This paper proposes an intelligent diagnosis method for plant machinery in multi-fault state using wavelet analysis, genetic programming (GP), and possibility theory. The wavelet analysis is used to extract feature spectra of multi-fault state from measured vibration signal for the diagnosis. Excellent symptom parameters for distinguishing fault states are automatically generated by GP. Because the value of symptom parameter calculated to express the feature of the vibration signal fluctuates even if machine state does not change, fuzzy diagnosis is necessary. After obtaining the excellent symptom parameters by GP called GP-SPs, the membership functions of GP-SPs are needed for fuzzy diagnosis. We also discuss the identification method of membership function of symptom parameters using probability theory and possibility theory, and show the inference method for identifying faults types. The methods proposed in this paper are verified by applying them to the diagnosis of rolling bearing in multi-fault state.

Dimensional or nondimensional symptom parameters are usually used for condition monitoring of plant machinery. However, it is difficult to extract the most important symptom parameters and the functions of those parameters by which machinery faults can be sensitively detected and the fault types can be precisely distinguished. In order to overcome this difficulty and to ensure highly accurate fault diagnosis, a new method, called "automated function generation of symptom parameters" using genetic algorithms (GA) is presented in this paper. By applying the method to real machinery diagnosis problems, it has been shown that the key symptom parameter function can be quickly generated. We give a diagnosis example of rolling bearings whose operating conditions are variable in terms of rotation speed and load.

The paper proposes a failure diagnosis method for machinery in unsteady operating condition using instantaneous power spectrum (IPS) and genetic programming (GP). The IPS is used to extract feature frequency of each machine state from measured vibration signal for distinguishing failures by the Relative Crossing Information (RCI). Excellent symptom parameters for detecting failures are automatically generated by GP. The method proposed in the paper is verified by applying it to the failure diagnosis of rolling bearing.

Here we present the new robust condition monitoring and diagnosis method based on the statistical hypothesis on vibration characteristics of the rotating machines in good condition. The hypothesis is that if the machine is in good condition, its probability density function of vibration signal follows the normal distribution in time domain. This method can lead to the robust failure diagnosis without any prior knowledge concerning vibration characteristics corresponding to specific failure to be detected.

The purpose of this study is to develop the basic theories and techniques for a plant inspection and diagnosis robot (IDR). The robot dealt with in this study will work in a large scale unmanned plant or a place with dangerous environment. It has the ability to monitor the condition of plant machinery with only few sensors, and quickly to discriminate machine failures, in order to guarantee both the quality and quantity of production against accident. The paper proposes a method for detecting a faulty part of a plant machine by the robot. A manipulator installed on the IDR is controlled by genetic algorithms (GA). A microphone installed on the manipulator tip is used to detect a failure signal. It is navigated to the nearby front of the faulty part by sound information and GA control. The method has been proved by practical applications.