Method and process model for pre-processing data from automated systems for monitoring the technical condition of combine harvesters

Modern self-propelled agricultural machines are characterized by their technological sophistication, complexity and high cost. Ensuring reliability and serviceability of such machines is the key task of technical service. The development of technical service technologies allows collecting, processing and forecasting the technical condition of machines based on the data obtained in the process of machine operation. Data collection is performed by built-in control systems to diagnose and detect malfunctions in the work of machine units and assemblies. The aim of the research was to develop and approbate the method of preliminary data processing obtained with the automatic system of technical condition monitoring and sensors installed on combine harvesters ACROS and to formalize the developed algorithm for further automation of data preparation process for technical analysis. Data quality was assessed according to the following criteria: data volume, data types, number of attributes, presence and number of omissions, presence of duplicates, presence of anomalies, matching categories, normalization and consistency of values, possible homogeneity, and segmentation. The tools used were Python, R, Pandas, NumPy, Matplotlib libraries. The authors established that raw data from analytical systems of technical condition control of combine harvesters were not suitable for analyzing and predicting the technical condition of units and assemblies due to many missing values. The process model construction based on the developed method of data pre-processing may be considered as a concept of an information system, which allows automating the data preparation process of technical condition control systems of combine harvesters for machine processing. The presented method allowed the authors to obtain structured and informative data and correct filling of omissions and to eliminate outliers and errors. The proposed process model provides transparency, control and optimization of data handling processes, will allow excluding errors and contradictions in their further analysis, and will provide repeatability of actions in the future while processing similar datasets received from the systems of technical condition control of combine harvesters.