System of the Automatic Preventive On-Line Monitoring and Diagnostics of Car Engines on the Basis of the New Methods of Preventive Diagnostics

<div class="section abstract"><div class="htmlview paragraph">This paper introduces architecture of an integrated system of preventive diagnostics and mathematical and computational methods, based on which such a system is being developed.</div><div class="htmlview paragraph">The present work aims to: <ul class="list disc"><li class="list-item"><div class="htmlview paragraph">describe the methods of preventive diagnostics based on mathematical models and computing algorithms, allowing to detect the hidden harbingers of engine dysfunctions and future failures;</div></li><li class="list-item"><div class="htmlview paragraph">describe the architecture of the system of preventive diagnostics, its further evolution;</div></li><li class="list-item"><div class="htmlview paragraph">describe the process of integration of preventive diagnostics system to the car engine.</div></li></ul></div><div class="htmlview paragraph">A necessary condition for the development system of preventive diagnosis is the condition of their economic efficiency, including the requirement of low cost sensors and computing systems of diagnostics.</div><div class="htmlview paragraph">The paper shows: <ul class="list disc"><li class="list-item"><div class="htmlview paragraph">A set of telemetry data for preventive diagnostics should include data of the crankshaft rotation angle sensor and data of vibrations of the engine case, received by means of the sensor control of vibrations installed additionally on its case;</div></li><li class="list-item"><div class="htmlview paragraph">The basis for the analysis of preventive signs of dysfunction and future failures of engine is the analysis of stochastic components in indications of established sensors, crankshaft rotation angle sensor and vibration sensor. The greatest number of hidden preventive diagnostic signs is defined by the analysis of stochastic components of signals from the vibration sensor.</div></li><li class="list-item"><div class="htmlview paragraph">Algorithms for detecting of preventive signs of dysfunctions of the engine represent the analysis of multidimensional empirical probability distribution functions of stochastic components in signals of crankshaft rotation angle sensor and vibration sensor.</div></li></ul></div><div class="htmlview paragraph">Along with the analysis of multidimensional distribution functions such characteristic as K-complexity, K-entropy of stochastic trajectory of process is introduced and analyzed.</div></div>