LAUSR

Abstract The paper proposes an energy consumption cross-level model for a demand driven machine tool working in a manufacturing system. The aim is to optimize performance of manufacturing system engaging more than one organizational level, in this case the machine and the system levels. The paper exploits a former study on Pareto optimal Minimum Energy-Time functions, representing the best possible machine setups under varying cycle-time. The functions are used to model machine productive states, whereas stochastic behaviours (failures, blocking, starvation) and inter-machine interactions are described using a Markovian General Threshold Model. Here, instead of using a constant power per machine state assumption, a functional relationship between processing rate and machine energy demand is used. A stand-by energy-saving policy, which exploits the flexibility in adopting a variable processing rate and a settable threshold level of the interoperational parts buffer as a trigger, is developed and benchmarked. The model is used to analyse an industrial case study of a three-station transfer machine. It takes into account a number of cost contributing factors, apart from sole energy demand, such as tooling cost, operator cost, inventory and costs related to potentially undelivered throughput. The application of the proposed cross-level optimization scheme to a case study showed an improvement both in energy-efficiency and in profitability of the production system under investigation. Moreover, it has been demonstrated that a model exploiting knowledge from both levels can avoid suboptimal setups resulting from treating single-level problems independently.