High-pressure oil jacking system working condition diagnosis method based on multi-source data
Through the working condition diagnosis method of high-pressure oil hoisting system based on multi-source data, the displacement detection unit and feedforward neural network are used to monitor and diagnose the health status of high-pressure oil hoisting systems in hydropower plants in real time, solving the difficulty in monitoring and monitoring in the existing technology. Diagnosing mechanical failures and control circuit failures ensures safe production in hydropower plants.
Patent Information
- Application Number
- CN202410960669.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-22
AI Technical Summary
The existing technology is difficult to effectively monitor and diagnose the operating conditions of the high-voltage oil hoisting system of hydropower plants, resulting in possible mechanical failures and control circuit failures, affecting the safe production of hydropower plants.
Using the operating condition diagnosis method of high-pressure oil hoisting system based on multi-source data, the monitoring system collects sensor data in real time, and uses displacement detection units and feedforward neural network to perform data preprocessing and fault diagnosis to judge the health status of the high-pressure oil hoisting system. .
Real-time operating condition diagnosis and fault judgment of the high-pressure oil hoisting system is realized, providing guidance for system maintenance, and ensuring the normal operation and safe production of the system.
Smart Images

Figure CN118998152A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of working condition diagnosis of a high-pressure oil control system of a hydropower plant, and in particular to a working condition diagnosis method of a high-pressure oil jacking system based on multi-source data. Background Art
[0002] The function of the high-pressure oil jacking system is to establish an oil film by injecting high-pressure oil between the thrust mirror plate and the thrust bearing pad surface of the turbine generator set during the start-up and shutdown process of the turbine generator set, so as to prevent the thrust pad from bearing too much weight due to slow rotation during the start-up and shutdown process, causing wear or burning of the pad surface or mirror surface. The high-pressure oil jacking system consists of a control panel, two control oil pumps, an oil filter, high-pressure oil system pipelines, an oil pressure sensor, an oil flow / oil pressure switch, and an oil pump power circuit. The high-pressure oil system is put into operation during the start-up and shutdown of the unit, when the unit speed is less than 90%, when the unit is in peristaltic motion, when the water machine is in backup motion, when a remote command is issued, and when the local operation is performed. Under normal conditions of high-pressure oil, a single pump is in operation. When the main pump fails, the main pump stops running, and the backup pump is put into operation to maintain the system operation.
[0003] Therefore, monitoring the operating conditions of the high-pressure oil jacking system and ensuring the normal functions of the control system and the oil pump motor are of great significance for ensuring the safe production of hydropower. Summary of the invention
[0004] The main purpose of the present invention is to provide a method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data, and to diagnose the working condition of the high-pressure oil jacking system based on sensor data collected in real time by a computer monitoring system of a hydropower plant, so as to provide guidance for system maintenance.
[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: a method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data, which is used for the high-pressure oil jacking system of a hydro-generator set, comprising the following steps: S1, the monitoring system host computer orders to start the high-pressure oil jacking system. If the conditions are met, it enters S2, otherwise the process ends; S2, the lower computer controller controls the high-pressure oil jacking system. If the feedback signal of the high-pressure oil control system is normal, it enters S3. Otherwise, it is determined that there is a control circuit fault in the high-pressure oil system, and the process ends; S3, the displacement detection unit detects the state of the high-pressure oil system. If the high-pressure oil jacking system is put into operation within the specified time T, the process proceeds to S4. Otherwise, it is determined that there is a mechanical failure in the high-pressure oil system and the process ends. S4, the host computer collects multi-source analog data of the high-pressure oil system and performs data preprocessing; S5. Input the preprocessed multi-source analog data into the fault diagnosis module to calculate the health status of the high-pressure oil jacking system.
[0006] In S1, the steps for the monitoring system host computer to order the high-pressure oil jacking system to be put into use are as follows: S11. Command the high-pressure oil jacking system to be put into operation through the human-machine interactive interface of the computer monitoring system, and the monitoring system sends the command to the PLC; S12, PLC receives the command from the host computer and determines whether the conditions of "the control mode of the high-pressure oil control system is remote", "the automatic control conditions of the local control unit are met", and "the high-pressure oil jacking system has no fault alarm" are met; if the conditions are met, enter S2, otherwise the process ends.
