Education & Training

See Pillar 1 – operators receive training giving them the necessary skills to maintain machinery and identify problems. In turn, maintenance technicians learn approaches to more proactive work while managers are encouraged to improve leadership skills.

With Industry 4.0: Digital Twin visualization provides an excellent opportunity to learn about the complexities of manufacturing on all levels: from components and machines to production lines and overall facility management.

Educational content can be online and available to employees 24/7. Novice personnel can be assigned experienced mentors who can have access to their activities and be available to answer queries.

7. Safety, Health & Environment

A safer work environment is created by identifying health risks and potential hazards and working to eliminate them. Uncomfortable conditions harm productivity and employees should not be expected to be productive while at risk.

With Industry 4.0: Sensors can measure air quality, radiation, temperature and other environmental conditions which may affect health and performance while the early detection of harmful gases, electrical surges and fire can save lives and prevent damage to equipment.

8. Administration

The TPM approach can be applied to systems that aren’t directly involved in manufacturing, including office administration. The significance of including administrative functions as one of the eight pillars is that this level of management – order processing, scheduling, workforce management, accounting – should be in sync with the other facets of the facility through effective communication, transparency and tried and tested protocols.

With Industry 4.0: Artificial Intelligence algorithms are very well suited to analysis and decision-making processes making this technology extremely advantageous to office automation.

Office automation is destined to develop and will include demand forecasting, intelligent pricing, and smart purchasing and outsourcing.

According to the Total Productive Maintenance approach, achieving excellence in each of the 8 pillars mentioned above is verification that a manufacturing facility is producing “World Class” results.

Take the free whitepaper with you

TPM Cornerstone Banner

OEE & TPM

TPM gave birth to one of the most widely used KPIs in manufacturing – Overall Equipment Efficiency (OEE).

OEE is an important metric in TPM, used to gauge the facility’s overall efficiency status.

If we take a look at the goals we previously outlined for TPM, it becomes apparent how these line up for OEE calculation:

TPM Goals Vs. OEE Calculation
TPM Goals Vs. OEE Calculation

 

Applying Smart Factory Solutions for Next-Level TPM

When implementing TPM in Industry 4.0, it’s a good idea to start with a proof of concept, analyze, and then scale up to bigger challenges. Deciding upon the right pilot is an important first step in the implementation process.

Consider these 3 levels of complexity when choosing a TPM pilot for your Industry 4.0 project:

1. Simple Improvements

Pros

Initiating a small improvement is a good opportunity to score a “win” in a short amount of time and doesn’t require a deep level of TPM knowledge. This is a good pilot type for recruiting stakeholders and building confidence in the process right from the start.

Cons

Making only a small improvement will result in a relatively low ROI for the project and won’t yield that much information on the TPM process.

2. Optimization

Pros

By addressing a bottleneck in your production line or relieving a constraint, you’ll see an immediate increase in total output.

Cons

Focused optimization may require some planned downtime for experimentation and analysis and there is the risk that you may not achieve a measured improvement on the original output rate.

3. Solving a problem

Pros

Solving a longstanding issue with a machine/process will garner support for TPM implementation and will be received well by operators.

Cons

ROI might be relatively small. Also, a complex problem might be too much of a challenge as a starting point, causing the project to lose momentum.

 

The Impact of Industry 4.0 on Total Productive Maintenance

As Industry 4.0 advances, further disrupting the way products and materials are manufactured and the market itself, new factory/plant management issues will arise.

To meet these challenges, managers will do well to make use of methodologies such as TPM to smoothen the transition into Industry 4.0, and to ensure bottom-line impact through improved output rates, quality, and customer satisfaction.

 


An intelligent maintenance system (IMS) uses sensors, hardware processors, cloud applications and advanced analytics to improve the performance of maintenance for a machine, production line or manufacturing facility.

What’s So Smart About Intelligent Maintenance Systems?

Being a significant expense of any manufacturing company (commonly between 10% to 15% of total operating costs), maintenance has long been the focus of production efficiency efforts.

Industry 4.0 technologies such as digital twin and artificial intelligence are proving that even the best results from traditional maintenance methods can be improved upon by at least 30%.

That figure is a serious game changer for any manufacturer.

