A data analysis method and system for the turning gear test results of a hydropower station

Abstract

The present invention discloses a data analysis method and system for the turning gear result of a hydropower station, which relates to the field of computer data analysis. The method includes obtaining the state data of the turning gear within a preset time and the real-time state detection information of the turning gear; performing data parsing on the state data of the turning gear within the preset time to obtain a first data set, a second data set, and a third data set; solving the first data set and the third data set through a preset first association model to obtain a first processed data set; solving the second data set and the third data set through a preset second association model to obtain a second processed data set; fusing the first processed data set and the second processed data set; and inputting the real-time state detection information of the turning gear into the turning gear state recognition model to obtain the adjustment feature information of the turning gear. This method establishes the data correlation of each part of the unit, reads the relevant data of each part of the unit during turning gear in real time, and achieves the purpose of guiding maintenance and adjustment.

Description

Technical Field

[0001] The present invention belongs to the field of computer data analysis, and particularly relates to a data analysis method and system for the turning test results of a hydropower station. Background Art

[0002] In a hydropower station, the turning test results reflect the swing of the axis and the corresponding state of the unit axis of the main rotating components such as the turbine runner, the main shaft of the hydrogenerator, and the generator rotor after operation. Maintenance personnel perform maintenance adjustments based on the analysis of the turning test results to ensure that the geometric center, rotation center, and center of gravity of the rotating part of the unit are on the same axis, ensuring the safe and stable operation of the unit. In the existing data analysis of the turning test results, no relationships are established among the swing, level, and air gap data of each part of the unit, resulting in the inability to provide effective maintenance guidance based on the numerical changes of multiple parameters, and the overall data analysis is time-consuming and labor-intensive. Therefore, it is necessary to design a data analysis method for the turning test results of a hydropower station to solve the above problems. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a data analysis method and system for the turning test results of a hydropower station. This solution is used to solve the problems of incomplete data analysis of the turning test results in the prior art, inability to establish relationships between data, and time-consuming and labor-intensive. It has the characteristics of being able to establish data correlations for each part of the unit to achieve real-time reading of relevant data for each part of the unit during the turning test, and achieving the purpose of guiding maintenance adjustments.

[0004] To achieve the above technical effects, the technical solution adopted by the present invention is:

[0005] A data analysis method for the turning test results of a hydropower station, comprising:

[0006] Obtaining the state data of the turning test within a preset time and the real-time state detection information of the turning test;

[0007] Performing data parsing on the state data of the turning test within a preset time to obtain an analysis data set of the turning test results. The analysis data set of the turning test results includes a first data set, a second data set, and a third data set. The first data set includes the roundness information of the rotor end face and the roundness information of the stator end face. The second data set includes the displacement information of the large shaft air supply pipe, the displacement information of the slip ring, and the displacement information of the guide water component. The third data set is the level information of the mirror plate;

[0008] Solving the first data set and the third data set through a preset first correlation model to obtain a first processed data set;

[0009] Solving the second data set and the third data set through a preset second correlation model to obtain a second processed data set;

[0010] Solve the first processed data set and the second processed data set through a preset multi-source data fusion model to obtain the fusion data of the turning gear result;

[0011] Input the state data of the turning gear within a preset time and the fusion data of the turning gear result into a preset recurrent neural network for training to obtain a turning gear state recognition model;

[0012] Input the real-time state detection information of the turning gear into the turning gear state recognition model to obtain the adjustment feature information of the turning gear, and the adjustment feature information of the turning gear is used to perform maintenance adjustment on the turning gear.

[0013] Preferably, solve the first data set and the third data set through a preset first correlation model to obtain the first processed data set. The rotor end face roundness information includes the rotor upper end face roundness information and the rotor lower end face roundness information, and the stator end face roundness information includes the stator upper end face roundness information and the stator lower end face roundness information, including:

[0014] Obtain the rotor roundness weight information and the stator roundness weight information of the current hydropower station;

[0015] Construct a matrix according to the rotor upper end face roundness information, the rotor lower end face roundness information, the stator upper end face roundness information, the stator lower end face roundness information, and the level information of the mirror plate within a preset time to obtain the first-level multi-parameter matrix information;

[0016] Solve the first-level multi-parameter matrix information through an inner product model to obtain the first-level multi-parameter inner product matrix information;

[0017] Solve the first-level multi-parameter inner product matrix information through a feature matrix model to obtain the first-level eigenvector matrix information and the first-level diagonal matrix information;

[0018] Solve the rotor roundness weight information, the stator roundness weight information, the first-level eigenvector matrix information, and the first-level diagonal matrix information through a first correlation model to obtain the first processed data set.

[0019] Preferably, solve the rotor roundness weight information, the stator roundness weight information, the first-level eigenvector matrix information, and the first-level diagonal matrix information through a first correlation model to obtain the first processed data set, where the first correlation model is:

[0020] ;

[0021] In the formula, represents the first processed data set, Indicates the rotor roundness weight information, Indicates the stator roundness weight information, Indicates the diagonal matrix when taking the largest eigenvalues from the first-level diagonal matrix information, Indicates the eigenvector matrix when taking the largest eigenvalues from the first-level eigenvector matrix information, Indicates the first preset constant coefficient.

