Hydropower station data monitoring system and method based on artificial intelligence

Abstract

The invention relates to the field of hydropower station monitoring, in particular to a hydropower station data monitoring system and method based on artificial intelligence, and the system comprises a variable sensing module, an intelligent network module, an additional pool module, an input addressing module and a water pressure recovery module. The intelligent network module is used for constructing state functions of first-order variables and second-order variables of a hydropower station, the additional pool module is used for establishing a second reservoir and connecting a water turbine through a bidirectional gate, the input addressing module is used for detecting abnormal working conditions and outputting abnormal variables, and the water pressure recovery module is used for reconstructing the second-order state functions. By means of the method, the water turbine can operate at the optimal rotating speed, the water flow, the power generation amount and the equipment operation condition are better controlled, and therefore the power generation efficiency is optimized, the safety and the operation and maintenance efficiency of equipment are improved, and stable operation of a hydropower station is ensured.

Description

Technical Field

[0001] The present invention relates to the field of hydropower station monitoring, and in particular to an artificial intelligence-based hydropower station data monitoring system and method. Background Art

[0002] Hydropower station data monitoring is the process of real-time and regular collection, transmission, storage, analysis and management of various data information generated during the operation of the hydropower station, which helps to understand the overall operation status of the hydropower station and the operating status of the motor. The monitoring of the working status of the hydropower station usually includes data collection by sensors and automatic control systems, as well as comparative analysis of water storage status, excitation motors and power generation to determine whether there are any abnormalities in the operation of the hydropower unit.

[0003] However, during the operation of the hydropower station, the hydraulic parameters of the reservoir and the power generation data of the turbine are constantly changing, and the parameters affecting the operation of the turbine have different manifestations under different operating conditions. It is difficult for sensors to accurately capture the abnormal operating conditions of the power station, and it is also difficult to find out some parameters that affect the operation of the turbine, resulting in monitoring lags and difficulty in analyzing monitoring data.

[0004] In addition, since the water inlet valve of the turbine is directly connected to the water reservoir, the speed of the turbine is not subject to human control without adjusting the gate. Even when some abnormalities are detected in the turbine power generation parameters, the power generation state of the turbine cannot be effectively adjusted, resulting in part of the power generation energy being wasted, reducing the power generation efficiency of the hydropower plant. Summary of the invention

[0005] The purpose of the present invention is to provide a hydropower station data monitoring system and method based on artificial intelligence to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: A hydropower station data monitoring system based on artificial intelligence, comprising: a variable sensing module, an intelligent network module, an additional water tank module, an input addressing module and a water pressure recovery module;

[0007] The variable sensing module is used to set sensors in the reservoir and the turbine motor of the hydropower station. The sensors read the hydrological data in the reservoir and the power generation data of the turbine motor, upload the data to the control center, and update the hydrological data in the reservoir in real time after the hydropower station is working;

[0008] The intelligent network module is used to use hydrological data as the first-order input variable, the excitation motor speed as the first-order output variable, the power generation data as the second-order input variable, and the power generation power of the turbine as the second-order output variable, open the reservoir gate for trial operation, use the power generation data and hydrological data detected during the trial operation to train the state classification network, output the state function of the first-order variable and the second-order variable, obtain the time domain attenuation speed from the coefficient matrix of each variable in the state function, and remove the sensor in the turbine motor after the trial operation is completed;

[0009] The additional water tank module is used to establish a second water tank on the basis of the original water tank of the hydropower station, and the hydrological data corresponding to the highest rate of change of the first-order state function is used as the hydrological characteristics of the second water tank. The first and second water tanks are connected to the turbine by using a two-way gate, and the control component for controlling the opening and closing of the gate is connected to the control center. The gate of the second water tank is initially in a closed state;

[0010] The input addressing module is used to input the hydrological data in the first water reservoir into the state function, output the theoretical power generation of the turbine, and when the difference between the actual power generation of the turbine and the theoretical power generation exceeds the threshold, switch the gate to the second water reservoir, stabilize the hydrological data of the second water reservoir, keep the speed of the excitation motor constant, record the decay rate of the generator power generation, compare the decay rate with the time domain decay rate of each variable, and output the closest variable as the abnormal variable;

[0011] The water pressure recovery module is used to calculate the theoretical value of the abnormal variable from the state function and the actual power generation, reconstruct the second-order state function with the theoretical value of the abnormal variable as a constant, calculate the speed of the excitation motor when the power generation is maximum according to the reconstructed equation, and control the two-way valve so that the second water reservoir can meet the hydrological data required when the speed of the excitation motor is stable.

