We live in a new era of industrialisation, in which the evaluation of data plays a key role. However, according to the German Industry 4.0 Index 2018, 94 percent of German industrial companies currently see no great benefit in predictive maintenance based on process and machine data. The new Commerzbank SME study shows a similar picture: only 32 percent of industrial companies use new technologies for individual production.
Why is that? One reason is probably that many responsible persons cannot imagine the opportunities Industry 4.0 offers their business. However, machine data is one of the most valuable resources of a company, for example, if they are used to reduce downtime, to make production more efficient or to optimise maintenance intervals. Another reason for these reservations is probably the unmanageable flood of data: only those who know which data are important can extract, analyse and exploit them. That's why connyun Industry 4.0 considered this issue from the point of view of the customer, and made it easier and faster for companies to access. The basis for this is Data Science Consulting.
The connyun Data Science team actively advises companies and demonstrates the opportunities that data can offer. The first step is for the Data Science experts to resolve prejudices and provide transparency. The options for cooperation are specified in the second step. The scale of a Data Science project is immaterial since even the smallest data packages hide valuable information. The third step is to clarify how data is exchanged and what results can be expected from the two to six weeks of data analysis. These first insights are often the start of a deeper investigation and evaluation.
Data Science in manufacturing must be pragmatic and can help to improve production forecasts, reduce machine and cell downtime, schedule maintenance more efficiently and better align downstream processes such as logistics. In practice, our connyun Data Science experts are mostly involved with the following topics:
Predictive maintenance: predictive maintenance saves costs
Frequently, industrial machines are maintained at fixed times, regardless of actual usage or wear. If maintenance comes too late, this can lead to unplanned shutdowns and, thus, high losses. To prevent this, many machines are serviced more frequently than necessary, which incurs unnecessary costs. Using data analysis in close cooperation with the our customers’ own experts, we can plan the ideal maintenance intervals for each individual component. This is done on the basis of machine usage, measured wear, historical usage profile and the current order information. Thus, the connyun Data Science team creates the conditions for intelligent, predictive maintenance, which can save costs on the principle "as often as necessary, as seldom as possible".
Anomaly detection: faults that are not recognisable to the naked eye
Most industrial plants today are technologically perfected and run with low errors. But just this high degree of perfection has its downsides: if a technical defect occurs - e.g. caused by an operator - this often remains hidden for a long time. The longer the defect remains undiscovered and thus unresolved, the higher the costs it causes, for example, through higher wear and repair costs, rejects, recalls or unplanned shutdowns. For the operators, such defects are very difficult to detect. However, the anomalies are clearly identified in the data. connyun's Data Science experts clearly identify and isolate these defects as deviations from standard behaviour using machine-learning. Building on this, improvement measures can be initiated immediately.
Planning, material flow, logistics: it gets better still
Experience belongs to the intellectual capital of every company. Any machine operator with decades of professional experience will confirm how important this learned knowledge is. This human competence is much more effective when combined and supported with data.
For example, let’s take "manual" production planning based purely on empirical values. In this case, the responsible production employee must be very capable at assessing capacities and resources, in order to plan order acceptance accordingly: can it be done or not? Mistakes here will lead either to incomplete utilisation of capacity or unfulfilled orders. Supporting this experience with objective data offers better planning, resulting in more precise and reliable statements in a much shorter time.
The same applies to logistics planning: predictions for completing batches are much more reliable if they are done using data analysis. This allows more accurate adaptation of storage and removal. This shortens storage times and prevents shipping waiting times.
All planning tasks that involve predictions, assessments, and experience become faster and more accurate when done on a data-driven basis. When must which tasks be fulfilled and using which machine, so that everything happens on time and no one has to wait? Where does raw material have to be replenished next, and how much time is left for other tasks? Those who are able to make precise statements here use capacity more efficiently, reduce machine downtime and increase productivity.
Data is raw material, knowledge is power: data mining
Data Science tools and methods are mature, proven, and affordable today. Manufacturing data analysis can be used directly and easily to reveal further potential even at peak times. Customers frequently ask themselves: Where are the hidden sources of error in my production? How do certain factors affect productivity? Why does the same mistake occur again and again? connyun's Data Science experts investigate these questions, and often establish that the clues are already there, but hidden within the data.
The secret of successful data mining is not just the analysis of complex data. In addition, connyun has technical expertise and in-depth industry knowledge of production and mechanical engineering. With this combination, our Data Science experts reach understandable and helpful solutions together with customers that save costs, increase productivity and create transparency.