Production Control

In the project CrimpProd-S the integration of decentralized, self-learning production control systems in the context of Industry 4.0 in combination with additive manufacturing is investigated. The overall target is the economic evolutionary integration of lot size 1 in existing production facilities. Therefore a procedure model and customized methods are developed to analyze the potentials regarding the existing company’s strategy, value chain as well as the business model. This integrated procedure guarantees the efficient and customer-oriented transformation.

There is a variety of empirically validated methods for implementing pull-systems in the manufacturing industry, but pull-based replenishment for the process industry remains a research gap. This article describes the development of a model for implementing a pull-system for an intracompany production network in the process industry and its validation in a case company.

The network analysis is a promising approach for assessing the behavior of material flow systems based on large data volumes. Previous studies focus mainly on static network analysis. This means that all the events that occur in a specific time period are aggregated to a single material flow network. As material flow systems are rapidly changing systems, a static view is not sufficient. The aim of this paper is to present existing concepts to obtain such structural changes and to assess their suitability for material flow networks by using real data.

The methods presented allow the splitting of classic monolithic numerical controls of industrial robots and machine tools into their functional units. The core functionalities can then be brought onto different computers in even separate places. Using techniques of augmented reality allows enriching a captured scene with additional information, as a virtual model of the industrial robot or the planned paths. Combining these approaches leads to a simplified programming task for industrial robots as the programs can be visualized in their context. This decreases setup time and improves quality.

Current control systems are limited from a technical viewpoint in areas such as scalability, reconfiguration time and computational complexity for algorithms. These limitations call for a new concept for control systems to address current and future requirements. One idea is that the physical location of the control system has to be moved of the machine to a cloud. In this way, the control system becomes scalable and can handle highly complex computational tasks while keeping the process expertise in the bay.

The volatile world urges companies to react more agile to fast changes. Monitoring is a major building block for corporate agility. A lack of understanding of causal interrelations between a multitude of volatility drivers demands a new monitoring approach. Big data is presented as an adequate tool and its agility support is discussed from a data and analytics point of view. Conclusively, guidelines for developing a monitoring system are discussed.

Volatility in market demand leads to temporary over- and under utilization of productive assets. Heijunka aims at de-coupling the production system from market volatility. The production program is spread as even as possible over time. This achieves high asset utilization, short lead times, and low inventories. There are validated Heijunka methods for the manufacturing industry, but for the process industry this remains a research gap. This article describes the development of a Heijunka model for the process industry in order to close that gap.

The integration of autonomous control methods into central PPC-systems enables a high logistics target achievement despite occurring dynamic influences. Thereby, the integration requires coupling strategies to specify the harmonisation of autonomous control and central planning methods. The logistics target achievement and the adherence to given planning parameters depend on the specific strategy and the related level of autonomy. This paper introduces a target system which enables the evaluation of coupling strategies considering both logistics target achievement and the adherence to given planning para-meters. Thereby, producing companies are enabled to consider both aspects for the selection of suitable planning and control methods.

One of the main purposes of the „Smart Virtual Worker“-project is the application of a digital human model for the simulation of industrial work tasks. This contribution focuses on the implementation of an autonomic action selection, such that a virtual agent is able to solve tasks under certain optimization criteria. To realize the autonomic action selection, we use the Q-Learning algorithm with different extensions. In this article, we describe these different learning algorithms and we briefly describe the performance of their implementation with regard to the industrial field.

Due to a variety of technological developments within the past decade, today’s industrial working environment is changing dramatically. Information and communication technologies as well as automation and robotics are penetrating the working environment. In the context of industrial 4.0, the interaction of systems increases. Sensor data, status information and control commands are exchanged, which creates the conditions for new types of production concepts such as Remote Manufacturing. Remote Manufacturing describes a delocalized production, which removes the spatial unity of production factors. It will be no longer necessary for the elementary production factors (labor, materials, equipment) and the dispositional production factors (planning, management, control, organization) to be located at one place. In the future, it will be possible to operate production facilities abroad and to monitor the process control level from Germany. In this paper, the economic environment and the ...