Refine
Document Type
- Conference Proceeding (3)
- Article (2)
Language
- English (5)
Has Fulltext
- yes (5)
Is part of the Bibliography
- no (5)
Keywords
- Gießerei (4)
- foundry (3)
- optimal scheduling (3)
- Ablaufplanung (2)
- Ganzzahlige lineare Optimierung (2)
- batch-wise parallel process (2)
- dwell-time (2)
- linear integer programming (2)
- neural network model (2)
- prediction methods (2)
Institute
This paper presents a databased approach for improving the precision of the moulding sand compressibility in the moulding sand mixer of a foundry. In this approach, the deviation between the measured and the target compressibility is reduced by controlling the water addition. The complex dynamic behaviour of the process variables and their influence on the water addition is modelled with a long short-term memory (LSTM) network. Another LSTM network as control path simulates the impact of the water addition on the compressibility. Simulation and experimental results with the applied model for water prediction in a feedforward control yield relevant improvements of the moulding sand compressibility.
This paper proposes an extended Petri net formalism as a suitable language for composing optimal scheduling problems of industrial production processes with real and binary decision variables. The proposed approach is modular and scalable, as the overall process dynamics and constraints can be collected by parsing of all atomic elements of the net graph. To conclude, we demonstrate the use of this framework for modeling the moulding sand preparation process of a real foundry plant.
We present a feedback-corrected optimal scheduling approach to reduce the demand of electrical energy of batch processes, exemplified at the sand preparation in foundry. The main energy driver in the exemplary foundry is the idle time of the batch-wise working sand mixers. In this novel approach, we use linear integer programming to minimize the demand of energy of the sand mixers by scheduling the batches in real-time. For the optimization we use a physical model of the sand preparation, which takes dwell-times of the processes as dead-time systems into account. In this paper, we present the steps to make the optimal scheduling approach applicable for the production process. The application at the real production plant proves the performance of the suggested approach. Compared to the conventional control, the feedback-corrected optimal scheduling approach leads to an reduction in energy consumption of approximately 6.5 % without modifying the process or the aggregates.
This paper presents a novel approach for modelling the energy consumption of the coupled parallel moulding sand mixers of a foundry as an optimal control problem. The minimization of energy consumption is optimized by scheduling the mixing processes in a linear integer programming scheme. The sand flow through the foundry’s sand preparation is characterized by a physical model. This model considers the sand demand of the moulding machine as disturbance, the stored sand masses in the mixer hoppers and machine hoppers, respectively. The novel approach of handling dwell-times for dosing, mixing and transport processes using dead-time systems and constraint pushing allows the application of a linear model. The formulation of the optimal control problem aims at real-time application as model predictive control at the production plant. Initial application results indicate an improvement in energy consumption of approximately 8%.
We present a novel long short-term memory (LSTM) approach for time-series prediction of the sand demand which arises from preparing the sand moulds for the iron casting process of a foundry. With our approach, we contribute to qualify LSTM and its combination with feedback-corrected optimal scheduling for industrial processes.
The sand is produced in an energy intensive mixing process which is controlled by optimal scheduling. The optimal scheduling is solved for a fixed prediction horizon. One major influencing factor is the sand demand, which is highly disturbed, for example due to production interruptions. The causes of production interruptions are in general physically unknown. We assume that information about the future behavior of the sand demand is included in current and past process data. Therefore, we choose LSTM networks for predicting the time-series of the sand demand.
The sand demand prediction is performed by our multi model approach. This approach outperforms the currently used naive estimation, even when predicting far into the future. Our LSTM based prediction approach can forecast the sand demand with a conformity up to 38 % and a mean value accuracy of approximately 99%. Simulating the optimal scheduling with sand demand prediction leads to an improvement in energy savings of approximately 1.1% compared to the naive estimation. The application of our novel approach at the real production plant of a foundry proves the simulation results and verifies the capability of our approach.