[0022] Preferably, the second data set and the third data set are solved through a preset second correlation model to obtain a second processed data set, including:

[0023] Obtain the displacement weight information of the large shaft air supply pipe, the slip ring displacement weight information, and the guide vane assembly displacement weight information of the current hydropower station;

[0024] Construct a matrix based on the displacement information of the large shaft air supply pipe, the slip ring displacement information, the guide vane assembly displacement information, and the horizontal information of the mirror plate within a preset time to obtain the second-level multi-parameter matrix information;

[0025] Solve the second-level multi-parameter matrix information through an inner product model to obtain the second-level multi-parameter inner product matrix information;

[0026] Solve the second-level multi-parameter inner product matrix information through an eigenmatrix model to obtain the second-level eigenvector matrix information and the second-level diagonal matrix information;

[0027] Solve the displacement weight information of the large shaft air supply pipe, the slip ring displacement weight information, the guide vane assembly displacement weight information, the second-level eigenvector matrix information, and the second-level diagonal matrix information through the second correlation model to obtain the second processed data set.

[0028] Preferably, the displacement weight information of the large shaft air supply pipe, the slip ring displacement weight information, the guide vane assembly displacement weight information, the second-level eigenvector matrix information, and the second-level diagonal matrix information are solved through the second correlation model to obtain the second processed data set, where the second correlation model is:

[0029] ;

[0030] In the formula, Indicates the second processed data set, Indicates the displacement weight information of the large shaft air supply pipe, Indicates the slip ring displacement weight information, Indicates the guide vane assembly displacement weight information, Indicates taking from the second-level diagonal matrix information The diagonal matrix when there are the largest eigenvalues, and the eigenvector matrix when there are the largest eigenvalues in the second-level eigenvector matrix information,

[0031] Preferably, when obtaining the displacement weight information of the water guide component, the calculation of the displacement weight information of the water guide component includes:

[0032] obtaining the upper guide displacement weight information, the lower guide displacement weight information, the water guide displacement weight information, the thrust bearing head displacement weight information, and the lower flange displacement information of the rotor;

[0033] calculating the upper guide displacement weight information and the lower guide displacement weight information to obtain a first judgment coefficient;

[0034] calculating the water guide displacement weight information, the thrust bearing head displacement weight information, and the lower flange displacement information of the rotor to obtain a second judgment coefficient;

[0035] solving the first judgment coefficient and the second judgment coefficient through a preset weight model to obtain the displacement weight information of the water guide component.

[0036] Preferably, the state data of the turning gear within a preset time and the fusion data of the turning gear result are input into a preset recurrent neural network for training to obtain a turning gear state recognition model, including:

[0037] performing normalization calculation on the state data of the turning gear within a preset time to obtain the normalized data of the turning gear state;

[0038] performing normalization calculation on the fusion data of the turning gear result to obtain the normalized data of the turning gear result;

[0039] inputting the normalized data of the turning gear state and the normalized data of the turning gear result into a preset recurrent neural network for training to obtain a turning gear state recognition model. In the recurrent neural network, there is an input layer, multiple hidden layers, and an output layer. Each neuron in the hidden layer is a recurrent unit. The input layer is fully connected to the first hidden layer, the last hidden layer is fully connected to the output layer, and a memory unit is arranged between the output layer and the input layer.

[0040] Preferably, a data analysis system for the turning gear result of a hydropower station includes:

[0041] an acquisition module for acquiring the state data of the turning gear within a preset time and the real-time state detection information of the turning gear;

[0042] The first processing module is used to parse the state data of the barring gear within a preset time to obtain an analysis data set of the barring gear result. The analysis data set of the barring gear result includes a first data set, a second data set, and a third data set. The first data set includes rotor end face roundness information and stator end face roundness information. The second data set includes large shaft gas supply pipe displacement information, slip ring displacement information, and water guide assembly displacement information. The third data set is the horizontal information of the mirror plate;

[0043] The second processing module is used to solve the first data set and the third data set through a preset first association model to obtain a first processed data set;

[0044] The third processing module is used to solve the second data set and the third data set through a preset second association model to obtain a second processed data set;

[0045] The fourth processing module is used to solve the first processed data set and the second processed data set through a preset multi-source data fusion model to obtain a fusion data of the barring gear result;

[0046] The fifth processing module is used to input the state data of the barring gear within a preset time and the fusion data of the barring gear result into a preset recurrent neural network for training to obtain a barring gear state recognition model;

[0047] The recognition module is used to input the real-time state detection information of the barring gear into the barring gear state recognition model to obtain adjustment feature information of the barring gear. The adjustment feature information of the barring gear is used to perform maintenance adjustment on the barring gear.

[0048] Preferably, in the second processing module, the rotor end face roundness information includes rotor upper end face roundness information and rotor lower end face roundness information, and the stator end face roundness information includes stator upper end face roundness information and stator lower end face roundness information, including:

[0049] The first acquisition unit is used to acquire the rotor roundness weight information and stator roundness weight information of the current hydropower station;

[0050] The first calculation unit is used to construct a matrix according to the rotor upper end face roundness information, the rotor lower end face roundness information, the stator upper end face roundness information, the stator lower end face roundness information, and the horizontal information of the mirror plate within a preset time to obtain first-level multi-parameter matrix information;

[0051] The second calculation unit is used to solve the first-level multi-parameter matrix information through an inner product model to obtain first-level multi-parameter inner product matrix information;

[0052] A third computing unit is configured to solve the first-level multi-parameter inner product matrix information through a feature matrix model to obtain first-level eigenvector matrix information and first-level diagonal matrix information;

[0053] A fourth computing unit is configured to solve the rotor roundness weight information, the stator roundness weight information, the first-level eigenvector matrix information, and the first-level diagonal matrix information through a first correlation model to obtain a first processed data set.