[0012] Further, the variable sensing module includes: a hydrological sensing unit and a motor sensing unit;

[0013] The hydrological sensing unit is used to detect hydrological data in the reservoir, and the hydrological data includes: water storage capacity, water level difference, gate opening and water storage loss rate;

[0014] The motor sensor unit is used to detect the power generation data of the water turbine motor, and the power generation data includes: the rotation speed of the excitation motor, the valve pressure of the unit, the guide vane opening, the oil temperature and the shoe surface temperature.

[0015] Further, the intelligent network module includes: an operation test unit, a state identification unit and a variable attenuation unit;

[0016] The operation test unit is used to open the gate of the water storage tank to test the motor, record the power generation data of the motor during the test run, and close the gate after the test run is completed;

[0017] The state recognition unit is used to train the neural network model using the power generation data during the trial operation, and to fit the state function of the input and output variables using the trained model;

[0018] The variable attenuation unit is used to obtain the coefficient matrix in the state function, and obtain the time domain attenuation speed of each variable from the coefficient matrix.

[0019] Further, the additional water tank module includes: a water tank unit and a valve control unit;

[0020] The water reservoir unit is used to divide an independent second water reservoir in the hydropower station, and the second water reservoir plays the role of diverting, testing and buffering the power generation process of the hydropower station;

[0021] The valve control unit is used to connect the first and second water storage tanks and the turbine using a controllable two-way valve to adjust the water flow pressure received by the turbine.

[0022] Further, the input addressing module includes: a pattern recognition unit, a variable control unit and an abnormality locating unit;

[0023] The pattern recognition unit is used to output the excitation motor speed through the first-order state function, output the turbine theoretical power through the second-order state function, and identify abnormal power generation conditions;

[0024] The variable control unit is used to stabilize the speed of the excitation motor of the turbine by using the second water reservoir;

[0025] The abnormality locating unit is used to obtain the attenuation rate of the turbine power according to the time domain variation of the actual power generation of the turbine, compare it with the time domain attenuation rate of each variable, and locate the abnormal variable.

[0026] Further, the water pressure recovery module includes: a state reconstruction unit and a hydraulic communication unit;

[0027] The state reconstruction unit is used to reconstruct the second-order state function with the theoretical value of the abnormal variable as a constant to calculate the maximum generating speed of the excitation motor;

[0028] The hydraulic communication unit is used to establish hydrological data corresponding to the maximum power generation speed in the second water storage tank until the power generation stage ends.

[0029] A hydropower station data monitoring method based on artificial intelligence includes the following steps:

[0030] Step S1. After the hydropower station is installed, a test run is carried out to detect the hydrological data of the reservoir and the power generation data of the turbine during the test run, and the neural network model is trained using the historical hydrological data and power generation data to obtain a trained state classification model;

[0031] Step S2. Taking the hydrological data and the excitation motor speed as the first-order input and output variables, and the power generation data and the power generation power as the second-order input and output variables, the hydrological data and the power generation data detected during the trial operation are input into the state classification model to obtain the first-order and second-order state functions, and the time domain attenuation speed of each power generation data is obtained from the coefficient matrix of the second-order state function;

[0032] Step S3. A second reservoir is established on the basis of the original first reservoir of the hydropower station, the water level height when the output of the first-order state function changes the most is used as the water level height of the second reservoir, and a two-way gate is used to connect the first and second reservoirs to the turbine;

[0033] Step S4. After the turbine is working, the hydrological data in the first reservoir is input into the first and second state functions in real time, and the theoretical power generation of the turbine is output. When the difference between the actual power generation of the turbine and the theoretical power generation exceeds a threshold, the two-way gate is switched to the second reservoir, and the water level of the second reservoir is stabilized;

[0034] Step S5. Record the decay rate of the generator power, compare the decay rate with the time domain decay rate of each variable, output the closest variable as the abnormal variable, and adjust the water level of the second reservoir to maximize the power generation.

[0035] Further, step S1 includes:

[0036] Step S11. The hydropower station is put into trial operation. Sensors are set in the reservoir and the turbine motor to detect the hydrological data in the reservoir and the power generation data of the turbine motor at the same frequency. The hydrological data include: water storage capacity, water level difference, gate opening and water storage loss rate. The power generation data include: excitation motor speed, unit valve pressure, guide vane opening, oil temperature and tile surface temperature.

[0037] Step S12. extracting historical hydrological data and power generation data from the database, wherein the historical hydrological data and power generation data are obtained from data recorded during the historical period of installed capacity operation of the hydropower station;

[0038] Step S13. Use historical hydrological data and power generation data to train a neural network model, and use the neural network model to fit the correlation between data pieces. The trained neural network model is output as a state classification model. The selected neural network models include: convolutional neural network, perceptron network, long short-term memory network and autoencoder network.

