Digitalization also offers great potential in electroplating technology. In particular, the use of accumulated data to improve electrolyte management, predictive system maintenance or quality improvement represents an exciting field of application. The article provides an overview of the opportunities and obstacles to the introduction of Industry 4.0, explains frequently used terms and introduces the topic of data acquisition, data analysis and simulation. Finally, the first SME-compatible solutions for the digital transformation in electroplating technology are presented using practical examples from projects at the Fraunhofer IPA, the IFF at the University of Stuttgart and the IWF at the TU Braunschweig.
1 Introduction
The digitalization of industry is now internationally referred to as Industry 4.0. Its origins can be traced back to a future-oriented project launched in 2011 as part of the German government's high-tech strategy [1]. A central component of this is cyber-physical systems that link physical production processes with data from the cyber environment. Industry 4.0 is therefore not only having a decisive impact on production processes, but many business models are also being increasingly questioned or further developed in the context of digitalization [16].
The production of electroplated surfaces requires the interaction of a large number of influencing variables, ranging from the system technology and peripherals (e.g. pumps, exhaust air, media supply) to the interacting chemical (inorganic and organic components, degradation products, pH value) and physical (temperature, current density, current form) coating parameters. The number of players involved in the creation of the coatings is similarly large. The coating system consists of tanks, possibly trolleys, rectifiers, peripherals such as pumps and ventilation technology. Depending on the component manufacturer, there may be different interfaces. In addition, there is the electrolyte, which is usually purchased from a process supplier. The coating company is therefore faced with the major task of networking the large number of individual systems and components with each other and with its own systems and devices in order to exploit the above-mentioned potential.
Obstacles on the way to implementation can also be derived from the considerations of plasma technology. These are the factors of time, high complexity and the necessary investment costs with a benefit that is usually unclear in absolute terms [6]. This cost-benefit conflict can be countered by developing general implementation concepts, preparing the content of the subject area and providing specific project examples.
2 Cyber-physical production systems for electroplating technology
Fig. 1: Cyber-physical production system as a core element of Industry 4.0 (based on [20])Cyber-physical production systems (CPPS) are a core element of Industry 4.0 and enable the coupling of physical elements with virtual elements of a production system. Figure 1 outlines the structure of such a system using the example of electroplating technology. Specifically, systems and machines can be coupled with simulation or data-based model environments, for example. Data acquisition systems ensure the coupling from the physical to the virtual environment and control signals and/or decision support for the physical environment can be derived from the virtual environment [20].
CPPS have great potential for electroplating technology in particular, as they allow the high complexity (combination of discrete and process-oriented production) of the processes to be better controlled. Mapping the production system in a model environment enables a much better understanding of the system and new approaches to improving productivity can be achieved. In particular, these approaches can support the improvement of energy and resource efficiency, robustness and the adaptability of processes [18].
2.1 Physical system: electroplating plant
Manual production is common for large parts, special applications and low-volume parts. The most commonly used type of system for larger quantities is currently certainly the automatic electroplating machine in conjunction with racks or drums. The racks/drums are attached to product carriers and are handled automatically within the system. The loading and unloading of the product carriers or racks/drums is carried out manually or via automation, whereby the proportion of automated loading is tending to increase, but is still far from being used as standard across the board. A system is a collection of containers for the electrolytes, which are connected via a transport system. It has all the equipment directly required for operation (e.g. sensors on containers, heating/cooling, pumps/filters, rectifiers, etc.), which are switched by a higher-level control system.
2.2 Data acquisition
Fig. 2: Current communication structure in the factory [14]In production companies, data is generated at various points, which can serve as the basis for CPPS. The automation pyramid in Figure 2 provides an overview of the various systems for recording data in production companies, starting with the specific recording of system statuses using sensors and extending to company-wide data recording in the ERP. The data volumes and resolutions recorded differ depending on the level.2.3 Decision support and control systems
2.4 Cyber system: smart data and big data
The virtual cyber system, often referred to as a digital twin, is a digital representation of the physical system. For this purpose, model-based and data-based approaches are essentially pursued, which are presented in more detail below.
2.4.1 Smart data: model-based approaches
Furthermore, simulations can be divided into discrete and continuous systems. In discrete systems, the variables only change at certain points in time, while continuous simulations allow the mapping of temporal changes [2]. An example of a discrete manufacturing system is a production line in which incoming parts at a machine change the state of the model (part hung in product carrier). Continuous systems can be process systems in which, for example, a chemical is consumed. Rates of change in such a system can be described by differential equations or difference equations.
2.4.2 Big data: data-based approaches
Data-based approaches are now a widely used process for evaluating large data sets. Initially, this was primarily used for the evaluation of internet searches, social networks and financial markets. Current developments in Industry 4.0 mean that these approaches are increasingly being used in industrial production.
