This work is devoted to the problem of diagnosing the suspension of vehicles. The problem of monitoring the condition of the suspension is now the most urgent due to the constant growth of the vehicle fleet and the tightening of requirements for safe operation. Timely and accurate monitoring of the condition of the suspension can prevent the failure of entire vehicle components, as well as avoid such serious consequences as a road traffic accident.

The proposed diagnostic method makes it possible to single out "useful" sounds from the total number of suspension noises, after a comparative analysis, and indicate the node whose sound differs from the standard, serviceable one. This solution in diagnostics can significantly reduce the overall labor intensity by eliminating partial or complete disassembly of the suspension. As a result, despite the simplification, the accuracy of fault detection will only increase.

The aim of the work is to study the vibroacoustic signals emitted by the suspension units. With the help of sensors, signals are read out, then mathematical processing takes place using a wavelet transform on a computer. As a result of the research, a diagnostic method has been developed that allows detecting hidden defects of suspension assemblies and determining the degree of wear. The scientific novelty lies in the fact that the diagnostic process becomes automated; all signals taken by the sensors are processed in a computer or a special scanner. Information about the state of certain nodes is displayed on the display, in contrast to existing methods, where diagnostics is carried out visually or by ear. Thus, it becomes possible to avoid mistakes.

Experiments were carried out on real cars, it was possible to read the reference values ​​of serviceable units with the help of sensors, as well as the values ​​of faulty ones. After comparing the data, it became clearly visible that the developed method can really help in diagnostics. Now we are adapting this method for diagnosing internal combustion engines and we already have our results. The next stage of development will be the use of neural networks for processing the read vibroacoustic signals. This will speed up the process and make it as automated as possible.