[0039] Further, step S2 includes:

[0040] Step S21. Take the hydrological data as the first-order input variable and the excitation motor speed as the first-order output variable, and fit the state function of the input variable and the output variable through the state classification model to obtain the first-order state function, which is expressed as:

[0041]

[0042] Where D(x1,x2,…,xn) represents the first-order state function, n represents the number of first-order input variables, x1, x2,…xn represent the 1st to nth first-order input variables respectively, X1 T To Xn T Respectively represent the transposed matrices of the detection value matrices of the 1st to nth variables during the trial run, A1 to An represent the coefficient matrices of the 1st to nth first-order input variables, c∈{1,2,…,n}, m represents the number of detections during the trial run, X c (1) to X c (m) represents the detection values ​​of the cth input variable from the 1st to the mth time respectively;

[0043] Step S22. Use the power generation data as the second-order input variable and the power generation as the second-order output variable, and output the second-order state function G(r1, r2, ..., rn) from the state classification model, where r1 to rn represent the 1st to nth second-order input variables respectively;

[0044] Step S23. Calculate the time domain decay rate function F(t) corresponding to each power generation data, where F(t) = [B flor(t / t0) +B flor(t / t0)+1 ]·mod(t / t0) / t0, where t represents time, t0 represents the sampling interval during the trial run, flor is the floor function, mod is the remainder function, B flor(t / t0) and B flor(t / t0)+1 They respectively represent the values ​​of the flor(t / t0)th and flor(t / t0)+1th elements in the coefficient matrix of the second-order state function.

[0045] Further, step S3 includes:

[0046] Step S31. Establish a second reservoir in the hydropower station, the second reservoir plays the role of diverting, testing and buffering the power generation process of the hydropower station, calculate the partial derivative of the first-order state function D(x1, x2, ..., xn) with respect to the water level height variable, and the water level height when the partial derivative function takes the maximum value is used as the water level height of the second reservoir;

[0047] Step S32. Use a two-way gate to connect the first and second water reservoirs and the turbine, so that the water level in the second water reservoir can be adjusted through the two-way gate connected to the first water reservoir.

[0048] Further, step S4 includes:

[0049] Step S41. The turbine is connected to the first water reservoir to work, and the hydrological data in the first water reservoir is obtained. The hydrological data is input into the first-order state function to obtain the theoretical value of the excitation motor speed. The excitation motor speed and the initial sampling values ​​of the remaining power generation data during the trial operation are input into the second-order state function to obtain the theoretical value of the generator power generation;

[0050] Step S42. When |P1-P2|≥P0, it is determined that there is an abnormality in power generation, and the process goes to step S43, where P1 represents the theoretical value of power generation, P2 represents the actual power generation, and P0 represents the preset threshold value;

[0051] Step S43. Switch the gate leading to the abnormal turbine from the first water reservoir to the second water reservoir, and at the same time control the connecting gates in the first and second water reservoirs to keep the water level in the second water reservoir constant.

[0052] Further, step S5 includes:

[0053] Step S51. Adjust the hydrological parameters of the second reservoir to keep them consistent with the changes in the hydrological parameters of the reservoir during the trial operation, continuously detect the power generation of the turbine, and output the power generation change function P(t);

[0054] Step S52: Calculate the correlation between the time domain attenuation of the power generation change function and the time domain attenuation of the variable:

[0055]

[0056] Among them, Q represents the correlation between the generated power and the time-domain attenuation of the variable, and T represents the time from the hydrological stabilization of the second reservoir to the present;

[0057] Step S53. Execute step S52 for each second-order variable, wherein the second-order variable corresponding to the minimum value of Q is output as the abnormal variable of the hydro-motor;

[0058] Step S54. Adjust the water level of the reservoir, detect the change in the generated power, record the water level when the generated power reaches the maximum value, and before the turbine is shut down for maintenance, adjust the water level of the second reservoir to be the same as the recorded water level.

[0059] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0060] 1. The present invention can perform input and output tests on the generator, use the state classification network to process the test data, construct the second-order state function of the water storage stage and the power generation stage, calculate the time domain attenuation rate of each input variable by the second-order state function, model the detection data of the hydropower station, reduce the dependence of the detection data on the timeliness of the sensor, and better control the water flow, power generation and equipment operation, thereby optimizing the power generation efficiency.

[0061] 2. The present invention can establish a second water reservoir in the hydropower station, and use a two-way gate to connect the turbine with the first and second water reservoirs. The opening and closing of the two-way gate is adjusted according to the calculated water pressure conditions to allow excess water to flow into the second water reservoir, thereby adjusting the rotation speed of the turbine, so that the turbine can operate at the optimal rotation speed, thereby improving the efficiency of power generation.

