In the Volvo Aero focused factory where Turbine Exhaust Casings (TECs) are made, an initial programme of experiments was conducted, addressing fifteen different scenarios. These reflect the combination of three forms of logistic control, and five different events during which their performance would be evaluated.
The logistic control methods included a representation of the present-day MRP approach (a ‘push’ scheduling strategy), a Kanban approach (using orders to ‘pull’ products through the factory), and an Optimised Production Technology approach (combining elements of ‘pull’ and ‘push’ within the same workflow, centred around a bottleneck process). Each of these was subjected to five different patterns of events:
  1. A ‘steady state’ scenario, where supply and demand are consistent throughout the model run, and there are no problems in the production system
  2. A ‘demand changes’ scenario, in which there is a temporary surge in demand that the manufacturing system must respond to
  3. A ‘quality problems’ scenario, in which some parts fail a key inspection, and replacements must be introduced to the workflow
  4. A ‘material shortages’ scenario, in which a key component for one of the product types is delayed
  5. A ‘machine breakdown’ scenario, in which on of the production resources is rendered unavailable for a time
 
Supply chain simulation Process chain evaluation Internal logistics simulation Supply chain
IT infrastructure Glossary Tools
 
Sample results
Tools Back Next Home Logout Sample results Areas of application Aims Innovation Introduction to logistic control Performance metrics Further reading Case studies Model construction Introduction to internal logistics Conclusions and recommendations Validation
In each case, the performance metrics described in the previous page were used, producing results showing the accumulation of work-in-progress within the system, the tied-up capital, fill rate and so on. The performance of the three systems of logistic control can thus be compared. For example, the figures below show the makespan (time in the manufacturing system), for a business unit operating under MRP, Kanban and OPT. The scenario in each case is a shortage of components required for one TEC type:
MRP - driven logistics Pull / Kanban - driven logistics OPT - driven logistics
Under MRP, the disruption caused by the component shortage is never entirely recovered from. This is mainly caused by a buildup of work-in-progress of other types, while the key component is unavailable.
Under a ‘pull’ system, the disruption caused by the component shortage is not as bad (in terms of the fill rate) and the level of work-in-progress is kept under tight control since the kanbans to do allow additional material into the system. It recovers well.
Under an OPT approach, component lead times are kept reasonably constant, and the level of work-in-progress is held steady. There is still a backlog of unfilled orders at the end of the model run period, however. It may be that further ‘tuning’ of the system is necessary, building up buffer stocks of components to allow the system to cope with this kind of disruption.
The results of all fifteen experiments, showing performance against every metric, can be seen in the project technical report, ‘Alternative logistic concepts and their effect within the company and the supply chain’ - see the bibliography page for a link to this document.