Industrial analytics thought leadership-series.
How will industrial analytics, Industry 4.0, IIoT and digital transformation change businesses? And ultimately the world? We ask industry leaders to share their view on these technologies, how we use them, and where they will take us.
Industry will change analytics; analytics will change industry
Megatrends in the world will give industrial analytics at molecular level a bright future. We will see a global standardization in process analysis giving us continuous manufacturing with small scale and flexibility to meet the need for personalized products. Professor emeritus Rudolf Kessler underlines the transdisciplinary element in industrial analytics and the key role of science.
“Production and industry is the key driver for advanced industrial analytics. A high percentage of financing of all industrial AI has production as the true driving force. I see Europe playing a key role in process analysis and industrial analytics in the future because of the European capabilities in manufacturing,” says professor emeritus Rudolf Kessler.
He has personal work experience from Mercedes Benz Research Labs, and Rudolf Kessler sees the European strength in the many cultures and a big heritage of skills. Process analysis is transdisciplinary both in terms of technology and people. Process chemists, process engineers, chemometricians and many other technologists must work together putting emphasis on the science of each discipline. The fact-based approach of linearity, causality and first principle thinking will be a vital European contribution to the technology, according to Kessler.
Data is more than numbers
“After the financial crisis the dangers of ruthless capitalism is very clear. Also, in terms of technology we should be aware of using the free liberal market with responsibility. In Germany we talk about “soziale Marktwirtschaft” (“social market economy”) with the emphasis of doing business in a socially sustainable way.”
“Working with data gives you an ethical responsibility. Some AI-people do not care about the data, and to me that is wrong. Data is nothing alone. We need to know and understand the data coming in to understand and interpret the answers coming out. Garbage in is garbage out,” he says.
The interdisciplinary aspect
Focusing on the interdisciplinary aspects in industrial analytics is the key to leverage the progress in hardware and software, according to Kessler. Key concepts like PAT, DoE, QbD relies on the joint effort of scientists, process experts and data scientists.
“I’ve been working in the chemical engineering society to get people with different skills, traditions and roles to work together and learn from one another. Next meeting is in Copenhagen 2020.”
“In an even greater scale I think it’s important for our societies to have the dialogue about technology with scientists, vendors, industry representatives and politicians all contributing,” says Rudolf Kessler.
The industrial viewpoint on AI
From the starting point in sensors and spectroscopy in combination with data science Rudolf Kessler has along the way broadened his perspective to process analysis in a wider sense.
“Also, as a scientist nowadays you need to do marketing. You need to go to mainstream to get money for your research. The same trend is apparent all over the places and even in bigger companies. The “dinosaurs” in corporate life need to have an AI strategy,” he says.
Kessler distinguishes between deep learning and machine learning. The first may be relevant to analyze the mood of people in the face recognition in China, while the latter will become relevant in industrial analytics and much closer to Kessler’s field of interest.
Informatics captured the phrases
“Many data is not just good. You need correct data and wise data. Causality is important and to establish the causal relations in data you need robustness every step of the way. It’s so complex, and your plan is easily corrupted.”
Rudolf Kessler has been working in industrial environments all his life. His main expertise is on the sensors hardware relevant to implement industry 4.0 and industrial IoT in the processing and manufacturing industry. The starting point of Industry 4.0 and industrial IoT was more on the hardware side than on data analytics alone.
“The informatics people “captured” the phrases and transferred analytics into data analytics. In the early stages we talked about “instrumental analytical chemistry”. I still “fight” for a broader view and for a strong dialogue between the processing industry, the sensor manufacturers and the data analytics people,” he says.
The future of industry
The trend in industry is building new production systems for smart personalized products, the trend in sensor manufacturing is sensors based on information on a molecular level and the trend in analytics is the science-based search for causality. According to Rudolf Kessler this in part answers to some global trends.
“Customers demand personalized products, which leads to small lot sizes and thus a need for decentralized structures in manufacturing. This coincides with the demand from an aging population with need for personalized care technology and specific medical treatment. Another trend is the urbanization with many people moving from rural areas to urban cities creating more mega cities with populations over 10 million,” he says.
According to the German Research Center for Artificial Intelligence (DFKI), the future for industrial automation is the smart factory. The smart factory will be flexible with modifiable and expandable units, networked with components connecting to other components, self-organizational with components performing tasks related to its context and it will be user-oriented with a great emphasis on ease-of-use.
Process analysis in leading role
“It’s about smart materials being processed in smart factories with smart sensors to produce smart products. The smart factory will allow small batches of goods that you can individualize for the customer,” says Rudolf Kessler.
“In global scale we will see industry go back into the city centers to integrate individuals to the new way of producing. When manufacturing turns to small and decentralized systems, process analysis will play an important role in designing and controlling the smart manufacturing,” says Rudolf Kessler.
Find the right solution for your analytics needs.
Get in touch if you have questions about our products, platform, how to get started or how best to address your analytics needs.