Searching for traces on the shop floor with process mining
On the way to the digital twin of production
Process mining is more in vogue than ever and aims to reconstruct inefficiencies on the basis of digital traces in IT systems, thus clearing the way for optimization. Since data-based processes are assumed in this detective work, this approach is located in the ERP environment. But especially when it comes to analysis on the store floor, using an MES like PROXIA is the better choice.
Time and again, downtimes during the valuable operating time of the systems on the store floor – the MES sounds the alarm! It aims to draw attention to a reduction in Overall Equipment Efficiency (OEE). Often, in such a case, the blame is placed only on the workers or the production equipment. However, it may be that everything was done correctly in these sub-processes and the problems occurred elsewhere, for example in the provision of materials or in other sub-processes of the company organization. Only late, often too late or in many cases not at all, the true cause is recognized.
This is exactly where process mining comes in, combining process-oriented business process management with non-process-oriented data mining. Compared to data mining, process mining has the advantage of being able to assign and visualize the collected raw data to specific (sub-) processes. This makes it possible to monitor and improve the overall process in its entire granularity. Especially with an increasing degree of automation and increasing correlation of formerly autonomous individual processes, accumulating downtimes can be detected by the exact reconstruction of dependencies. Cause-effect relationships can thus be traced without having to lose oneself in lengthy analyses. After the correlations have been identified, the collected data can be further analyzed using process mining in conjunction with data mining methods in order to derive decision rules, for example.
CIP – on the way to the digital twin of productionProcess Mining supports the continuous improvement process (CIP), the supreme discipline in production. If we look at the classic PDCA cycle (plan-do-check-act), it is noticeable that an important phase is missing from the consideration; however, it is assumed in the models. The “realize” phase is the section where the PDCA cycle actually begins. In this very phase, the MES of PROXIA Software AG offers the appropriate tools to detect deviations and identify problems in the processes. In the “check” phase, additional tools such as the PROXIA Measures Manager are available. With it, it is possible to examine implemented measures for specific goals or changes in condition. For example, the change in the downtime pattern of equipment can be studied after the material supply process has been adjusted. This means that the question of whether the changeover to a supermarket concept with pre-picking has a positive effect on the productivity of downstream processes can be answered with a secure data basis. At the same time, the impact on the overall order throughput time can be evaluated. By reviewing various metrics, in the context of a measure, this allows potential conflicting goals to be identified. A performed measure remains as a digital image in the PROXIA system. In this way, it is possible to repeatedly check the quality of measures once they have been implemented.
Location of Process Mining – ERP or MES?
While the ERP system is used to monitor and control the type and quantity of orders for a period of time, PROXIA MES is used to determine processing times, allocation of resources and the order processing sequence in production. In this system symbiosis, the ERP system is usually referred to as the “leading system”. This statement is also true for the management of master data and the holistic control of the value creation process. However, many ERP systems do not sufficiently immerse themselves in the “microcosm” of production. At this point, PROXIA’s MES takes over the operational process and ensures agile and optimized control of the value stream. In addition, the MES supports the ERP systems by a permanent information feedback in the control of the superordinate value creation process. This results in a clear separation of responsibilities and tasks. PROXIA’s MES functional scope includes production planning, sequence planning, order control, machine data/operating data acquisition, maintenance control and quality management – as well as the provision of data mining functionalities. Generally speaking, data mining from mass (non-process) data promotes correlations by identifying new cross-connections and trends in a data-based manner. This is where the reporting module of PROXIA Software AG provides support. Depending on how long the system has been digitally tracking the production processes, representative data stretches can be used for consideration. For example, it is possible to determine points in time or events at which parameters have changed. By holistically collecting and recording data from production, it is now possible to determine the causes around these events.
Many current ERP installations are only partially suitable for the detailed management of dynamic production processes. This is due to the data model alone, because ERP systems often aim to manage costs and materials and are used to allocate personnel, material and overhead costs. Modeling complex dynamic processes and visualizing them in a user-friendly way is not one of their tasks. In contrast, the PROXIA Timeline analysis tool has a completely different focus: here, the entire production process is displayed graphically. In this way, a deviation can be recognized at first glance without having to study comprehensive figures. Even more complex questions such as, “Did the transfer of subsets within production work smoothly?” “How is the image of my ghost layer?” are answered graphically. This relieves the administrative workload in production because less time has to be spent on production controlling.