Method for monitoring the technical condition of a diesel generator when in operation

A method for monitoring the technical condition of a diesel generator when in operation, involving measurement of the vibration acceleration values at three mutually orthogonal planes with the use of vibration sensors installed at monitoring points on the diesel generator, characterized in that primary measurements of the vibration acceleration values are preliminarily performed at the monitoring points of a knowingly fault-free operational diesel generator, and then subsequent measurements of the vibration acceleration values are performed at the monitoring points on the diesel generator in the course of its operation with the prescribed intervals, with additional measurement of temperature and intensity of an ultrasonic signal at the same monitoring points and determination of the root mean square values of the ultrasonic signal intensity, temperature and vibration acceleration, and the root mean square values of vibration velocity and vibratory displacement are also calculated on the basis of the measured values;

• the obtained values are presented in the form of matrices, then the obtained root mean square values are normalized;
• calculation of the co-variation matrices and their singular decomposition are performed in order to determine the eigen vectors and the eigen values, then the obtained data are projected on the eigenvector components with formation of clusters corresponding to the measurements at each point within the space of eigenvector components, then the specified interval formed as the range of distance between the previous measurement clusters is determined, and a conclusion is made on absence of any faults in the diesel generator if more than 50% of the current measurement clusters fall within the specified interval, or presence of any malfunctions of the diesel generator if less than 50% of the current measurement clusters fall within the specified interval, or presence of any fault in the diesel generator if more than 50% of clusters are below the specified interval limits.