Article Preview
Top1. Introduction
The best running enterprises are running enterprises. The manufacturing of goods and the provisioning of services without interruptions is important to create revenue and increase (Stenström et al., 2016). Therefore, enterprises use maintenance to minimize or avoid downtime (Stenström et al., 2016).
The impact of maintenance strategies on firm performance has already been analyzed by Swanson (Swanson, 2001). Two basic strategies can be differentiated. Reactive and proactive strategies (Bateman, 1995; Stenström et al., 2016). Reactive or also run-to-failure approaches try to optimize the reaction to failures of plants according to Mobley (2002). This reactive (or corrective) maintenance is composed of post-hoc activities, executed after a warning or error occurred (Tsang et al., 2006; Nadj et al., 2016). A reactive approach to maintenance implies that an enterprise must keep personnel available and maintain a spare parts inventory for reacting to failures. Even if the parts inventory can be reduced by cooperating with equipment vendors that provide immediate delivery, the vendors demand premiums for quick reaction times. As many failures require deep skills for handling, a significant training effort is necessary. Furthermore, the staff available must be able to cope even with multiple concurrent failures. Additionally, the failures may happen at unfavourable times thus create high overtime labour cost. Mobley (2002) and Stenström et al. (2016) estimate that reactive maintenance operations are approximately three times more expensive than scheduled ones. Also, Swanson (2001) found that reactive strategies have a negative impact on firm performance. Even after a quick repair the manufacturing of products and the delivery of services are interrupted resulting in lost revenue and profit. Furthermore, contractual penalties may also become due because production and availability targets are not met.
On the other hand, Swanson (2001) and Stenström et al. (2016) showed that proactive strategies in maintenance have a significant positive impact on performance. They can be differentiated into preventive and predictive strategies (Bateman, 1995). Preventive approaches have a breadth first approach. They apply maintenance actions to all parts. On the contrary, predictive strategies follow a depth-first approach. They try to identify those parts most likely to fail and fix them before failure.
The first predictive maintenance management approaches were time-driven (Mobley, 2002). They use models such as bathtub curves and statistical trend data to determine the schedule maintenance tasks. Furthermore, the observations of employees were used to detect indicators of a future failure, e.g. vibrations of machinery (Mobley, 2002). The new potential created by digital technologies was already identified by different authors (Kusiak et al., 2009; Ding and Kamaruddin, 2015). Newer predictive maintenance approaches schedule maintenance activities on as-needed basis by obtaining and analysing the actual operating conditions (Kusiak et al., 2009). For this purpose, a new type of information system, so-called predictive maintenance systems were designed (Canito et al., 2017; Chiu et al., 2017; Fernandes et al., 2018). They use the actual operation condition of plant equipment to optimize total plant operation.