In-field failure assessment of tractor hydraulic system operation via pseudospectrum of acoustic measurements

Hydrological working conditions are critical indicators for the maintenance of tractor mechanical systems. This research presents a pseudospectrum approach to analyze the particle pollution level of the hydraulic filter in its aging process and thus on-time prediction and diagnosis of weak or faulty conditions. The practical experiments of pseudospectrum analysis are performed on oil filter sound records. For mobile application purposes, besides deployment of advanced audio recording equipment, three popular brands of cell phones are used. The soundtracks are recorded in different incremental stages of fluid contamination by pollutive particles until being choked based on the ISO4406 standard. The pseudospectrum of the oil filter acoustics leads to two numerical features: the low-band energy and 12.5 kHz relative peak energy (12.5RPE). This manuscript proposes these two features for efficient on-line monitoring of the liquid cleanliness level as well as the prediction and diagnosis of choked condition of the tractor hydraulic system.