[0062] 3. When the power generation power drops to a threshold value, the present invention can switch the gate to the second water reservoir, make the speed of the excitation motor the same as the test speed, and judge the abnormal state data according to the recovery speed of the power generation efficiency, thereby improving the reliability and stability of the equipment, improving the safety and operation and maintenance efficiency of the equipment, and ensuring the smooth operation of the hydropower station. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0064] Figure 1 It is a structural schematic diagram of a hydropower station data monitoring system based on artificial intelligence of the present invention;

[0065] Figure 2 It is a schematic diagram of the steps of a hydropower station data monitoring method based on artificial intelligence of the present invention. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] See also Figure 1 ,The present invention provides a technical solution: a hydropower station data monitoring system based on artificial intelligence, comprising: a variable sensing module, an intelligent network module, an additional water tank module, an input addressing module and a water pressure recovery module;

[0068] The variable sensing module is used to set sensors in the reservoir and the turbine motor of the hydropower station. The sensors read the hydrological data in the reservoir and the power generation data of the turbine motor, upload the data to the control center, and update the hydrological data in the reservoir in real time after the hydropower station is working;

[0069] The variable sensing module includes: a hydrological sensing unit and a motor sensing unit;

[0070] The hydrological sensing unit is used to detect hydrological data in the reservoir, and the hydrological data includes: water storage capacity, water level difference, gate opening and water storage loss rate;

[0071] The motor sensor unit is used to detect the power generation data of the water turbine motor, and the power generation data includes: the rotation speed of the excitation motor, the valve pressure of the unit, the guide vane opening, the oil temperature and the shoe surface temperature.

[0072] The intelligent network module is used to use hydrological data as the first-order input variable, the excitation motor speed as the first-order output variable, the power generation data as the second-order input variable, and the power generation power of the turbine as the second-order output variable, open the reservoir gate for trial operation, use the power generation data and hydrological data detected during the trial operation to train the state classification network, output the state function of the first-order variable and the second-order variable, obtain the time domain attenuation speed from the coefficient matrix of each variable in the state function, and remove the sensor in the turbine motor after the trial operation is completed;

[0073] The intelligent network module includes: an operation test unit, a state identification unit and a variable attenuation unit;

[0074] The operation test unit is used to open the gate of the water storage tank to test the motor, record the power generation data of the motor during the test run, and close the gate after the test run is completed;

[0075] The state recognition unit is used to train the neural network model using the power generation data during the trial operation, and to fit the state function of the input and output variables using the trained model;

[0076] The variable attenuation unit is used to obtain the coefficient matrix in the state function, and obtain the time domain attenuation speed of each variable from the coefficient matrix.

[0077] The additional water tank module is used to establish a second water tank on the basis of the original water tank of the hydropower station, and the hydrological data corresponding to the highest rate of change of the first-order state function is used as the hydrological characteristics of the second water tank. The first and second water tanks are connected to the turbine by using a two-way gate, and the control component for controlling the opening and closing of the gate is connected to the control center. The gate of the second water tank is initially in a closed state;

[0078] The additional water tank module includes: a water tank unit and a valve control unit;

[0079] The water reservoir unit is used to divide an independent second water reservoir in the hydropower station, and the second water reservoir plays the role of diverting, testing and buffering the power generation process of the hydropower station;

[0080] The valve control unit is used to connect the first and second water storage tanks and the turbine using a controllable two-way valve to adjust the water flow pressure received by the turbine.

[0081] The input addressing module is used to input the hydrological data in the first water reservoir into the state function, output the theoretical power generation of the turbine, and when the difference between the actual power generation of the turbine and the theoretical power generation exceeds the threshold, switch the gate to the second water reservoir, stabilize the hydrological data of the second water reservoir, keep the speed of the excitation motor constant, record the decay rate of the generator power generation, compare the decay rate with the time domain decay rate of each variable, and output the closest variable as the abnormal variable;

[0082] The input addressing module includes: a pattern recognition unit, a variable control unit and an abnormality locating unit;

[0083] The pattern recognition unit is used to output the excitation motor speed through the first-order state function, output the turbine theoretical power through the second-order state function, and identify abnormal power generation conditions;

[0084] The variable control unit is used to stabilize the speed of the excitation motor of the turbine by using the second water reservoir;

[0085] The abnormality locating unit is used to obtain the attenuation rate of the turbine power according to the time domain variation of the actual power generation of the turbine, compare it with the time domain attenuation rate of each variable, and locate the abnormal variable.