[0054] Preferably, the third processing module includes:

[0055] A second acquisition unit is configured to acquire the displacement weight information of the large shaft air supply pipe, the slip ring displacement weight information, and the guide vane assembly displacement weight information of the current hydropower station;

[0056] A fifth computing unit is configured to construct a matrix based on the displacement information of the large shaft air supply pipe, the slip ring displacement information, the guide vane assembly displacement information, and the horizontal information of the mirror plate within a preset time to obtain second-level multi-parameter matrix information;

[0057] A sixth computing unit is configured to solve the second-level multi-parameter matrix information through an inner product model to obtain second-level multi-parameter inner product matrix information;

[0058] A seventh computing unit is configured to solve the second-level multi-parameter inner product matrix information through a feature matrix model to obtain second-level eigenvector matrix information and second-level diagonal matrix information;

[0059] An eighth computing unit is configured to solve the displacement weight information of the large shaft air supply pipe, the slip ring displacement weight information, the guide vane assembly displacement weight information, the second-level eigenvector matrix information, and the second-level diagonal matrix information through a second correlation model to obtain a second processed data set.

[0060] The beneficial effects of the present invention are as follows:

[0061] The present invention first analyzes and classifies the data of each part of the unit, then correlates the first data set with the third data set, and correlates the second data set with the third data set; then fuses the processed data to ensure the redundancy or complementarity of different processed data in space or time to obtain a key analysis of the priority items of the turning result; finally, inputs the real-time state detection information of the turning into the turning state recognition model to obtain the adjustment feature information of the turning, so as to achieve the purpose of guiding the overhaul adjustment. This method establishes the data correlation of each part of the unit, realizes the real-time reading of the relevant data of each part of the unit during turning, and automatically completes the data analysis and the auxiliary decision-making of the axis adjustment by using the analysis algorithm, so that the whole turning is visualized and simplified, and the goal of saving the project overhaul cost and shortening the project duration is achieved. Brief Description of the Drawings

[0062] Figure 1 It is a schematic flow diagram of the data analysis method for the turning result of a hydropower station in the embodiment of the present invention;

[0063] Figure 2 It is a topological structure diagram when the data analysis system for the turning result of a hydropower station in the embodiment of the present invention is applied;

[0064] Figure 3 It is a schematic diagram of the data calculation principle in the embodiment of the present invention;

[0065] Figure 4 It is a schematic structural diagram of the data analysis device for the turning result of a hydropower station in the embodiment of the present invention;

[0066] In the figure: The data analysis device 800 for the turning result of a hydropower station, a processor 801, a memory 802, a multimedia component 803, an I / O interface 804, and a communication component 805. Detailed Embodiments

[0067] Embodiment 1:

[0068] As Figure 1 shown, this embodiment provides a data analysis method for the turning result of a hydropower station; Figure 1 It is shown in that this method includes steps S1 to S7, specifically:

[0069] S1: Obtain the state data of the turning within a preset time and the real-time state detection information of the turning;

[0070] In step S1, the turning form can be 8-point fixed-point turning or continuous turning, and the preset time is determined according to the training amount of the preset recurrent neural network;

[0071] In terms of data acquisition, for displacement information, it is collected by setting a contact displacement sensor; for roundness information, a capacitance sensor is set for data acquisition, and it can adapt to units with different gaps through different magnetic backplates; for the main shaft rotation signal, a position sensor is set for acquisition to facilitate the calculation of the turning angle; for the horizontal position information, an electronic level image combining instrument is used for measurement.

[0072] Furthermore, the capacitance sensor can adopt the PC25 series sensor. The PC25 series sensor is designed based on the capacitance principle and can measure the distance between its surface and the measured object. The measurement signal is sent from the sensor end to the PC25 series signal conditioning module through a high-temperature resistant three-layer shielded coaxial cable, and after being corrected by a precise digital linearization algorithm, it is sent to the calculation module.

[0073] S2: Parse the status data of the barring gear within a preset time to obtain a parsed data set of the barring gear result. The parsed data set of the barring gear result includes a first data set, a second data set, and a third data set. The first data set includes rotor end face roundness information and stator end face roundness information. The second data set includes large shaft gas supply pipe displacement information, slip ring displacement information, and water guide assembly displacement information. The third data set is the horizontal information of the mirror plate;

[0074] Among them, the water guide assembly displacement information includes upper guide displacement information, lower guide displacement information, water guide displacement information, thrust head displacement information, and rotor lower flange displacement information.

[0075] When performing data parsing and transmission in this method, the gateway is set as follows:

[0076] ①. Data interface: Pulse is set to collect 2 channels, with a rate of 1 kHz and an input voltage of 0 - 10V; 485 communication collects 6 channels, with a maximum rate of 9600 and optoelectronic isolation;

[0077] ②. Wireless communication: Lora mode.

