AGV Factory Layout Simulation

The subject of the study was an investigation into the use of Automatic Guided Vehicles (AGVs) in a factory which cuts and prepares high quality, thin card in a continuous-process, batch-produced environment. The factory used conveyor systems but wanted to improve throughput flexibility by changing over to AGVs. The AGVs would perform several tasks, including transporting large rolls of card to the 'slitting' machines as well as pallets of 'cut-to-size' card to the appropriate destination. It had been recommended that 7 AGVs could satisfy the requirements and the discrete event investigation was to confirm this as part of a risk management exercise.

The investigation commenced by modelling in detail the existing factory structure, building walls, pillars and machinery. The proposed AGV track system was added, as was the vehicle data to give them realistic performance; such details as vehicle dimensions, turning circle, speed, acceleration, power consumption and battery charging requirements. A great deal of care was taken to ensure an accurate representation of the environment was created. The discrete event simulation was then given its parameters and run.

The results showed up several potential problems. A main, two-track aisle was not wide enough to allow one AGV to pass another vehicle manoeuvring at a dispatch point and the aisle width was restricted because of several roof pillars. The solution was to have one track each side of the pillars. This had a knock-on effect of intruding into a goods 'buffer' storage area. This area was necessary and, because of space restrictions, an office wall had to be demolished to meet the floor space requirement.

Another problem, relating to the need to recharge the AGVs, was that once the battery charge dropped to a certain reserve level, a vehicle would automatically return to the charge station. However, because of the random nature of the work pattern, they were often all charging at the same time. Once the recharge sequence had started it could not be stopped, resulting in system down time. This was overcome by modifying the response of the AGVs to ensure that a number of partially charged vehicles were always on stand-by. The analysis also revealed that at higher levels of throughput, bottlenecks could occur at delivery points situated on the end of track spurs. The smooth interaction of the whole system was improved by using additional track loops to replace the spurs.

Finally, it had been recommended that seven AGVs were used for the system which provided the capability to overcome quickly bottleneck situations. By studying the simulation it was shown that over longer periods of time, six vehicles offered a better overall efficiency and level of investment The exercise proved its worth by highlighting cases where a simple sequence of events and situations combined to create previously unforeseen AGV traffic flow problems. Discrete event analysis provided the opportunity to trial immediately alternative solutions. The analysis also had an impact upon the business case indicating that a significant cost saving could be achieved by purchasing fewer AGVs. View the movie.