To achieve maintenance efficiency, a number of issues should be addressed:

  • In the case of a specific machine’s failure, what is its impact on the factory/plant’s throughput?
  • In the case of multiple machine failures, which maintenance job should be executed first?
  • What is the effect of the various failure types on production throughput and quality?
  • Which maintenance activities are possible to perform without affecting the production schedule?
  • What’s the most efficient use of factory/plant resources - labor, materials and equipment - for performing maintenance?

In this post, we'll look at how Intelligent Maintenance Systems help answer these questions, and in turn, improve maintenance efficiency.

 

So, what is an Intelligent Maintenance System?

“Intelligent maintenance system” is an umbrella term for a number of approaches that share a common goal: to improve the efficiency of maintenance activities through the use of digital technologies.

An intelligent maintenance system (IMS) uses sensors, hardware processors, cloud applications and advanced analytics to improve the performance of maintenance for a machine, production line or manufacturing facility.

 

How manufacturing facilities benefit from utilizing an IMS

The challenge in answering the questions above lies in the fact that they all depend on a large number of interlinked parameters; parameters that change over time, both in the short and long-term.

Fluctuations in environmental factors, the quality of raw materials, asset health status, workforce, market demand, and many others affect the required rate and type of maintenance.

So, how can manufacturers gain control over ever-changing operational environments?

The answer: by being able to predict.

One of Industry 4.0’s most powerful use cases is predictive maintenance which leverages data captured from sensors, PLCs, data historians, ERPs, MESs etc. to form failure predictions.

Machine learning and other AI algorithms such as artificial neural networks are used to process this data constantly. The algorithms search for correlations that can help determine the root cause of recurring problems that lead to unplanned downtime.

artifical neural network schematic
Schematic of an artificial neural network. ANNs are used to discover causal correlations between root causes and failures.

The ROI of IMS

An IMS with predictive maintenance capabilities will autonomously alert relevant personnel about issues and only recommend maintenance activities when necessary.

In this way, maintenance efficiency is significantly improved, offering manufacturers a number of benefits that positively affect the bottom line:

Lower maintenance costs - Repairs are performed when needed instead of according to a predetermined schedule (which in many cases leads to redundant checks/maintenance activities).

Increased uptime - With IMS, downtime can be scheduled, and as a result is usually much shorter than what’s needed for reactive repairs.

Reduced labor costs - Maintenance teams are smaller since tasks are planned beforehand.

Lower equipment costs - Maintenance focuses on the problematic components only. This prevents wear and tear to surrounding parts during the repair and prevents unnecessary replacements.

Lower inventory expenses - Since problems are predicted, orders are made only for materials and components that will be needed in the near future.

Lower chance of secondary damage - Because problems are detected early on, they can be dealt with before more extensive damage is done to equipment.

Increase in Remaining Useful Life (RUL) - The root cause of issues is pinpointed making the need for disassembly less frequent.

Quality 4.0 - By improving asset health, deviations in performance become far less frequent, leading to consistently high levels of output quality.

 

Maintenance doesn’t have to be a sore point

By deploying an Intelligent Maintenance System, today’s manufacturers have the opportunity to overcome maintenance challenges and gain control over complex production issues.

 

Contact us to find out how your operation can reduce unplanned downtime and prevent quality deviations using intelligent maintenance.

For more information about everything Industry 4.0, head on over to our free resource library.

 


In manufacturing, digital twin solutions represent “real-world” equipment, machines, factories, and plants.

5 Ways Digital Twin Solutions
Predict and Resolve Manufacturing Issues

A digital twin is a digital replica of a physical asset, device, system, place, or person.

In manufacturing, digital twins are used to represent “real-world” equipment, machines, factories, and plants.

Sensors placed upon machinery across an entire production line, for example, feed data through an IIoT network to be accessed via a dashboard. The data is displayed as a visual model - the twin - and acts as a live representation of that line.

This offers the ability to drill down into specific assets for improved root cause analysis and a host of new use cases for improved manufacturing operations.

Digital twins are helping companies surpass previous performance levels in various aspects of manufacturing, from the product design phase to supply chain management.

Here's how:

1. Monitoring

A digital twin merges live data from its physical counterpart with an interactive visual interface. This offers an unsurpassed level of monitoring.

Asset and production line performance data is presented in the context of the physical device, usually with a 3D visualization of the part/machine/line.

This enhanced visual interaction improves the ability to formulate actionable insights based upon the presented data.

