How Industry 4.0 Contributes to Operational Excellence | Lecture notes from Jose Ignacio Erausquin | Madrid, 5/22/2019

At the invitation of our Spanish partner Asenta, Michel Baudin gave a lecture in Madrid on the ways Industry 4.0 does or can contribute to operational excellence. Industry 4.0 was presented as a stack of technologies — from direct machine control to knowledge management — with each layer relying on the layers below. Following are the notes taken by Asenta’s Jose Ignacio Erausquin, organized by layer.

1. Machine control
   • Companies should care more about the value of their machine control programs which is their own know-how. The value of CNC and PLC programs exceeds that of the machines themselves as a differentiator from competitors
   • Let’s use all that technology offers to use systems that allow us to fully automate CNC Programming. Companies like Plethora are showing the way.
   • Currently we have plenty of opportunities to use very cheap and quite powerful microcontrollers that can help us perform simple automation in Kaizen projects and incremental improvement.
2. Process monitoring systems
   • Human-machine Interfaces (HMI) are key for a friendly automation and to keep automation to the service of workers and not the other way around. Think about improving these interfaces when retrofitting machines.
   • Think about people when automating processes. Many things can be done today based on technology to help people, avoid mistakes…
   • Automation may be used for information flows, as well as material flows. Simple systems like RFID and other can help track products, interchange information, provide input for orders, and so on.
3. Operational support
   • The highest difficulty/challenge in the factories usually is collecting and organizing information from multiple sources. Not all ERPs are good at everything, as they have been originally created for different purposes.
   • A recurring issue when discussing the application of technology is that priorities change from business to business and that we may need different solutions depending on the case. For example, should we use paper or screens for instructions? The answer depends on the speed at which instructions change in the business. Screens are more effective in rapidly changing environments.
   • Let’s try to avoid the “horseless carriages” syndrome of not taking full advantage of the potential of the technology. Don’t just copy what you used to do and apply technology to automate it.
   • Be sure that application of technology is really helping people to do better work or to do it more easily. Sometimes technology allows us to do very impressive things but this is not always practical. For instance, augmented reality can help operators, but at what development cost?
   • Simplify operations before automating. Don’t build an electronic version of the 48-page traveller, simplify it first. For instance integrating some operations in a flow line production eliminates the need for step-by-step tracking through this line. Maybe technology is no longer needed to understand the position of the workpiece.
4. Engineering/Management support
   • One of the biggest problems is that support systems are not managed in a common way by Engineering and IT. Typically, Engineering and Operations like bottom-up approaches to adapt information systems to their needs, but IT will try to reduce proliferation of systems and not allow departments to stray from a centralized approach.
   • There is new technologies to manage and analyze Big Data (BD), meaning multiple terabytes of new data per day, but usually this is not needed in manufacturing. Maybe you can use a sensor and collect process temperature in a furnace every millisecond, but this doesn’t provide the kind of information we need in manufacturing. So the very important thing is not so much the ability to manage BD analysis, but to properly select which data we need to manage the process.
   • A manufacturing plant requires a common information model to provide simple answers to simple questions from muliple systems. The top-down imposition of a single system usually don’t work well, as we can see from the proliferation of Excel spreadsheets everywhere. This requires Engineering, Operations and IT to work together to define and map the company information. This important task should never be completed just by software development specialists.
   • These maps should identify the different types of data and the different types of technologies needed, such as databases, security tools, visualization tools, etc.
   • IT systems should not only focus on collecting data. They should be more about analyzing them consistently and providing the results of these analyses in a very friendly way to support decision taking.
5. Knowledge management
   • Just because data is available don’t necessarily mean that we are sharing knowledge. Probably knowledge management (to learn from others) is one of the most difficult tasks in a large company. Some things work better than others in knowledge transfer, i.e. forums. If we think knowledge transfer is important in pour company we have to create such forums in top of making information available.
   • There is a long way from getting data available to get knowledge. These are the main challenges to have to face to make it possible.
     1. You have to be able to get the data. Define the required data and make it accessible. Choose only required data, not everything you can collect. Typically, in the manufacturing world, e-commerce methods (Big Data) will not be required.
     2. Ensure data quality. Data should “mean” something without the need of personal interpretation, which introduce bias and subjectivity. For instance, using “smart” numbering systems is a source of errors and don’t make much sense in the era of computers. Information should be structured in databases pulling together all the relevant elements. Data credibility is an issue in most organizations and we must make sure that data collection is reliable and error-free.
     3. Analyze data to discover. There are so many data analysis tools that it’s almost impossible to be an expert in all of them. To select the proper data analysis tools we have to understand a little more about who invented such method, what kind of problem he wanted to solve and identify if we have a similar problem. In manufacturing problem solving, we often need to understand how a process moves over time, and the data often comes in the form of time series.
     4. Make manufacturing data talk. Make it easy for decision makers to understand at a glance what is happening. User-friendly communication is key to facilitate action. Avoid fancy charts that confuse readers. This includes radar charts, 3D stacked bar charts, 3D pie charts,… Adapt the charts and the communication system to user needs. If the way we use to represent reality is too complex, then users don’t understand well what happened.

Conclusions and key takeaways from the lecture

1. Automate the production of value, not waste
2. Automation must be developed at the service of people, not the other way around
3. Bottom-up approaches respond better to real needs than top-down approaches
4. Continuous Improvement always work better that “breakthrough automation”. Today managers are “desperately” in the search of a “Silver bullet” to solve all their problems at once, but it doesn’t exist.
5. Pull knowledge together. The different skills need to be combined. Let Engineering and IT work together.
6. Digitalization and Industry 4.0 are not the aim in itself, but can be effectively used to pursue operational excellence.

#industry40, #jidoka, #autonomation, #humanmachineinteface, #HMI, #ERP, #BigData, #machinecontrols, #excel