This paper describes the methodology developed to determine the size of a new fleet of combat helicopters for the Royal Australian Navy. The Australian Government's 2009 Defence White Paper described "as a matter of urgency" a requirement for a fleet of helicopters "to provide eight or more aircraft concurrently embarked on ships at sea". The objective therefore is to find the minimum fleet size that enables the fleet to meet both this minimum daily embarked requirement, as well as annual requirements for a specified number of flying hours for both embarked helicopters and the remaining ashore-based fleet. The fleet sizing problem incorporates the helicopter fleet, ships, personnel and a home base with maintenance facilities. Individual helicopters (or tails) pass through various states throughout their life. These include the serviceable state when they are able to fly, or a range of maintenance states which will make them unable to fly. Maintenance types include regular inspections, phased maintenance and deep maintenance. Many of these states can occur when a tail is either embarked or ashore. Unscheduled maintenance adds a random element to the problem and influences which aircraft may be serviceable on a day-to-day basis. A discrete-event simulation approach was chosen to address this problem, due to the requirements to track tails in various states, test for state transitions and incorporate random effects. The fleet sizing model begins using the nominated fleet size as an input, along with the required number of embarked aircraft and the annual embarked and ashore flying hours. Other inputs include maintenance data regarding the time between services and their duration. The fleet is initialized with the required number of tails embarked and the remaining ashore tails in various serviceability and maintenance states. The model runs for the expected fleet life of 30 years, stepping through one day at a time. At the beginning of each day, it is determined whether it is a work day for ashore flying crews or maintenance crews. Embarked crews work every day of the year. Embarked serviceable aircraft will fly a certain number of hours depending on their operational tempo. A procedure is used to balance the flying hours of the embarked fleet while accounting for their expected and realised maintenance, in order to meet the annual requirement. If it is an ashore work day, the next step is to allocate flying hours to aircraft, and maintenance crews to maintenance lines. As a baseline, a flying allocation method is used that seeks to minimise queuing of aircraft for phased maintenance. This is compared with a greedy technique that seeks to meet the daily requirement first, but at the expense of lower serviceability. Similar methods are used to allocate maintenance resources. The baseline in this case is allocating the same number of personnel to each maintenance line for both flight line maintenance (which handles regular inspections and unscheduled maintenance) and phased maintenance. An alternative is a greedy technique applied to flight line maintenance, where personnel are pooled and the aircraft with the least number of maintenance manhours is serviced first. This will increase the serviceability of the fleet, but may leave aircraft requiring a large amount of maintenance unserviceable for a long time. Once the allocation of resources is completed, aircraft will fly or be maintained accordingly. At the end of the day, tests for state transitions occur. Aircraft will move in and out of scheduled or unscheduled maintenance, or in and out of a serviceable state, subject to criteria being met. Disembarking aircraft need to be replaced immediately. The model can also include tails going on exercises or being lost from the fleet. To demonstrate the capability of the model in sensitivity and trade-off analysis, results are shown for a generic helicopter against an indicative requirement. The greedy maintenance allocation heuristic is shown to be superior, even superseding the differences between the flying allocation methods. When a modification program is included, a large decrease in ashore hours occurs due to queuing for deep maintenance during the mid-life upgrade, which can be overcome by increasing maintenance capacity. These results clearly show the influence that resource allocation and maintenance capacity can have when predicting fleet size.