[0086] The water pressure recovery module is used to calculate the theoretical value of the abnormal variable from the state function and the actual power generation, reconstruct the second-order state function with the theoretical value of the abnormal variable as a constant, calculate the speed of the excitation motor when the power generation is maximum according to the reconstructed equation, and control the two-way valve so that the second water reservoir can meet the hydrological data required when the speed of the excitation motor is stable.

[0087] The water pressure recovery module includes: a state reconstruction unit and a hydraulic communication unit;

[0088] The state reconstruction unit is used to reconstruct the second-order state function with the theoretical value of the abnormal variable as a constant to calculate the maximum generating speed of the excitation motor;

[0089] The hydraulic communication unit is used to establish hydrological data corresponding to the maximum power generation speed in the second water storage tank until the power generation stage ends.

[0090] like Figure 2 As shown, a hydropower station data monitoring method based on artificial intelligence includes the following steps:

[0091] Step S1. After the hydropower station is installed, a test run is carried out to detect the hydrological data of the reservoir and the power generation data of the turbine during the test run, and the neural network model is trained using the historical hydrological data and power generation data to obtain a trained state classification model;

[0092] Step S1 includes:

[0093] Step S11. The hydropower station is put into trial operation. Sensors are set in the reservoir and the turbine motor to detect the hydrological data in the reservoir and the power generation data of the turbine motor at the same frequency. The hydrological data include: water storage capacity, water level difference, gate opening and water storage loss rate. The power generation data include: excitation motor speed, unit valve pressure, guide vane opening, oil temperature and tile surface temperature.

[0094] Step S12. extracting historical hydrological data and power generation data from the database, wherein the historical hydrological data and power generation data are obtained from data recorded during the historical period of installed capacity operation of the hydropower station;

[0095] Step S13. Use historical hydrological data and power generation data to train a neural network model, and use the neural network model to fit the correlation between data pieces. The trained neural network model is output as a state classification model. The selected neural network models include: convolutional neural network, perceptron network, long short-term memory network and autoencoder network.

[0096] Step S2. Taking the hydrological data and the excitation motor speed as the first-order input and output variables, and the power generation data and the power generation power as the second-order input and output variables, the hydrological data and the power generation data detected during the trial operation are input into the state classification model to obtain the first-order and second-order state functions, and the time domain attenuation speed of each power generation data is obtained from the coefficient matrix of the second-order state function;

[0097] Step S2 includes:

[0098] Step S21. Take the hydrological data as the first-order input variable and the excitation motor speed as the first-order output variable, and fit the state function of the input variable and the output variable through the state classification model to obtain the first-order state function, which is expressed as:

[0099]

[0100] Where D(x1,x2,…,xn) represents the first-order state function, n represents the number of first-order input variables, x1, x2,…xn represent the 1st to nth first-order input variables respectively, X1 T To Xn TRespectively represent the transposed matrices of the detection value matrices of the 1st to nth variables during the trial run, A1 to An represent the coefficient matrices of the 1st to nth first-order input variables, c∈{1,2,…,n}, m represents the number of detections during the trial run, X c (1) to X c (m) represents the detection values ​​of the cth input variable from the 1st to the mth time respectively;

[0101] Step S22. Use the power generation data as the second-order input variable and the power generation as the second-order output variable, and output the second-order state function G(r1, r2, ..., rn) from the state classification model, where r1 to rn represent the 1st to nth second-order input variables respectively;

[0102] Step S23. Calculate the time domain decay rate function F(t) corresponding to each power generation data, where F(t) = [B flor(t / t0) +B flor(t / t0)+1 ]·mod(t / t0) / t0, where t represents time, t0 represents the sampling interval during the trial run, flor is the floor function, mod is the remainder function, B flor(t / t0) and B f l or(t / t0)+1 They respectively represent the values ​​of the flor(t / t0)th and flor(t / t0)+1th elements in the coefficient matrix of the second-order state function.

[0103] Step S3. A second reservoir is established on the basis of the original first reservoir of the hydropower station, the water level height when the output of the first-order state function changes the most is used as the water level height of the second reservoir, and a two-way gate is used to connect the first and second reservoirs to the turbine;

[0104] Step S3 includes:

[0105] Step S31. Establish a second reservoir in the hydropower station, the second reservoir plays the role of diverting, testing and buffering the power generation process of the hydropower station, calculate the partial derivative of the first-order state function D(x1, x2, ..., xn) with respect to the water level height variable, and the water level height when the partial derivative function takes the maximum value is used as the water level height of the second reservoir;

[0106] Step S32. Use a two-way gate to connect the first and second water reservoirs and the turbine, so that the water level in the second water reservoir can be adjusted through the two-way gate connected to the first water reservoir.