[0007] In S2, the lower computer controller uses the output quantity module and the output relay to input the high-pressure oil jacking system. If the contactor feedback signal of the high-pressure oil control system is normal, it enters S3.
[0008] In S3, the steps for the displacement detection unit to detect the high-pressure oil system are as follows: S31, detecting the movement of the rotor through the displacement detection unit, and if the movement duration is less than T, it is determined that the rotor is lifted, and the displacement detection unit transmits a signal to the PLC DI module; S32, PLC calculates the high-pressure oil control system input time T dif , calculated as the timestamp T corresponding to the current time n The timestamp T of the moment when the PLC receives the command from the host computer to start the high-pressure oil jacking system n-1 The difference between dif >T, it is determined that the high-pressure oil system has a mechanical failure and the process ends; otherwise, it enters S4.
[0009] In S4, the steps for the host computer to collect multi-source analog data of the high-pressure oil system and perform data preprocessing are as follows: S41, the host computer collects the sensor data installed in the high-pressure oil jacking system, including the high-pressure oil pipeline outlet pressure p t , oil pump motor temperature K t , motor current I t , bearing vibration displacement D t data; S42, respectively normalize the collected data, the formula is: ; Where: To normalize the data, To lower the range of this data, To scale the data, is the normalized data; The normalized data are combined into feature vectors [e1,e2...e n ] T ; Where: e is the normalized data of a sensor, n is the number of sensors; S43, the weights of the feedforward neural network are trained using a training set constructed from manually annotated historical data of the monitoring system. ; in: is the sensor data after the same preprocessing steps as S42, This is the working condition marked value. When the high pressure oil control system is working normally 1 when abnormal; 0 when abnormal.
[0010] When training the neural network, the optimizer was set to LBFGS and the number of iterations was set to 200 rounds. In S5, the fault diagnosis module is a feedforward neural network including an input layer, a double hidden layer and an output layer; wherein the input layer includes n neurons, the first hidden layer includes 5 neurons, the second hidden layer includes 4 neurons, the output layer includes 1 neuron, and the activation function uses ReLU.
[0011] In S5, the method for calculating the health status of the high-pressure oil jacking system is: The eigenvector [e1,e2...e n ] T Input to the fault diagnosis module, get the output value output, calculate the confidence C = sigmoid (output), and calculate the health status of the high-pressure oil control system HC = C × 100%; where: e is the normalized sensor data, and n is the number of sensors.
[0012] The present invention has the following beneficial effects: The present invention provides a method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data. Based on the displacement detection unit data and sensor data collected in real time by a computer monitoring system of a hydropower plant, the method judges and diagnoses the working condition of the high-pressure oil jacking system, provides guidance for system maintenance; and prevents problems before they occur by calculating the health status of the high-pressure oil control system. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.
[0014] Figure 1 It is a flow chart of the present invention.
[0015] Figure 2 This is the feedforward neural network diagram of the fault diagnosis module. DETAILED DESCRIPTION
[0016] Embodiments of the present invention are described in detail below. Examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, but should not be construed as limiting the present invention.
[0017] See also Figure 1 , 2 A method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data, which is used for a high-pressure oil jacking system of a hydro-generator set, comprises the following steps: S1, the monitoring system host computer orders to start the high-pressure oil jacking system. If the conditions are met, it enters S2, otherwise the process ends; S2, the lower computer controller controls the high-pressure oil jacking system. If the feedback signal of the high-pressure oil control system is normal, it enters S3. Otherwise, it is determined that there is a control circuit fault in the high-pressure oil system, and the process ends; S3, the displacement detection unit detects the state of the high-pressure oil system. If the high-pressure oil jacking system is put into operation within the specified time T, the process proceeds to S4. Otherwise, it is determined that there is a mechanical failure in the high-pressure oil system and the process ends. S4, the host computer collects multi-source analog data of the high-pressure oil system and performs data preprocessing; S5. Input the preprocessed multi-source analog data into the fault diagnosis module to calculate the health status of the high-pressure oil jacking system.