By combining next-level monitoring capabilities offered by digital twin with machine learning algorithms, manufacturers are able to perform automated root cause analysis to prevent recurring asset failures and quality deviations.

The Seebo Predictive Maintenance Software Platform
The Seebo digital twin dashboard. Predictive maintenance and quality alerts are provided in the context of the production process.

2. Maintenance

The improved monitoring capabilities mentioned above feed directly into maintenance efficiency.

Operators and technicians are provided with detailed information about the health of every asset. This leads to insight that can be acted upon directly to improve OEE.

A digital twin isn’t only a graphical model. Predictive analytics powered by machine learning and other AI algorithms dissect the data, search for correlations, and formulate predictions about remaining useful life (RUL).

Digital twin solutions take maintenance from a reactive to a predictive approach. Repairs are only performed when needed and components aren’t switched out unnecessarily.

Since maintenance is planned and only done when necessary, technical teams need not be so large and arrive on-site with the right tools and parts, as well as precise instructions for the repair procedure.

3. Training

Digital twin is an excellent tool for professional training due to its visual interface and the fact that it mirrors real-life scenarios from the production floor.

Digital twin can be used for broad-topic training such as site orientation and safety protocols or specific technical training such as repair and installation procedures.

4. Communication

A digital twin can play a vital part in helping employees share knowledge about production issues.

Automatic alerts about predicted failures or quality deviations can be viewed by all the relevant personnel. Tips and advice can be shared, including highly specific technical details thanks to the visual nature of the twin.

This information is archived and accessible to personnel in other facilities, streamlining root cause analysis and preventing the unnecessary repetition of mistakes.

5. Strategy

A digital twin solution can be extremely useful in testing new concepts for optimization without needing to disrupt production.

Major changes can be made to the operation without going into downtime.

A digital twin can provide insight across all stages of the product life cycle by:

  • Refining assumptions through the use of predictive analytics.
  • Establishing a digital thread that connects individual systems to improve traceability.
  • Visualizing the use of products in the field, in real time.

 

How to find the right digital twin solution for your operation

Manufacturers looking to deploy a digital twin should evaluate candidate solutions against the following key capabilities:

Simple visual interface - The digital twin platform should be clear and intuitive, easy to navigate, and simple to interact with for personnel of all backgrounds: process engineers, maintenance and quality teams, plant managers etc.

Detailed and process-based - The twin should be capable of displaying a wide range of data from individual components to production lines consisting of multiple machines as well as data on HVAC, inventory, and supply chain systems.

Big Data Processing - The twin, along with the industrial IoT network set up in the facility, must be able to handle large volumes of data simultaneously for real-time processing.

Analytical - The digital twin solution should offer a variety of AI algorithms that are easy to apply and that output actionable insights that are relevant to the needs of your operation.

Customizable - Every manufacturing facility is different. Dashboards, alerts, and other system settings should be configurable to suit the nature of the work at your plant or factory.

Download our free digital twin whitepaper, or get in touch for a 1-on-1 demo to see how a digital twin can significantly improve your operation’s uptime and output quality.


Tips for Choosing Predictive Maintenance Software

Looking for Predictive Maintenance Software? First, read this...

Predictive maintenance (PdM) is probably the business use case most responsible for drawing manufacturers to Industry 4.0 adoption.

Preventive maintenance for production line assets (common practice until the advent of Industry 4.0) is a major cost burden for manufacturers. In fact, 40% of all preventive maintenance costs are spent on assets with a negligible effect on actually preventing failures.

It's no surprise that the ability of predictive maintenance to predict unplanned downtime events and quality deviations is proving to be a game changer.

This is due to the 8 main benefits of predictive maintenance:

  1. Reduction in lost production time - PdM allows for planned downtime which is usually much shorter than what's needed for reactive repairs, and can be scheduled for times that are convenient and less costly.
  2. Reduced maintenance costs - Repairs are done when needed, instead of routine maintenance which in many cases is redundant.
  3. Lower labor costs - Technicians are called upon for specific and focused tasks.
  4. Reduced equipment costs - Only the problematic components are dealt with, preventing unnecessary replacements and the wear and tear of adjacent parts caused by repair.
  5. Lower chances of secondary damage - PdM identifies problems early on before they escalate and cause more extensive damage to equipment.
  6. Reduced inventory expenses - With PdM, orders can be made only for the parts and materials that are needed.
  7. Longer lasting machinery - Since disassembly is carried out less frequently, equipment lasts longer, increasing remaining useful life (RUL).
  8. Reduced risk-based costs - Fewer unplanned repairs reduce safety risks and the chance of damage being done to other parts or equipment.

