Improving the Performance of Waste Handling Cranes by Optimisation-based Decision Rules

Abstract

Waste-to-energy is the process of generating energy from municipal waste. The most conventional energy recovery method is incineration, which takes place at a waste-to-energy plant. The incoming waste is collected in a bunker, where an overhead crane stacks and mixes the material before feeding it to the furnaces. The more homogeneous the waste reaches the furnaces, the more stable the incineration process. In automated mode, crane movements are controlled by the generation and execution of crane orders.

The waste crane scheduling problem is formulated as a mixed integer linear program, describing the scheduling of crane orders based on order characteristics such as type, origin and destination. The objective of the problem is to optimise the crane performance, which is interpreted as the minimisation of the total crane driving time while performing as many mixing orders as possible.

In the base model, simulated order schedules are resequenced. Given the type, origin and destination of the orders, the starting times are reconsidered. The obtained reductions in total driving time compared to the simulation model vary from 2% to 7%, depending on the crane strategy, whether stacking or mixing is allowed, and the maximum time between the generation and completion of feeding orders. Since the order types are given as input, the number of mixing orders is fixed.
In the extended model, the determination of the order characteristics is included in the scheduling process. Unfortunately, the size of the model increases so fast over time that 15 minutes proved to be the maximum length of a scheduling period for which a solution could be obtained. With a rolling horizon approximation, a stacking period of 10 hours has been scheduled.

Decision rules on the generation of orders are derived from the output of the optimisation models. These decision rules are implemented in the simulation model. The obtained results are assessed based on key performance indicators mainly related to the mix quality. The rules resulted in up to 35% more mixing orders, which translates into material being mixed more often, not necessarily in more material being mixed.