[0018] The present invention judges and diagnoses the working condition of the high-pressure oil jacking system based on the displacement detection unit data and sensor data collected in real time by the computer monitoring system of the hydropower plant, and provides guidance for system maintenance; and prevents problems before they occur by calculating the health status of the high-pressure oil control system.
[0019] In S1, the steps for the monitoring system host computer to order the high-pressure oil jacking system to be put into use are as follows: S11. Command the high-pressure oil jacking system to be put into operation through the human-machine interface of the computer monitoring system, and the communication software of the monitoring system sends the command to the PLC through the Modbus TCP protocol; S12, PLC receives the command from the host computer and determines whether the conditions of "the control mode of the high-pressure oil control system is remote", "the automatic control conditions of the local control unit are met", and "the high-pressure oil jacking system has no fault alarm" are met; if the conditions are met, enter S2, otherwise the process ends.
[0020] In S2, the lower computer controller uses the output quantity module and the output relay to input the high-pressure oil jacking system. If the contactor feedback signal of the high-pressure oil control system is normal, it enters S3.
[0021] In S3, the steps for the displacement detection unit to detect the high-pressure oil system are as follows: S31, detecting the movement of the rotor through the displacement detection unit, and if the movement duration is less than T, it is determined that the rotor is lifted, and the displacement detection unit transmits a signal to the PLC DI module; S32, PLC calculates the high-pressure oil control system input time T dif , calculated as the timestamp T corresponding to the current time n The timestamp T of the moment when the PLC receives the command from the host computer to start the high-pressure oil jacking system n-1 The difference between dif >T, it is determined that the high-pressure oil system has a mechanical failure and the process ends; otherwise, it enters S4.
[0022] The displacement detection unit can be a displacement sensor or a vibration sensor. Taking the vibration sensor as an example, the vibration sensor sets a certain vibration threshold and is installed above the generator rotor. When the vibration sensor detects vibration and the vibration duration is less than T, it is judged that the rotor is lifted, and the vibration sensor transmits the signal to the PLC DI module.
[0023] In S4, the steps for the host computer to collect multi-source analog data of the high-pressure oil system and perform data preprocessing are as follows: S41, the host computer collects the sensor data installed in the high-pressure oil jacking system, including the high-pressure oil pipeline outlet pressure p t , oil pump motor temperature K t , motor current I t , bearing vibration displacement D t data; S42, respectively normalize the collected data, the formula is: ; Where: To normalize the data, To lower the range of this data, To scale the data, is the normalized data; The normalized data are combined into feature vectors [e1,e2...e n ] T ; Where: e is the normalized data of a sensor, n is the number of sensors; S43, the weights of the feedforward neural network are trained using a training set constructed from manually annotated historical data of the monitoring system. ; in: is the sensor data after the same preprocessing steps as S42, This is the working condition marked value. When the high pressure oil control system is working normally 1 when abnormal; 0 when abnormal.
[0024] When training the neural network, the optimizer was set to LBFGS and the number of iterations was set to 200 rounds. In S5, see Figure 2 The fault diagnosis module is a feedforward neural network including an input layer, a double hidden layer and an output layer; the input layer contains n neurons, the first hidden layer contains 5 neurons, the second hidden layer contains 4 neurons, the output layer contains 1 neuron, and the activation function uses ReLU.
[0025] In S5, the method for calculating the health status of the high-pressure oil jacking system is: The eigenvector [e1,e2...e n ] T Input to the fault diagnosis module, get the output value output, calculate the confidence C = sigmoid (output), and calculate the health status of the high-pressure oil control system HC = C × 100%; where: e is the normalized sensor data, and n is the number of sensors.