[0107] Step S4. After the turbine is working, the hydrological data in the first reservoir is input into the first and second state functions in real time, and the theoretical power generation of the turbine is output. When the difference between the actual power generation of the turbine and the theoretical power generation exceeds a threshold, the two-way gate is switched to the second reservoir, and the water level of the second reservoir is stabilized;

[0108] Step S4 includes:

[0109] Step S41. The turbine is connected to the first water reservoir to work, and the hydrological data in the first water reservoir is obtained. The hydrological data is input into the first-order state function to obtain the theoretical value of the excitation motor speed. The excitation motor speed and the initial sampling values ​​of the remaining power generation data during the trial operation are input into the second-order state function to obtain the theoretical value of the generator power generation;

[0110] Step S42. When |P1-P2|≥P0, it is determined that there is an abnormality in power generation, and the process goes to step S43, where P1 represents the theoretical value of power generation, P2 represents the actual power generation, and P0 represents the preset threshold value;

[0111] Step S43. Switch the gate leading to the abnormal turbine from the first water reservoir to the second water reservoir, and at the same time control the connecting gates in the first and second water reservoirs to keep the water level in the second water reservoir constant.

[0112] Step S5. Record the decay rate of the generator power, compare the decay rate with the time domain decay rate of each variable, output the closest variable as the abnormal variable, and adjust the water level of the second reservoir to maximize the power generation.

[0113] Step S5 includes:

[0114] Step S51. Adjust the hydrological parameters of the second reservoir to keep them consistent with the changes in the hydrological parameters of the reservoir during the trial operation, continuously detect the power generation of the turbine, and output the power generation change function P(t);

[0115] Step S52: Calculate the correlation between the time domain attenuation of the power generation change function and the time domain attenuation of the variable:

[0116]

[0117] Among them, Q represents the correlation between the generated power and the time-domain attenuation of the variable, and T represents the time from the hydrological stabilization of the second reservoir to the present;

[0118] Step S53. Execute step S52 for each second-order variable, wherein the second-order variable corresponding to the minimum value of Q is output as the abnormal variable of the hydro-motor;

[0119] Step S54. Adjust the water level of the reservoir, detect the change in the generated power, record the water level when the generated power reaches the maximum value, and before the turbine is shut down for maintenance, adjust the water level of the second reservoir to be the same as the recorded water level.

[0120] Example: The current water level of the first reservoir of the hydropower station is 50m, the water level during the trial operation is 20m to 10m, the actual power of the turbine motor is detected to be 150kW, and the flow rate of the first reservoir is 10m 2 / s, calculated according to the first-order state function, the expected power is 250kW. If the difference exceeds the threshold, the second water reservoir is switched to adjust the water level and flow rate of the second water reservoir to the same as those in the trial operation. The correlations between the change function P(t) of the generated power and the valve pressure, guide vane opening, oil temperature and bearing surface temperature of the unit are detected to be 0.2, 0.18, 2.0 and 1.5 respectively, and the valve pressure of the unit is output as an abnormal variable.

[0121] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0122] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A hydropower station data monitoring method based on artificial intelligence, characterized in that: The method comprises the following steps: Step S1. After the hydropower station is installed, a test run is carried out to detect the hydrological data of the reservoir and the power generation data of the turbine during the test run, and the neural network model is trained using the historical hydrological data and power generation data to obtain a trained state classification model; Step S2. Taking the hydrological data and the excitation motor speed as the first-order input and output variables, and the power generation data and the power generation power as the second-order input and output variables, the hydrological data and the power generation data detected during the trial operation are input into the state classification model to obtain the first-order and second-order state functions, and the time domain attenuation speed of each power generation data is obtained from the coefficient matrix of the second-order state function; Step S3. A second reservoir is established on the basis of the original first reservoir of the hydropower station, the water level height when the output of the first-order state function changes the most is used as the water level height of the second reservoir, and a two-way gate is used to connect the first and second reservoirs to the turbine; Step S4. After the turbine is working, the hydrological data in the first reservoir is input into the first and second state functions in real time, and the theoretical power generation of the turbine is output. When the difference between the actual power generation of the turbine and the theoretical power generation exceeds a threshold, the two-way gate is switched to the second reservoir, and the water level of the second reservoir is stabilized; Step S5. Record the decay rate of the generator power, compare the decay rate with the time domain decay rate of each variable, output the closest variable as the abnormal variable, and adjust the water level of the second reservoir to maximize the power generation.