[0078] S3: Solve the first data set and the third data set through a preset first correlation model to obtain a first processed data set;

[0079] To clarify the specific processing process of the first processed data set, the rotor end face roundness information includes rotor upper end face roundness information and rotor lower end face roundness information, and the stator end face roundness information includes stator upper end face roundness information and stator lower end face roundness information. Step S3 includes S31 to S35, specifically:

[0080] S31: Obtain the rotor roundness weight information and stator roundness weight information of the current hydropower station;

[0081] In step S31, the rotor roundness weight information is determined according to the maintenance standard of the current hydropower station. First, obtain the tolerance range of the rotor roundness; then calculate the average value of the end point values of the tolerance range to obtain the average value information of the rotor roundness; finally, calculate the ratio of the average value information of the rotor roundness to a preset roundness threshold to obtain the rotor roundness weight information;

[0082] The stator roundness weight information is determined according to the tolerance range of the rotational fit between the rotor and the stator in the current hydropower station. Specifically: Subtract the end point values of the rotational fit tolerance range, and then multiply the difference value by the rotor roundness weight information to obtain the stator roundness weight information.

[0083] S32: Construct a matrix based on the roundness information of the upper end face of the rotor, the roundness information of the lower end face of the rotor, the roundness information of the upper end face of the stator, the roundness information of the lower end face of the stator, and the level information of the mirror plate within a preset time to obtain the first-level multi-parameter matrix information;

[0084] In step S32, the first-level multi-parameter matrix information is:

[0085] ; (1)

[0086] In the above formula (1), represents the first-level multi-parameter matrix information, represents the characteristic data corresponding to the roundness information of the upper end face of the rotor, represents the characteristic data corresponding to the roundness information of the lower end face of the rotor, represents the characteristic data corresponding to the roundness information of the upper end face of the stator, represents the characteristic data corresponding to the roundness information of the lower end face of the stator, represents the characteristic data corresponding to the level information of the mirror plate.

[0087] S33: Solve the first-level multi-parameter matrix information through an inner product model to obtain the first-level multi-parameter inner product matrix information;

[0088] In step S33, the inner product model is:

[0089] ; (2)

[0090] In the above formula (2), represents the first-level multi-parameter inner product matrix information, represents the first-level multi-parameter matrix information, represents the transposed matrix information of the first-level multi-parameter matrix.

[0091] S34: Solve the first-level multi-parameter inner product matrix information through a characteristic matrix model to obtain the first-level eigenvector matrix information and the first-level diagonal matrix information;

[0092] In step S34, the characteristic matrix model uses an existing analytical algorithm to decompose the first-level multi-parameter inner product matrix information to obtain the first-level eigenvector matrix information and the first-level diagonal matrix information , where the first-level diagonal matrix information , where ;

[0093] S35: Solve the rotor roundness weight information, the stator roundness weight information, the first-level eigenvector matrix information, and the first-level diagonal matrix information through the first correlation model to obtain a first processed data set.

[0094] In step S35, the first correlation model is:

[0095] ; (3)

[0096] In the above formula (3), represents the first processed data set, represents the rotor roundness weight information, represents the stator roundness weight information, represents the diagonal matrix when taking the largest eigenvalues in the first-level diagonal matrix information, represents the eigenvector matrix when taking the largest eigenvalues in the first-level eigenvector matrix information, represents a first preset constant coefficient, which is related to the maintenance standard of the current hydropower station.

[0097] S4: Solve the second data set and the third data set through a preset second correlation model to obtain a second processed data set; to clarify the specific processing process of the second processed data set, this step includes steps S41 to S45, specifically:

[0098] S41: Obtain the displacement weight information of the large shaft air supply pipe, the slip ring displacement weight information, and the guide vane assembly displacement weight information of the current hydropower station;

[0099] In step S41, the calculation principles of the displacement weight information of the large shaft air supply pipe and the slip ring displacement weight information are the same as those of the rotor roundness weight information in step S31.

[0100] In step S41, when obtaining the displacement weight information of the guide vane assembly, the calculation of the displacement weight information of the guide vane assembly includes steps S411 to S414, specifically:

[0101] S411: Obtain the upper guide displacement weight information, the lower guide displacement weight information, the water guide displacement weight information, the thrust head displacement weight information, and the lower flange displacement information of the rotor;

[0102] In step S411, the calculation principles of the upper guide displacement weight information, the lower guide displacement weight information, the water guide displacement weight information, the thrust head displacement weight information, and the lower flange displacement information of the rotor are the same as those of the rotor roundness weight information in step S31.

[0103] S412: Calculate the upper guide displacement weight information and the lower guide displacement weight information to obtain a first judgment coefficient;

[0104] In step S412, calculate the average value of the upper guide displacement weight information and the lower guide displacement weight information to obtain a first judgment coefficient.

[0105] S413: Calculate the water guide displacement weight information, the thrust head displacement weight information, and the lower flange displacement information of the rotor to obtain a second judgment coefficient;

[0106] In this step, calculate the average value of the above three to obtain a second judgment coefficient.

[0107] S414: Solve the first judgment coefficient and the second judgment coefficient through a preset weight model to obtain the water guide assembly displacement weight information.

[0108] ; (4)

[0109] In the above formula (4), represents the water guide assembly displacement weight information, represents the first judgment coefficient, represents the first judgment coefficient, and both represent preset adjustment parameters, which are related to the maximum displacement of the water guide assembly.

