Accuracy of identification of the state of the internal combustion engine with the customized model using the measuring expert system

Abstract

Рассмотрены вопросы уменьшения погрешности идентификации технического состояния двигателя внутреннего сгорания и его составных частей как объекта экспертизы путем настраивания параметров применяемой в измерительной экспертной системе виртуальной модели ДВС. Для настройки модели предложено применение градиентного метода, обеспечивающего наиболее быструю минимизацию погрешности идентификации Purpose and methods. Improving the accuracy of identification for the technical condition of the internal combustion engine (ICE) in operational conditions using the engine measurement expert system (EMSE) is addressed by adjusting the computer dynamic model of the internal combustion engine. Results. Algorithmic schemes of computer models for the state of the ICE are obtained using the equations of its dynamics, which takes into account the factors such as the movement of the fuel supply body, the force on the hook - the load. The structural schemes of modeling at the input of a step-by-step action are presented. A promising method of tuning the model in the EMSE is proposed, which consists of measuring its working processes, in particular the angular acceleration of the crankshaft, for a specific brand of ICE. Then the corresponding set of models of its technical condition is obtained: normal, permissible, limit, pre-accident and emergency. By adjusting the values of the coefficients of these models in the EMSE, they achieve their coincidence with the actual state of the ICE. The identification error is minimized using the gradient method of steepest descent. The presence of several computer models is a practical advantage in the examination of the technical condition of the tested engines allowing its effective implementation in operational conditions. In this case, based on the experience of operation, the computer model closest to the actual state of the ICE is adjusted. At the same time, the efficiency of localization of ICE malfunctions increases, since the coefficients reflecting the state of the engine components and systems are consistently adjusted. Conclusions. The application of the proposed methodology using the criterion of minimizing the identification error by the gradient method allows implementation of this effective method for identifying the state of the ICE. It increases the reliability of determining the technical state of the ICE and its components by adjusting the computer model