The CRoss-Industry Standard Process for Data Mining (CRISP-DM) is an established process for data mining [17]. The individual phases can be seen in Figure 3. A fundamental component of this process is to depict the data in a model and use it to identify links and correlations in the data. For this purpose, statistical methods are systematically applied to large data sets [5].
Advanced machine learning makes use of data mining methods, among others. This in turn is divided into the areas of supervised learning and unsupervised learning. The aim of supervised learning is to develop a model that can provide predictions for input data for which no data set exists. Well-known examples of supervised learning are artificial neural networks (ANN) and decision trees. For the area of unsupervised learning, kMeans is a well-known algorithm. The aim of unsupervised learning is to find patterns and links in a given data set. kMeans, for example, attempts to assign all data sets to a group (cluster). The number of groups to be searched for must be specified in advance [10].
Algorithms for machine learning are now available in proprietary and non-proprietary software and programming languages and are well integrated. Software such as KNIME offers a good overview of the evaluation with its GUI. For continuous evaluation and real-time applications, however, programming languages such as C/C++, Python, R or Java are preferable.
The models of the cyber system should be used to positively influence the physical system. To this end, control signals can be derived directly from the cyber system or a visualization can be used to influence employees in their decisions.
3 Application examples - How Industry 4.0 can work in electroplating technology
Fig. 3: Phases of the CRISP-DM reference model [17]In the following, an already implemented and a planned project in the field of I4.0 will be presented.
Electroplating 4.0 - Benefits of a digital twin of an electroplating plant
The challenge here was to represent the model of the system and the electrolyte as accurately as possible so that deviations between the real system and the model are as small as possible. The determining factor for the consumption parameters in a drum coating system is primarily the carry-over through the drum and components. In cooperation with the contract coating company, the components to be coated were divided into carry-over categories after prior empirical measurement, which were then stored in the ERP system.
A decisive factor for the successful implementation and use of the simulation model was the subsequent integration of real order data from the MES into the digital image in order to be able to track the change in the electrolyte due to the real coating processes.
Compared to the usual consumption control via ampere hours, which assumes an average value of the carryover in relation to the coated area, this results in significant deviations between the two operating modes, especially in the case of accumulations of individual component types. During the validation of the algorithms, the simulation model was able to achieve a mathematical accuracy of the concentration of an electrolyte component over the course of a month with an average deviation of 2.8 % from the chemically analyzed concentration.
The digital image of the coating system and the electrolyte with its specific consumption parameters, in combination with the real application data, enables the calculated determination of electrolyte concentrations. This is of particular interest with regard to electrolyte components that are difficult to determine or can only be determined irregularly. A digital twin of the electrolyte can be used in future to improve process stability, achieve tighter tolerances and thus save costs for chemicals. Going one step further, the development of a digital twin of the electrolyte can be used to develop a system for predictive electrolyte maintenance. This approach is to be developed as part of a planned research project.
Fig. 4: Course of the simulated and analytically determined concentration [13]
Predictive maintenance in electroplating technology
Maintenance includes all technical and administrative measures to maintain the functional condition of a unit under consideration and to restore it in the event of a failure [7]. In operational practice, maintenance plays a very important role due to high costs and downtimes. With regard to economic factors, inadequate maintenance measures can lead to losses in production capacity of up to 20 % [9]. In a survey of experts, around 80 % of respondents therefore confirmed that maintenance is of high and increasing importance [12].
As part of a planned research project, an interdisciplinary team from industry and science is tackling this challenge in order to simplify the maintenance process of the system and electrolyte and thus increase system and electrolyte availability. With the help of simulation and big data approaches, the project aims to enable the plant operator or a service partner to take predictive maintenance measures in line with requirements. From a process engineering perspective, the concept of electrolyte management in particular is to be further developed with the help of simulation and data evaluation. An important part of this is also the investigation of which sensors and data acquisition are actually necessary, as a large amount of data is often collected but neither centrally recorded nor analyzed and evaluated. The development of suitable models and algorithms should enable predictive maintenance of electrolyte and system technology, so that in future a mode of operation with narrower tolerance limits can be established with a higher capacity and, above all, with the shortest possible downtimes.
4 Summary
The aim of Industry 4.0 is to improve the productivity, flexibility and quality of production processes and to create opportunities for new business models through the intelligent networking of systems and processes. Electroplating technology is faced with the difficult task of meeting the challenges of digital transformation in the face of limited personnel capacity, unknown investment costs and the question of 'where to start'.
The IWF at TU Braunschweig and Fraunhofer IPA are working with various SMEs from the electroplating industry to develop useful and practicable solutions for electroplating that can be used to take the first steps towards implementing Industry 4.0. The current focus is on the beneficial use of data that is already being generated.