[0110] S42: Construct a matrix based on the large shaft makeup air pipe displacement information, the slip ring displacement information, the water guide assembly displacement information, and the horizontal information of the mirror plate within a preset time to obtain the second-level multi-parameter matrix information;

[0111] In this step S42, the second-level multi-parameter matrix information is:

[0112] ; (5)

[0113] In the above formula (5), represents the second-level multi-parameter matrix information, represents the large shaft makeup air pipe displacement information, represents the slip ring displacement information, represents the water guide assembly displacement information, represents the characteristic data corresponding to the horizontal information of the mirror plate.

[0114] S43: Solve the second-level multi-parameter matrix information through an inner product model to obtain the second-level multi-parameter inner product matrix information;

[0115] In step S43, the inner product model is:

[0116] ; (6)

[0117] In the above formula (6), represents the second-level multi-parameter inner product matrix information, represents the second-level multi-parameter matrix information, represents the transposed matrix information of the second-level multi-parameter matrix.

[0118] S44: Solve the second-level multi-parameter inner product matrix information through the eigenmatrix model to obtain the second-level eigenvector matrix information and the second-level diagonal matrix information;

[0119] In step S44, the eigenmatrix model uses an existing analytical algorithm to decompose the second-level multi-parameter inner product matrix information to obtain the second-level eigenvector matrix information and the second-level diagonal matrix information , where the second-level diagonal matrix information , where ;

[0120] S45: Solve the large shaft compensating air pipe displacement weight information, the slip ring displacement weight information, the water guide component displacement weight information, the second-level eigenvector matrix information, and the second-level diagonal matrix information through the second correlation model to obtain the second processing data set.

[0121] In step S45, the second correlation model is:

[0122] ; (7)

[0123] In the above formula (7), represents the second processing data set, represents the large shaft compensating air pipe displacement weight information, represents the slip ring displacement weight information, represents the water guide component displacement weight information, represents the diagonal matrix when taking the largest eigenvalues in the second-level diagonal matrix information, represents the eigenvector matrix when taking the largest eigenvalues in the second-level eigenvector matrix information, represents the second preset constant coefficient.

[0124] S5: Solve the first processing data set and the second processing data set through a preset multi-source data fusion model to obtain the fusion data of the barring gear result;

[0125] In step S5, the preset multi-source data fusion model can use an existing D-S fusion algorithm or entropy theory algorithm.

[0126] S6: Input the fusion data of the state data of the barring gear within a preset time and the barring gear result into a preset recurrent neural network for training to obtain a barring gear state recognition model;

[0127] In step S6, it includes steps S61 to S63, specifically:

[0128] S61: Perform normalization calculation on the state data of the barring gear within a preset time to obtain the normalized data of the barring gear state;

[0129] S62: Perform normalization calculation on the fusion data of the barring gear result to obtain the normalized data of the barring gear result;

[0130] In steps S61 and S62, the normalization calculation uses the existing Z-Score model.

[0131] S63: Input the normalized data of the barring gear state and the normalized data of the barring gear result into a preset recurrent neural network for training to obtain a barring gear state recognition model. In the recurrent neural network, it includes an input layer, multiple hidden layers, and an output layer. Each neuron in the hidden layer is a recurrent unit. The input layer is fully connected to the first hidden layer, the last hidden layer is fully connected to the output layer, and a memory unit is set between the output layer and the input layer.

[0132] In step S63, each neuron in the hidden layer is a recurrent unit. When performing recurrent learning in each recurrent unit, it includes S631 to S634, specifically:

[0133] S631: Obtain the input weights of the forward neural network, the input weights of the backward neural network, and the hidden layer function;

[0134] S632: Learn the input weights of the forward neural network and the hidden layer function through a preset forward learning model to obtain the output data of the forward recurrent unit;

[0135] In step S62, the forward learning model uses the forward propagation model in LSTM.

[0136] S633: Learn the input weights of the backward neural network and the hidden layer function through a preset backward learning model to obtain the output data of the backward recurrent unit;

[0137] In step S63, the backward learning model uses the backward propagation model in LSTM.

[0138] S634: Connect the output data of the forward loop unit and the output data of the backward loop unit through a preset connection function to obtain the output data of the current loop unit.

[0139] In step S634, the preset connection function is:

[0140] ; (8)

[0141] In the above formula (8), represents the output data of the current loop unit at time, represents a preset activation function, represents a preset weight matrix corresponding to the output data of the forward loop unit and the output data of the backward loop unit, represents the output data of the forward loop unit, represents the output data of the backward loop unit.

[0142] S7: Input the real-time state detection information of the barring gear into the barring gear state recognition model to obtain the adjustment characteristic information of the barring gear, and the adjustment characteristic information of the barring gear is used for overhaul adjustment of the barring gear.

[0143] Since data association has been performed in the previous data analysis, in this step, the corresponding components can be adjusted according to the adjustment characteristic information of the barring gear. The adjustment characteristic information of the barring gear is set in multiple levels, such as a ten-level gradient setting. Specifically:

[0144] In the first to third levels, the roundness of the rotor end face and the roundness of the stator end face are the priority adjustment objects, and the displacement of the large shaft gas supply pipe, the displacement of the slip ring, the displacement of the water guide assembly, and the horizontal position of the mirror plate are set as adjustment objects of the same level;

[0145] In the fourth to seventh levels, the displacement of the large shaft gas supply pipe, the displacement of the slip ring, and the displacement of the water guide assembly are the priority adjustment objects, and the roundness of the rotor end face, the roundness of the stator end face, and the horizontal position of the mirror plate are set as adjustment objects of the same level; among them, in the fifth and sixth levels, the adjustment level of the displacement of the water guide assembly is higher than the adjustment levels of the displacement of the large shaft gas supply pipe and the displacement of the slip ring;

[0146] In the eighth to tenth levels, the horizontal position of the mirror plate is the priority adjustment object, and the roundness of the rotor end face, the roundness of the stator end face, the displacement of the large shaft gas supply pipe, the displacement of the slip ring, and the displacement of the water guide assembly are set as adjustment objects of the same level.

