A data-driven method to predict future bottlenecks in a remanufacturing system with multi-variant uncertainties

The remanufacturing system is remolding the manufacturing industry by bringing scrapped products back to such a condition that reintegrated performance is just as good as new. The remanufacturing environment is featured by a far deeper level of uncertainty than new manufacturing, such as probabilistic routing files, and highly variable processing time. The stochastic disturbances result in the production bottlenecks, which constrain the productivity of the job shop. The uncertainties in the remanufacturing process cause the bottlenecks to shift when the workshop is processing. Considering this outstanding problem, many researchers try to optimize the production process to mitigate dynamic bottlenecks toward a balanced state. This paper proposes a data-driven method to predict bottlenecks in the remanufacturing system with multi-variant uncertainties. Firstly, discrete event simulation technology is applied to establish a simulation model of the remanufacturing production line and calculate the bottleneck index to identify bottlenecks. Secondly, a data-driven method, auto-regressive moving average (ARMA) model is employed to predict the bottlenecks in the system based on real-time data captured by the Arena software. Finally, the proposed prediction method is verified on real data from the automobile engine remanufacturing production line.