Abstract
In the capital-intensive technological industries, equipments are important assets and high value-added sources. After long term using, equipments may inevitably break down, appropriate maintenance is therefore needed for early prevention. Equipment maintenance can usually be divided into two types, i.e., breakdown maintenance and preventive maintenance. Breakdown maintenance is executed when the equipment failure has already happened unpredictably. Preventive maintenance, on the other hand, is executed before the occurrence of failure in order to reduce the frequency of equipments’ breakdown. Even though preventive maintenance is a better approach to ensure good equipment status, it interrupts the regular production schedule and, nevertheless to say, it costs money. For these reasons, the frequency of preventive maintenance frequency is usually based on the cost consideration. There are many expenditure items when executing equipment maintenance. The existing cost-estimation methods sum up these items into one and then set up a maximum acceptable value. In this research, however, we treat different expenditure of items as different-dimensional data, which are not comparable and therefore cannot be summed up into one unique value. In light of this, we use Data Envelopment Analysis (DEA) method to find out the rationalized expenditure item by item based on historical data. Taipower’s 161kv GIS equipment’s casual maintenance was selected as a case study. The results revealed that three out of the eight collected data sets did not exhibit reasonable maintenance expenditure. By applying the DEA method, the unreasonable expenditure items embedded in these three data sets can be located along with their suggested values. The method proposed in this research can further be applied in the future to determine whether the quotation provided by the maintenance supplier is acceptable or not.