[0147] Combining the above steps, in this method, first, the data of each part of the unit is parsed and classified. Then, the first data set and the third data set are associated, and the second data set and the third data set are associated. Next, the processed data is fused to ensure the redundancy or complementarity of different processed data in space or time, so as to obtain a key analysis of the priority items of the turning gear result. Finally, the real-time status detection information of the turning gear is input into the turning gear status recognition model to obtain the adjustment feature information of the turning gear, achieving the purpose of guiding the overhaul adjustment. This method establishes the data correlation of each part of the unit, realizes the real-time reading of the relevant data of each part of the unit during turning gear, and automatically completes the data analysis and the auxiliary decision-making of the axis adjustment by using the analysis algorithm, thus making the whole turning gear visualization and simplification, and achieving the goal of saving the project overhaul cost and shortening the project duration.

[0148] Embodiment 2:

[0149] As Figure 2 shown, it is a topological structure diagram when a data analysis system for the turning gear result of a hydropower station is applied. Specifically: First, the parameters at different component positions are collected. Then, the parameters at different component positions are input into this analysis system. This analysis system outputs the adjustment feature information of the turning gear. Finally, the overhaul adjustment of the corresponding position of the turning gear is carried out according to the adjustment feature information of the turning gear.

[0150] This embodiment provides a data analysis system for the turning gear result of a hydropower station. The system includes:

[0151] An acquisition module, configured to acquire the status data of the turning gear within a preset time and the real-time status detection information of the turning gear;

[0152] A first processing module, configured to perform data parsing on the status data of the turning gear within a preset time to obtain a parsed data set of the turning gear result. The parsed data set of the turning gear result includes a first data set, a second data set, and a third data set. The first data set includes the rotor end face roundness information and the stator end face roundness information. The second data set includes the displacement information of the large shaft air supply pipe, the displacement information of the slip ring, and the displacement information of the guide vane assembly. The third data set is the horizontal information of the mirror plate;

[0153] A second processing module, configured to solve the first data set and the third data set through a preset first association model to obtain a first processed data set;

[0154] A third processing module, configured to solve the second data set and the third data set through a preset second association model to obtain a second processed data set;

[0155] A fourth processing module, configured to solve the first processed data set and the second processed data set through a preset multi-source data fusion model to obtain the fusion data of the turning gear result;

[0156] A fifth processing module, configured to input the fusion data of the state data of the barring gear within a preset time and the barring gear result into a preset recurrent neural network for training to obtain a barring gear state recognition model;

[0157] An identification module, configured to input the real-time state detection information of the barring gear into the barring gear state recognition model to obtain the adjustment feature information of the barring gear, and the adjustment feature information of the barring gear is used for overhaul adjustment of the barring gear.

[0158] In an implementation method disclosed by the present invention, in the second processing module 3, the rotor end face roundness information includes the rotor upper end face roundness information and the rotor lower end face roundness information, and the stator end face roundness information includes the stator upper end face roundness information and the stator lower end face roundness information, including:

[0159] A first acquisition unit, configured to acquire the rotor roundness weight information and the stator roundness weight information of the current hydropower station;

[0160] A first calculation unit, configured to construct a matrix according to the rotor upper end face roundness information, the rotor lower end face roundness information, the stator upper end face roundness information, the stator lower end face roundness information, and the level information of the mirror plate within a preset time to obtain first-level multi-parameter matrix information;

[0161] A second calculation unit, configured to solve the first-level multi-parameter matrix information through an inner product model to obtain first-level multi-parameter inner product matrix information;

[0162] A third calculation unit, configured to solve the first-level multi-parameter inner product matrix information through a feature matrix model to obtain first-level eigenvector matrix information and first-level diagonal matrix information;

[0163] A fourth calculation unit, configured to solve the rotor roundness weight information, the stator roundness weight information, the first-level eigenvector matrix information, and the first-level diagonal matrix information through a first association model to obtain a first processing data set.

[0164] In an implementation method disclosed by the present invention, in the third processing module 4, it includes:

[0165] A second acquisition unit, configured to acquire the displacement weight information of the large shaft gas supply pipe, the slip ring displacement weight information, and the guide vane assembly displacement weight information of the current hydropower station;

[0166] A fifth calculation unit, configured to construct a matrix according to the large shaft gas supply pipe displacement information, the slip ring displacement information, the guide vane assembly displacement information, and the level information of the mirror plate within a preset time to obtain second-level multi-parameter matrix information;

[0167] The sixth calculation unit is configured to solve the second-level multi-parameter matrix information through an inner product model to obtain second-level multi-parameter inner product matrix information;

[0168] The seventh calculation unit is configured to solve the second-level multi-parameter inner product matrix information through an eigenmatrix model to obtain second-level eigenvector matrix information and second-level diagonal matrix information;

[0169] The eighth calculation unit is configured to solve the large shaft air supply pipe displacement weight information, the slip ring displacement weight information, the water guide assembly displacement weight information, the second-level eigenvector matrix information, and the second-level diagonal matrix information through a second correlation model to obtain a second processed data set.

