Map – Machine Learning for Agile Process Management in Mechanical and Equipment Engineering

The following example describes the budatec use case.

budatec GmbH receives an order that contains customer-specific adjustments to a standard machine. An interactive shop floor management system is required so that the employees who process the order are provided with all the necessary information. This enables the communication of adjustments to the standard product. Asian customers in particular also expect to be able to make changes or adjustments to the order even after the order has been placed. This must be communicated immediately to the employees on the shop floor so that they can assemble the correct components. After the employees on the shop floor have all the information for the order displayed on their mobile devices, they begin assembling the machine. During this process, a part is noted as missing. To document this, the comment field in the mobile application is used to forward the information to project management, who also use the interactive situation report. News about raw material shortages that affect said part are displayed in the environmental analysis that is integrated into the interactive situation report. Project management can use this information to initiate appropriate measures for future orders. The following video shows the challenges and how the environment analysis and the assistance system help shop floor employees overcome them.

In order to be able to implement the application described here, various preliminary work must be carried out. The aim of the project was to describe the manufacturing process of a product in detail. In an iterative process, both a process model and detailed documentation of the manufacturing process were created. Furthermore, it was evaluated how this information could be automatically transferred to an ERP system. The result is a standardized procedure that enables budatec GmbH to automatically transfer process descriptions to an ERP system. This approach significantly reduces the previously manual effort required to fill the ERP system.

Environmental Analysis

Established environment analysis tools are not linked to process management. Accordingly, the aim of the MAP project was to develop such a link. The procedure for creating an AI-supported environmental analysis is described below. First, news sites are specified for searching. The news articles are then scraped and manually labeled. Based on this, the AI ​​algorithm is trained. Once the algorithm is trained, it outputs news articles relevant to the company, which are communicated graphically in the interactive situation report. The AI-supported environmental analysis creates transparency in the company environment for senior employees and managers.

The link below will take you to our Process Assistant. There you will find all the necessary information about the requirements, effort, etc. for introducing AI-based environment analysis.

Assistance system for employees on the store floor

SMEs in particular pursue a service-oriented approach. The components that are common in large companies, such as the Manufacturing Execution System (MES), the Enterprise Resource Planning (ERP) system and the Product Lifecycle Management (PLM) system, are not present in most SMEs. Outsourcing IT infrastructure enables SMEs to react more flexibly to changing market situations or customer requests. In order to succeed, a concept is required that automatically describes customer-specific manufacturing processes. Shop floor IT already enables various internal company departments to be networked, but setting up the model requires a lot of time. To counteract this, the MAP project developed an assistance system through which the processes are partially automatically generated. Another advantage is that employees on the shop floor are supported and can report deviations – e.g. missing parts – directly to the system.

The link below will take you to our Process Assistant. There you will find all the necessary information about the requirements, effort, etc. for the introduction of the worker assistance system and the interactive situation report.

The following example describes the KSB use case.

Due to the numerous variants that KSB offers, it is a challenge for product managers to keep track of them. This also makes it more difficult to identify the most profitable product variants currently and in the future. Two assistance tools provided to product managers are intended for support. The Soley Impact Manager displays links to product and process data. Based on this information, product managers can identify which product variants are particularly successful. In addition to analyzing this historic data, a look into the future is also necessary. To do this, product managers use the strategic foresight method. By identifying new policies at an early stage, for example, product managers can incorporate this information into their product portfolio. Thanks to the Soley Impact Manager and strategic foresight, product managers have a well-founded decision-making basis on which the product portfolio can be built. The following video repeats the challenges and shows how the Soley Impact Manager and strategic foresight contribute to transparency of the product portfolio.

Strategic Foresight

New regulations or legal changes have particular impact on KSB's product portfolio. In order to identify upcoming changes at an early stage, a commercial tool for environmental analysis is used. The results depend heavily on the search query, which requires terms and their synonyms to be assessed for inclusion or exclusion from the search. Previously, this was done manually, requiring a lot of effort. In order to reduce this complexity, the MAP project developed an interface that helps build the search string. To do this, the user selects a product family and receives direct suggestions as to which terms can be included in the search. Synonyms are stored behind these terms so that they are directly included in the search. The developed interface quickly creates a search string that is inserted into the environment analysis tool.

Product Portfolio Optimization with Product Mining

The product portfolio is the central aspect for industrial manufacturing companies in their pursuit of growth, profitability, innovation, sustainability and resilience. Soley's product mining platform makes it possible to make consistently faster, better and fact-based decisions to optimize the product portfolio, even with high levels of complexity. For this purpose, a graph-based Enterprise Digital Twin is created, which is based on technical, commercial and logistical data. This digital twin is enriched with engineering knowledge in the form of complexity analyzes and value patterns, and provides an innovative, transparent insight into the connections within the product portfolio. Delivered in an enterprise-grade, cloud-based solution that includes features such as digital decision support and task management, Soley helps the world's leading companies actively drive business transformation. In the MAP project, the approach was expanded to include networking with data from process mining. This makes it possible to incorporate process-related aspects, such as lead times or process deviations, into the evaluation and optimization of individual elements of the product portfolio.

The MAP project was supported by Sino-German junior research teams. These teams each worked together for six months, and took on additional work for the MAP project. The Process Assistant (PA) (Link in German only) is based on a process model that was created using the MO²GO modeling software, and maps the organization and support of such young research teams. Furthermore, the barriers identified in the collaboration are shown and described in the PA. To reduce and eliminate these barriers, best practice measures were identified and implemented within the supervision of the Sino-German research teams. These measures each relate to specific processes and barriers, which are also shown in the PA.