Vacuum filling machine fault self-diagnosis system based on BP network

Network and Tongbiao Fault Self-diagnosis System for Vacuum Filling Machine Based on 6 Network LIU Wei, DENG Qin, YAN Peng, LIU Jun School of Mechanical and Precision Instrument, Xi'an University of Technology, Xi'an 710048, China Zhenzhong resin potting and sealing machine is used for epoxy of electronic devices. Resin potting guide equipment. Structure 1.

Into and 8 are the tank plunger pump and proportioning anvil system containing epoxy resin and curing agent, respectively. They are installed in the tank bottom and the tank lid; the proportioning system completes the setting of the filling rate and the curing agent ratio, and the plunger The pump is responsible for proportionally and quantitatively pressing the 6 materials into the mixer; the 6 mixers are fully mixed; the feeding of the workpiece to the workpiece is only the main structure of the infusion work, and the vacuum of the temperature control system is also available. The deaeration system stirring system, etc. are not given in the date of receipt 2002 This equipment is a reciprocating equipment. It is difficult to use a wide range of other reciprocating motion equipment, such as iron, chromatogram, etc. Here. We use the method of characterization of the inlet and outlet pressure waveforms of the static mixer. Successfully identify the plunger wear path plugging Forced valve anomaly Miscellaneous etc. 卞 To fault 1 Pressure waveform and feature extraction Zhuang static mixer outlet dipping material inlet 8 Feed force transmitter installed. And send the transmitter, 5 to the control counter to the machine. We compiled the corresponding program on the control computer to complete the signal acquisition and storage feature extraction feature to identify the fault report and other work. The pressure waveforms in the case of faults and the pressures in case of faults are not properly used when waveform 2 is 78. We take 3 pressures at 0.5, 1.0, 2.0, 3.0, 6.0, 65, and 7.08 to form a 21-dimensional characteristic vector, respectively. Yi = 12,12,. You 21 test that the pressure value sequence at the above moment can fully reach the operation of the various departments of the equipment. Inlet and outlet status Coring = 1 atmosphere 1唧 = 10 atm pressure pair; establishment and training of the network 28 Corresponds to the 21-dimensional characteristic address and 8 dimensions, 1 volume 2 揠 fault measurement must be forced valve abnormal pipeline blockage manual valve different, mixer deposition plug plunger 杲 wear leakage, perfusion forced valve abnormal pipeline leakage 欠, 7, network, 21 input nerves, 8 output neurons, hidden layer 25 neurons, hidden layer and output layer neurons, 3. Stimulate two numbers.

Simulate various fault conditions, perform expensive experiments on the test machine and prepare the fault to the anvil pair to save the file to avoid the standard algorithm's slow convergence, which may lead to local minimum. This method is used to solve the simulated annealing learning algorithm. Optimal Randomized Algorithm for Metal Annealing Processes.

Statistical thermodynamics shows that in the equilibrium state, the probability that the system can enthalpy can be divided into 3, 1, and force ranges at a given temperature. From Equation 1, it can be seen that when the system temperature is high, the probability of the system being in a higher energy state is higher, and when the system temperature is lower, the probability of the system in a higher energy state is very low, such as the system can be confined Seen as a feasible solution in the optimization calculation. The point with the lowest energy is then converted into the minimum point of the objective function. It can therefore be used for the learning of artificial neural networks. Specific process temperature speed, 6 adaptive coefficients, 1; 161 iteration steps, stochastic adaptation neighborhood factors; 1 quantum 51 to find the output of a given sample and bounded network and the corresponding energy function subroutine feature fault vector After training the data in the file, we will get a neural network 3 fault self-diagnosis system that can be called a state recognizer or, hence, a classifier. Put the trained state recognizer in the device control computer. , Work with the process control program at the same time and recognize the operating status at any time. If necessary, manual intervention can be performed. Correcting or retraining can be used to realize the advantages of human-machine integration. The combination of bit control and machine construction constitutes the fault self-diagnosis system of the filling machine. Conclusion It has been proved by practice that the state identifier established by us has a good ability to identify single faults and multiple faults that have been trained, but lacks many untrained faults. Lenovo capabilities. The recognition rate is not high. In order to solve this problem, it is necessary to improve the network structure, not just enough to learn 4 algorithm 1 Zhao Linming. Hu Haoyun. Multilayer forward artificial neural network

Zhengzhou Yellow River Conservancy Press, 1999.

2 Wu Jinpei, Xiao Jianhua. Intelligent fault diagnosis and expert system Tim. Beijing Science Press, 1997.

He and Ji, Chen Changzheng. Intelligent Diagnosis Based on Neural Network. Beijing Metallurgical Industry Press, 2000.

Xu, Wang Hong, Wang Wenhui. Artificial neural network principle and application. Shenyang Northeastern University Press, 2000.

Diesel Sensor For DELPHI

The fuel rail sensor, commonly referred to as the Fuel Pressure sensor.
It is part of the vehicle's fuel system and is designed to monitor fuel pressure on fuel tracks. The Pressure Sensor sends this signal to the electronic control unit, which then adjusts the vehicle's fuel and time.
When the sensor has an issue it can cause problems with the performance of the vehicle.

Fuel pressure regulator,Fuel pressure sensor,sensor,Fuel rail Pressure sensor

RUIAN DELLFEI AUTO PARTS CO.,LTD. , https://www.dellfei.com