[0170] In this system, the system networking adopts a combination of wired / wireless methods to maximize the signal stability and usability requirements. Among them, the wireless transmission protocol of this system adopts LoRa technology. When LoRa technology is adopted, this system has the following advantages:

[0171] ①. Long distance;

[0172] The signal strength that LoRa can demodulate is one hundred-thousandth of that of Bluetooth and ZigBee. In addition, the ultra-high sensitivity of LoRa comes from the modulation itself, does not depend on narrowband (Sigfox uses ultra-narrowband technology), does not depend on retransmission (NB-IOT uses retransmission technology), and does not depend on coding redundancy.

[0173] ②. Strong anti-interference ability;

[0174] LoRa can achieve long-distance transmission. In addition to the sensitivity advantage, there is also a very important factor, which is its strong anti-interference ability. LoRa has an ultimate anti-interference technology that can still communicate with a signal strength 20dB lower than the noise, which is not available in existing traditional communication technologies.

[0175] In addition, LoRa also has a very good ability to cope with stronger sudden random interferences. If faced with a strong interference with a burst length less than half of the LoRa symbol length or a duty cycle of the interference <50%, LoRa can still stably demodulate and ensure that its sensitivity deterioration <3dB.

[0176] ③. Low power consumption;

[0177] LoRa modulation has the characteristics of not depending on narrowband, retransmission, and coding redundancy. Therefore, LoRa modulation is a very efficient modulation method with a very low operating current. Its static current is 1pA; the receiving current is less than 5mA; when the transmitting power is 17dBm, the current is only 45mA.

[0178] Such asFigure 3 As shown, it is a schematic diagram of the data calculation principle of this data analysis system.

[0179] It should be noted that regarding the system in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0180] Embodiment 3:

[0181] Corresponding to the above method embodiments, a data analysis device for the turbine turning result of a hydropower station is also provided in this embodiment. A data analysis device for the turbine turning result of a hydropower station described below can be correspondingly referred to the data analysis method for the turbine turning result of a hydropower station described above.

[0182] Figure 4 It is a block diagram of a data analysis device 800 for the turbine turning result of a hydropower station shown according to an exemplary embodiment. As Figure 4 shown, the data analysis device 800 for the turbine turning result of a hydropower station may include: a processor 801, a memory 802. The data analysis device 800 for the turbine turning result of a hydropower station may further include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0183] Among them, the processor 801 is used to control the overall operation of the data analysis device 800 for the hydroelectric power station turning test results, so as to complete all or part of the steps in the above-mentioned data analysis method for the hydroelectric power station turning test results. The memory 802 is used to store various types of data to support the operation of the data analysis device 800 for the hydroelectric power station turning test results. These data may include, for example, instructions for any application or method operating on the data analysis device 800 for the hydroelectric power station turning test results, as well as application-related data, such as contact data, sent and received messages, pictures, audio, video, and so on. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disc. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone, and the microphone is used to receive external audio signals. The received audio signals may be further stored in the memory 802 or sent through the communication component 805. The audio component further includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the above-mentioned other interface modules may be a keyboard, a mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the data analysis device 800 for the hydroelectric power station turning test results and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination of one or more of them. Therefore, the corresponding communication component 805 may include: a Wi-Fi module, a Bluetooth module, and an NFC module.

[0184] In an exemplary embodiment, the data analysis device 800 for the turbine turning test results of a hydropower station may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components, and is used to execute the above-mentioned data analysis method for the turbine turning test results of a hydropower station.

[0185] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When the program instructions are executed by a processor, the steps of the above-mentioned data analysis method for the turbine turning test results of a hydropower station are implemented. For example, the computer-readable storage medium may be the above-mentioned memory 802 including program instructions, and the above program instructions may be executed by the processor 801 of the data analysis device 800 for the turbine turning test results of a hydropower station to complete the above-mentioned data analysis method for the turbine turning test results of a hydropower station.

[0186] Embodiment 4:

[0187] Corresponding to the above method embodiment, a readable storage medium is also provided in this embodiment. A readable storage medium described below and a data analysis method for the turbine turning test results of a hydropower station described above can be correspondingly referred to each other.

[0188] A readable storage medium stores a computer program. When the computer program is executed by a processor, the steps of the data analysis method for the turbine turning test results of the above method embodiment are implemented.

[0189] The readable storage medium may specifically be various readable storage media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disc that can store program codes.

Claims

1. A data analysis method for the turning test results of a hydropower station, characterized in that, Including: Obtain the status data of the barring gear within a preset time and the real-time status detection information of the barring gear; Perform data analysis on the status data of the barring gear within a preset time to obtain an analysis data set of the barring gear result. The analysis data set of the barring gear result includes a first data set, a second data set, and a third data set. The first data set includes rotor end face roundness information and stator end face roundness information. The second data set includes large shaft gas supply pipe displacement information, slip ring displacement information, and guide water component displacement information. The third data set is the horizontal information of the mirror plate; Solve the first data set and the third data set through a preset first correlation model to obtain a first processed data set; Solve the second data set and the third data set through a preset second correlation model to obtain a second processed data set; Solve the first processed data set and the second processed data set through a preset multi-source data fusion model to obtain the fusion data of the barring gear result; Input the status data of the barring gear within a preset time and the fusion data of the barring gear result into a preset recurrent neural network for training to obtain a barring gear status recognition model; Input the real-time status detection information of the barring gear into the barring gear status recognition model to obtain the adjustment feature information of the barring gear, and the adjustment feature information of the barring gear is used for overhaul and adjustment of the barring gear.