2. The method for monitoring hydropower station data based on artificial intelligence according to claim 1, characterized in that: Step S1 includes: Step S11. The hydropower station is put into trial operation. Sensors are set in the reservoir and the turbine motor to detect the hydrological data in the reservoir and the power generation data of the turbine motor at the same frequency. The hydrological data include: water storage capacity, water level difference, gate opening and water storage loss rate. The power generation data include: excitation motor speed, unit valve pressure, guide vane opening, oil temperature and tile surface temperature. Step S12. extracting historical hydrological data and power generation data from the database, wherein the historical hydrological data and power generation data are obtained from data recorded during the historical period of installed capacity operation of the hydropower station; Step S13. Use historical hydrological data and power generation data to train a neural network model, and use the neural network model to fit the correlation between data pieces. The trained neural network model is output as a state classification model. The selected neural network models include: convolutional neural network, perceptron network, long short-term memory network and autoencoder network.

3. The method for monitoring hydropower station data based on artificial intelligence according to claim 2 is characterized in that: Step S2 includes: Step S21. Take the hydrological data as the first-order input variable and the excitation motor speed as the first-order output variable, and fit the state function of the input variable and the output variable through the state classification model to obtain the first-order state function, which is expressed as: Where D(x1,x2,…,xn) represents the first-order state function, n represents the number of first-order input variables, x1, x2,…xn represent the 1st to nth first-order input variables respectively, X1 T To Xn T Respectively represent the transposed matrices of the detection value matrices of the 1st to nth variables during the trial run, A1 to An represent the coefficient matrices of the 1st to nth first-order input variables, c∈{1,2,…,n}, m represents the number of detections during the trial run, X c (1) to X c (m) represents the detection values ​​of the cth input variable from the 1st to the mth time respectively; Step S22. Use the power generation data as the second-order input variable and the power generation as the second-order output variable, and output the second-order state function G(r1, r2, ..., rn) from the state classification model, where r1 to rn represent the 1st to nth second-order input variables respectively; Step S23. Calculate the time domain decay rate function F(t) corresponding to each power generation data. F(t)=[B flor(t / t0) +B flor(t / t0)+1 ]·mod(t / t0) / t0, where t represents time, t0 represents the sampling interval during the trial run, flor is the floor function, mod is the remainder function, B flor(t / t0) and B flor(t / t0)+1 They respectively represent the values ​​of the flor(t / t0)th and flor(t / t0)+1th elements in the coefficient matrix of the second-order state function.

4. The method for monitoring hydropower station data based on artificial intelligence according to claim 3 is characterized in that: Step S3 includes: Step S31. Establish a second reservoir in the hydropower station, the second reservoir plays the role of diverting, testing and buffering the power generation process of the hydropower station, calculate the partial derivative of the first-order state function D(x1, x2, ..., xn) with respect to the water level height variable, and the water level height when the partial derivative function takes the maximum value is used as the water level height of the second reservoir; Step S32. Use a two-way gate to connect the first and second water reservoirs and the turbine, so that the water level in the second water reservoir can be adjusted by connecting the two-way gate between the first water reservoir; Step S4 includes: Step S41. The turbine is connected to the first water reservoir to work, the hydrological data in the first water reservoir is obtained, the hydrological data is input into the first-order state function, the theoretical value of the excitation motor speed is obtained, the excitation motor speed and the initial sampling values ​​of the remaining power generation data during the trial operation are input into the second-order state function, and the theoretical value of the generator power generation is obtained; Step S42. When |P1-P2|≥P0, it is determined that there is an abnormality in power generation, and the process goes to step S43, where P1 represents the theoretical value of power generation, P2 represents the actual power generation, and P0 represents the preset threshold value; Step S43. Switch the gate leading to the abnormal turbine from the first water reservoir to the second water reservoir, and at the same time control the connecting gates in the first and second water reservoirs to keep the water level in the second water reservoir constant.

5. The method for monitoring hydropower station data based on artificial intelligence according to claim 4 is characterized in that: Step S5 includes: Step S51. Adjust the hydrological parameters of the second reservoir to keep them consistent with the changes in the hydrological parameters of the reservoir during the trial operation, continuously detect the power generation of the turbine, and output the power generation change function P(t); Step S52: Calculate the correlation between the time domain attenuation of the power generation change function and the time domain attenuation of the variable: Among them, Q represents the correlation between the generated power and the time-domain attenuation of the variable, and T represents the time from the hydrological stabilization of the second reservoir to the present; Step S53. Execute step S52 for each second-order variable, wherein the second-order variable corresponding to the minimum value of Q is output as the abnormal variable of the hydro-motor; Step S54. Adjust the water level of the reservoir, detect the change in the generated power, record the water level when the generated power reaches the maximum value, and before the turbine is shut down for maintenance, adjust the water level of the second reservoir to be the same as the recorded water level.

