Recognition of Operating States of a Wheel Loader for Diagnostics Purposes

In this paper, the operating states of a wheel loader were studied for diagnostics purposes using a real time simulation model of an articulated-frame-steered wheel loader. Test drives were carried out to obtain measurement data, which were then analyzed. The measured time series data were analyzed to find the sequences of operating states using two different data sets, namely the variables of hydrostatic transmission and working hydraulics. A time series is defined as a collection of observations made sequentially in time. In our proposed method, the time series data were first segmented to find operating states. One or more segments build up an operating state. A state is defined as a combination of the patterns of the selected variables. The segments were then clustered and classified. The operating states were further analyzed using the quantization error method to detect anomalies. The recognized operating states define the operation of the machine so the analysis can be focused on specific sections and situations in time series and to identify which kinds of operating situations generate anomalies. Simulated leakages in the main hydraulic components of the hydrostatic transmission and the working hydraulics were used as anomalies to study the changes in the recognized operating states and the magnitude of the quantization error.