 

Preventive Vs. Predictive Maintenance
Preventive Vs. Predictive Maintenance. While preventive maintenance has been common practice for decades, predictive maintenance requires less labour and is significantly better at preventing failures.

What is predictive maintenance software?

Predictive maintenance is a method of preventing machine failure by analyzing machine data to identify patterns and predict issues before they happen.

At the intersection of human-machine interaction within the smart factory, predictive maintenance software gives manufacturers monitoring and control over PdM capabilities.

For this reason, it’s critical that PdM software offer users visual, real-time interaction that’s accurate,  reliable, and can be customized.

What to know about your company’s needs

Every manufacturing facility operates differently. Clearly understanding your operation’s needs, priorities, and economic dynamics is key in implementing the predictive maintenance software that will deliver the highest ROI for your operations.

The opinions and experience of engineers, technicians and management personnel should be included when deciding upon which software to deploy.

SEE ALSO: Pick solutions, not platforms - how to choose an industrial IoT Platform

Answer these 8 questions to get a better idea of your company’s needs:

  1. What is the monthly cost of unplanned downtime to the business?
  2. What are the common root causes of production disruptions?
  3. What are the common root causes of quality deviations?
  4. What type of problems are you looking to predict - mechanical failures (e.g. motor or bearing failures) or process-centric issues (e.g. deviations from recipes)?
  5. What is the level of in-house data analytics expertise?
  6. Who will be using the software?
  7. Is data from the production lines (OT) accessible to external systems? Is it persisted in a database (e.g. data historian)?
  8. Is data from operational systems integrated with data from business systems (IT) and process flows to enable effective and accurate data analytics?

What to look for in a predictive maintenance software solution

Here are 6 key capabilities to consider when evaluating predictive maintenance software solutions:

Industrial artificial intelligence - AI algorithms integrated into the platform that can be used to drive efficiency through use cases that are relevant to the specific manufacturing type. For example, mechanical failures and process-related issues each require specific types of machine learning.

Simple and intuitive - Predictive maintenance software should be easy-to-use for operators, technicians and management. For example, a digital twin interface can display PdM insights and supporting data in the context of the production line, allowing for quick and intuitive root cause analysis.

Human in the loop - Predictive maintenance software should be able to receive input from production line experts in parallel to data from sensors. In this way, human experience can be leveraged for more accurate predictions as the algorithm learns from expert knowledge.

Actionable and prescriptive - Insights from the software should lead to action, with information on exactly what needs to be done, and how. By pinpointing a predicted failure, technicians perform the prescribed set of corrective actions by accessing and checking off standard operating procedure tasks.

Measurable outcomes - Predictive maintenance software should be able to report on its value and the improved business outcomes using quantifiable metrics.

Compatibility - It’s common for production lines to consist of industrial assets manufactured by a variety of OEMs. PdM software should have the ability to work with different types and brands of assets, seamlessly connecting data historians and PLCs while integrating IT systems (ERP, MES, QMS, etc.)

 

A condition monitoring dashboard of the Seebo platform showing predictive alerts.
The Seebo platform. Predictive maintenance and quality alerts are provided in the context of the production process.

Always check what’s in the box

The predictive maintenance software market is diverse both with regards to the solutions themselves, and the support and accompanied services offered by software vendors.

Make sure you know exactly what the PdM software does - what’s included, and what’s not.
Many software solutions require significant customization to be performed by the vendor’s professional services. Make sure to factor in the extra costs of additional service providers or software applications.

Planning is everything

Once all of the above has been taken into account, there should be one underlying principle guiding you to the right choice - quick and easy deployment.

Consider solutions that can be deployed to deliver value in under 3 months, and that can be continually and easily adapted as the business requires.

Leading vendors will be able to offer you a customized deployment timeline. The timeline should include all the various deployment stages to help you configure the PdM system through prototyping, validation and finally, solution delivery.

For more in-depth information about predictive maintenance, download our free whitepaper.

Get in touch to see how PdM can significantly cut your maintenance costs and improve the performance of your production line.