Claims
1. A method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data, which is used for the high-pressure oil jacking system of a hydro-generator set. It is characterized in that The following steps are involved: S1, the monitoring system host computer orders to start the high-pressure oil jacking system. If the conditions are met, it enters S2, otherwise the process ends; S2, the lower computer controller controls the high-pressure oil jacking system. If the feedback signal of the high-pressure oil control system is normal, it enters S3. Otherwise, it is determined that there is a control circuit fault in the high-pressure oil system, and the process ends; S3, the displacement detection unit detects the state of the high-pressure oil system. If the high-pressure oil jacking system is put into operation within the specified time T, the process proceeds to S4. Otherwise, it is determined that there is a mechanical failure in the high-pressure oil system and the process ends. S4, the host computer collects multi-source analog data of the high-pressure oil system and performs data preprocessing; S5. Input the preprocessed multi-source analog data into the fault diagnosis module to calculate the health status of the high-pressure oil jacking system.
2. A method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data according to claim 1, It is characterized in that In S1, the steps for the monitoring system host computer to order the high-pressure oil jacking system to be put into use are as follows: S11. Command the high-pressure oil jacking system to be put into operation through the human-machine interactive interface of the computer monitoring system, and the monitoring system sends the command to the PLC; S12, PLC receives the command from the host computer and determines whether the conditions of "the control mode of the high-pressure oil control system is remote", "the automatic control conditions of the local control unit are met", and "the high-pressure oil jacking system has no fault alarm" are met; if the conditions are met, enter S2, otherwise the process ends.
3. According to claim 1, a method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data, It is characterized in that In S2, the lower computer controller uses the output quantity module and the output relay to input the high-pressure oil jacking system. If the contactor feedback signal of the high-pressure oil control system is normal, it enters S3.
4. The method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data according to claim 1, It is characterized in that In S3, the steps for the displacement detection unit to detect the high-pressure oil system are as follows: S31, detecting the movement of the rotor through the displacement detection unit, and if the movement duration is less than T, it is determined that the rotor is lifted, and the displacement detection unit transmits a signal to the PLC DI module; S32, PLC calculates the high-pressure oil control system input time T dif , calculated as the timestamp T corresponding to the current time n The timestamp T of the moment when the PLC receives the command from the host computer to start the high-pressure oil jacking system n-1 The difference between dif >T, it is determined that the high-pressure oil system has a mechanical failure and the process ends; otherwise, it enters S4.
5. The method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data according to claim 1, It is characterized in that In S4, the steps for the host computer to collect multi-source analog data of the high-pressure oil system and perform data preprocessing are as follows: S41, the host computer collects the sensor data installed in the high-pressure oil jacking system, including the high-pressure oil pipeline outlet pressure p t , oil pump motor temperature K t , motor current I t , bearing vibration displacement D t data; S42, respectively normalize the collected data, the formula is: ; Where: To normalize the data, To lower the range of this data, To scale the data, is the normalized data; The normalized data are combined into feature vectors [e 1 ,e 2 ...e n ] T ; Among them: e is the normalized data of a sensor, n is the number of sensors; S43, the weights of the feedforward neural network are trained using a training set constructed from manually annotated historical data of the monitoring system. ; in: is the sensor data after the same preprocessing steps as S42, This is the working condition marked value. When the high pressure oil control system is working normally 1 when abnormal; 0 when abnormal.
6. A method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data according to claim 5, It is characterized in that When training the neural network, the optimizer was set to LBFGS and the number of iterations was set to 200 rounds.
7. The method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data according to claim 1, It is characterized in that In S5, the fault diagnosis module is a feedforward neural network including an input layer, a double hidden layer and an output layer; wherein the input layer includes n neurons, the first hidden layer includes 5 neurons, the second hidden layer includes 4 neurons, the output layer includes 1 neuron, and the activation function uses ReLU.
8. A method for diagnosing the working condition of a high-pressure oil jacking system based on multi-source data according to claim 7, It is characterized in that In S5, the method for calculating the health status of the high-pressure oil jacking system is: The eigenvector [e 1 ,e 2 ...e n ] T Input to the fault diagnosis module, get the output value output, calculate the confidence C = sigmoid (output), and calculate the health status of the high-pressure oil control system HC = C × 100%; where: e is the normalized sensor data, and n is the number of sensors.
