Research On Fault Detection Method Of Water Flow Standard Facility Based On Dynamic Signal

In view of the complex structure of water flow standard facility,many moving parts,easy to occur occasional faults in the process of instrument calibration,it is difficult to ensure the stability and reliability of the standard facility,this paper studies the fault detection method of water flow standard facility based on dynamic signal.The main research contents are as follows:1.The common faults and fault characteristics in the working process of the facility are analyzed.The dynamic signals of flow and pressure in the facility are selected as the analysis object.The frequency domain characteristics of the signals are analyzed by using fast Fourier transform to extract the fault frequency characteristics.Based on the principle of energy conservation of unsteady flow,the relationship between flow and pressure in the pipeline of the facility was analyzed,and a correlation analysis method based on pressure and flow signals was proposed.In addition,the signal processing methods such as band-pass filtering and singular value decomposition(SVD)were combined to judge the position of single fault source and multiple fault source conditions.The extraction of fault features lays a foundation for the establishment of fault detection model based on chaotic oscillator.2.A fault detection model of water flow standard facility based on Holmes-Duffing system was proposed in view of the characteristics of large noise and weak main signals in the calibration process of water flow standard facility.In view of the limitation of the traditional detection model that chaos occurs only under small frequency parameters,the detection model is improved by using the time scale transformation method,so that it can detect weak periodic signals of any specified frequency.In addition,simulation experiments are carried out on the signals of different frequencies and different noise levels in MATLAB.Simulation results show that this detection model can effectively detect the signal of the specified frequency and has strong anti-noise ability.3.A real flow experiment platform was built based on the water flow standard facility,and the response speed level and upstream and downstream phase relationship of each sensor were determined through the experiments.Aiming at the common faults of water flow standard facility,the fault simulation experiment is designed,and the characteristic frequencies of different faults are extracted.The chaotic fault detection model is verified by inputting signals with different fault characteristic frequencies.The flow and pressure signals corresponding to each characteristic frequency after verification are analyzed to determine the location of the fault source.The above experiments verify the feasibility of the proposed method.