6. A hydropower station data monitoring system based on artificial intelligence, characterized in that: The system includes the following modules: variable sensing module, intelligent network module, additional water tank module, input addressing module and water pressure recovery module; The variable sensing module is used to set sensors in the reservoir and the turbine motor of the hydropower station. The sensors read the hydrological data in the reservoir and the power generation data of the turbine motor, upload the data to the control center, and update the hydrological data in the reservoir in real time after the hydropower station is working; The intelligent network module is used to use hydrological data as the first-order input variable, the excitation motor speed as the first-order output variable, the power generation data as the second-order input variable, and the power generation power of the turbine as the second-order output variable, open the reservoir gate for trial operation, use the power generation data and hydrological data detected during the trial operation to train the state classification network, output the state function of the first-order variable and the second-order variable, obtain the time domain attenuation speed from the coefficient matrix of each variable in the state function, and remove the sensor in the turbine motor after the trial operation is completed; The additional water tank module is used to establish a second water tank on the basis of the original water tank of the hydropower station, and the hydrological data corresponding to the highest rate of change of the first-order state function is used as the hydrological characteristics of the second water tank. The first and second water tanks are connected to the turbine by using a two-way gate, and the control component for controlling the opening and closing of the gate is connected to the control center. The gate of the second water tank is initially in a closed state; The input addressing module is used to input the hydrological data in the first water reservoir into the state function, output the theoretical power generation of the turbine, and when the difference between the actual power generation of the turbine and the theoretical power generation exceeds the threshold, switch the gate to the second water reservoir, stabilize the hydrological data of the second water reservoir, keep the speed of the excitation motor constant, record the decay rate of the generator power generation, compare the decay rate with the time domain decay rate of each variable, and output the closest variable as the abnormal variable; The water pressure recovery module is used to calculate the theoretical value of the abnormal variable from the state function and the actual power generation, reconstruct the second-order state function with the theoretical value of the abnormal variable as a constant, calculate the speed of the excitation motor when the power generation is maximum according to the reconstructed equation, and control the two-way valve so that the second water reservoir can meet the hydrological data required when the speed of the excitation motor is stable.

7. The artificial intelligence-based hydropower station data monitoring system according to claim 6 is characterized by: The variable sensing module includes: a hydrological sensing unit and a motor sensing unit; The hydrological sensing unit is used to detect hydrological data in the reservoir, and the hydrological data includes: water storage capacity, water level difference, gate opening and water storage loss rate; The motor sensor unit is used to detect the power generation data of the water turbine motor, and the power generation data includes: the rotation speed of the excitation motor, the valve pressure of the unit, the guide vane opening, the oil temperature and the shoe surface temperature.

8. The artificial intelligence-based hydropower station data monitoring system according to claim 7 is characterized by: The intelligent network module includes: an operation test unit, a state identification unit and a variable attenuation unit; The operation test unit is used to open the gate of the water storage tank to test the motor, record the power generation data of the motor during the test run, and close the gate after the test run is completed; The state recognition unit is used to train the neural network model using the power generation data during the trial operation, and to fit the state function of the input and output variables using the trained model; The variable attenuation unit is used to obtain the coefficient matrix in the state function, and obtain the time domain attenuation speed of each variable from the coefficient matrix.

9. The artificial intelligence-based hydropower station data monitoring system according to claim 8, characterized in that: The additional water tank module includes: a water tank unit and a valve control unit; The water reservoir unit is used to divide an independent second water reservoir in the hydropower station, and the second water reservoir plays the role of diverting, testing and buffering the power generation process of the hydropower station; The valve control unit is used to connect the first and second water storage tanks and the turbine using a controllable two-way valve to adjust the water flow pressure received by the turbine.

10. The artificial intelligence-based hydropower station data monitoring system according to claim 9, characterized in that: The input addressing module includes: a pattern recognition unit, a variable control unit and an abnormality locating unit; The pattern recognition unit is used to output the excitation motor speed through the first-order state function, output the turbine theoretical power through the second-order state function, and identify abnormal power generation conditions; The variable control unit is used to stabilize the speed of the excitation motor of the turbine by using the second water reservoir; The abnormality locating unit is used to obtain the attenuation rate of the turbine power according to the time domain change of the actual power generation of the turbine, compare it with the time domain attenuation rate of each variable, and locate the abnormal variable; The water pressure recovery module includes: a state reconstruction unit and a hydraulic communication unit; The state reconstruction unit is used to reconstruct the second-order state function with the theoretical value of the abnormal variable as a constant to calculate the maximum generating speed of the excitation motor; The hydraulic communication unit is used to establish hydrological data corresponding to the maximum power generation speed in the second water storage tank until the power generation stage ends.

Images