Research status of fault diagnosis of large and medium-sized vertical water pump units

This paper analyzes the development process and status quo of the fault diagnosis technology of vertical water pump units. Several common fault detection methods of vertical water pump units are introduced in detail from the analytical model-based processing method, signal-based processing method, and knowledge-based processing method. The advantages and disadvantages are analyzed and summarized, providing theoretical methods for fault diagnosis of large and medium-sized vertical water pump units


Introduction
Troubleshooting of large and medium-sized vertical pump units is an indispensable part of ensuring the safe operation of pumping stations. Vertical water pump is a common type of water pump. If there is no regular troubleshooting, maintenance will cost more time, and money once the vertical pump fails and may cause serious safety incidents. Vertical pump unit faults mainly include abnormal vibration faults, bearing wear, cavitation, and so on.
Vibration is the most common form of vertical pump unit faults. In order to ensure the working efficiency and reliability of vertical pump unit, vibration monitoring should be carried out regularly to infer whether the vertical pump unit failed. However, the vibration situation caused by different faults is also different, so it is very necessary to process and analyze the detection signal, judge the cause of the fault and put forward corresponding improvement measures. Therefore, it is one of the effective methods to ensure the safe and stable operation of the pump station to monitor and analyze the vibration signal and infer whether the running condition of the vertical pump unit is good or not.

The development process and present situation of fault diagnosis technology
The development of vibration fault treatment of vertical water pump units in China can be divided into three stages: The first stage is the post-processing stage. At that time, due to people's weak awareness of equipment maintenance and overhaul and relatively simple tools, abnormal conditions of the equipment could not be found in time in most cases and could only be repaired after the failure.

Vibration fault diagnosis method of vertical water pump Unit based on analytic model
In order to describe the fault of the vertical pump system in practical engineering better, we need to analyze the signal detected by the sensor or estimate the physical parameters required to build the system. By combining the changes of physical parameters and the actual occurrence of the fault phenomenon, we can establish the vertical pump failure mathematical model. In this way, the fault can be analyzed, predicted, and located. The analytical model-based fault diagnosis methods of vertical pump units can be roughly divided into three kinds: state estimation method, equivalent space method, and parameter estimation method.

State estimation method
State estimation is a method to estimate the internal state change of a dynamic system based on the detection data.
To build an accurate mathematical model of vertical water pump, and through the sensor to measure the specific data of the equipment, automatically eliminate the system due to noise or other effects of the interference information. By comparing the two data, the difference can be used to judge the fault of the equipment. Through the state estimation method, the data can be detected and identified to predict the future working state of the vertical water pump.
However, the premise of using the state estimation method is to build an accurate mathematical model, but it is difficult to achieve at present. Therefore, the current research focuses on enhancing the anti-jamming ability of the model against a series of errors such as interference and noise caused by the installation of the detection system. small, and there will be no alarm or missed alarm. If the residual threshold is too low, a slight increase in the residual will cause an alarm, but in this case the pump unit may be intact. To sum up, it is necessary to pay attention to the choice of residual threshold when using equivalent spatial generator for fault diagnosis of vertical pump system. [3]

Parameter estimation method
When the vertical pump system fails, the system In general, X(f) is a complex function，we have In the formula, ( ) represents the amplitude distribution of the frequency curve decomposed by the signal， ( ) is called the phase spectrum，which can represent the phase distribution of the frequency curve. [5]

Principal component analysis
The

Wavelet analysis
Although spectrum analysis is a convenient and and is convenient for multi-component signal analysis. [8] The working flow chart of detecting faults of vertical pump unit by wavelet analysis is shown in Figure 2:

Vibration fault diagnosis method of vertical water pump unit based on knowledge
The

Expert system method
Expert system method is to store a large number of existing fault information, after collecting the detection signal of the vertical pump system, through comprehensive comparison, rapid analysis and processing. The possible fault location and its specific cause can be located through a series of calculation and reasoning. The repository of expert system stores a large number of data such as fault cause, symptom, manifestation, suggestion. Therefore, it only needs to compare the measured signal with the known data to calculate the most similar optimal solution. Schematic diagram of expert system method is shown in Figure 3.
Expert system method has been widely used in the fault diagnosis of vertical water pump, but this method has some limitations because the content of knowledge base is limited, therefore, it cannot deduce the fault data that has not appeared. [10] Fig. 3. Schematic diagram of expert system method there are certain requirements for sample selection. [11]

Fuzzy fault diagnosis method
The fuzzy fault diagnosis method is simply to establish wavelet analysis and other methods. [12]

Conclusion
At present, the vertical pump is more and more widely