HDPS-BPSO based predictive maintenance scheduling for backlash error compensation in a machining center

This paper presents a novel HDPS-BPSO maintenance scheduling strategy for backlash error compensation in a machining center through binary particle swarm optimization (BPSO) and data-driven regression methods. During the experiment, a hierarchical diagnosis and prognosis system (HDPS) was leveraged to predict the potential backlash error first. Then BPSO is applied to provide optimized maintenance strategies through capturing the trade-off between several factors such as maintenance cost, machining accuracy, and defective percentage. The target of proposed predictive maintenance strategy is to minimize the cost of potential failures and relevant maintenance performances. The numerical result in this case demonstrates the benefit of implementing predictive maintenance compared with preventive one.