2. The data analysis method for the turning result of a hydropower station according to claim 1, wherein Solve the first data set and the third data set through a preset first correlation model to obtain a first processed data set. The rotor end face roundness information includes rotor upper end face roundness information and rotor lower end face roundness information. The stator end face roundness information includes stator upper end face roundness information and stator lower end face roundness information, including: Obtain the rotor roundness weight information and stator roundness weight information of the current hydropower station; Construct a matrix based on the rotor upper end face roundness information, the rotor lower end face roundness information, the stator upper end face roundness information, the stator lower end face roundness information, and the horizontal information of the mirror plate within a preset time to obtain first-level multi-parameter matrix information; Solve the first-level multi-parameter matrix information through an inner product model to obtain first-level multi-parameter inner product matrix information; Solve the first-level multi-parameter inner product matrix information through a feature matrix model to obtain first-level eigenvector matrix information and first-level diagonal matrix information; Solve the rotor roundness weight information, the stator roundness weight information, the first-level eigenvector matrix information, and the first-level diagonal matrix information through a first correlation model to obtain a first processed data set.

3. The data analysis method for the turning result of a hydropower station according to claim 2, characterized in that, Solve the rotor roundness weight information, the stator roundness weight information, the first-level eigenvector matrix information, and the first-level diagonal matrix information through a first correlation model to obtain a first processed data set, where the first correlation model is: ； In the formula, represents the first processed data set, represents the rotor roundness weight information, represents the stator roundness weight information, represents the diagonal matrix when taking the largest eigenvalues in the first-level diagonal matrix information, represents the eigenvector matrix when taking the largest eigenvalues in the first-level eigenvector matrix information, represents the first preset constant coefficient.

4. The data analysis method for the turning gear result of a hydropower station according to claim 1, wherein Solve the second data set and the third data set through a preset second correlation model to obtain a second processed data set, including: Obtain the large shaft gas supply pipe displacement weight information, slip ring displacement weight information, and guide water component displacement weight information of the current hydropower station; Construct a matrix based on the displacement information of the large shaft gas supply pipe, the displacement information of the slip ring, the displacement information of the water guide assembly, and the horizontal information of the mirror plate within a preset time to obtain the second-level multi-parameter matrix information; Solve the second-level multi-parameter matrix information through the inner product model to obtain the second-level multi-parameter inner product matrix information; Solve the second-level multi-parameter inner product matrix information through the eigenmatrix model to obtain the second-level eigenvector matrix information and the second-level diagonal matrix information; Solve the displacement weight information of the large shaft gas supply pipe, the displacement weight information of the slip ring, the displacement weight information of the water guide assembly, the second-level eigenvector matrix information, and the second-level diagonal matrix information through the second correlation model to obtain the second processed data set.

5. The data analysis method for the turning result of a hydropower station according to claim 4, wherein Solve the displacement weight information of the large shaft gas supply pipe, the displacement weight information of the slip ring, the displacement weight information of the water guide assembly, the second-level eigenvector matrix information, and the second-level diagonal matrix information through the second correlation model to obtain the second processed data set, where the second correlation model is: ； In the formula, represents the second processed data set, represents the displacement weight information of the large shaft gas supply pipe, represents the displacement weight information of the slip ring, represents the displacement weight information of the water guide component, represents taking diagonal matrices when taking the largest eigenvalues in the second-level diagonal matrix information, represents taking eigenvector matrices when taking the largest eigenvalues in the second-level eigenvector matrix information, represents the second preset constant coefficient.

6. The data analysis method for the turning result of a hydropower station according to claim 4, wherein In obtaining the displacement weight information of the water guide assembly, the calculation of the displacement weight information of the water guide assembly includes: Obtain the upper guide displacement weight information, the lower guide displacement weight information, the water guide displacement weight information, the thrust head displacement weight information, and the rotor lower flange displacement information; Calculate the upper guide displacement weight information and the lower guide displacement weight information to obtain the first judgment coefficient; Calculate the water guide displacement weight information, the thrust head displacement weight information, and the rotor lower flange displacement information to obtain the second judgment coefficient; Solve the first judgment coefficient and the second judgment coefficient through a preset weight model to obtain the displacement weight information of the water guide assembly.

7. The data analysis method for the turning result of a hydropower station according to claim 1, wherein Input the fusion data of the state data of the barring gear within a preset time and the barring gear result into a preset recurrent neural network for training to obtain a barring gear state recognition model, including: Perform normalization calculation on the state data of the barring gear within a preset time to obtain the normalized data of the barring gear state; Perform normalization calculation on the fusion data of the barring gear result to obtain the normalized data of the barring gear result; Input the normalized data of the barring gear state and the normalized data of the barring gear result into a preset recurrent neural network for training to obtain a barring gear state recognition model. In the recurrent neural network, there is an input layer, multiple hidden layers, and an output layer. Each neuron in the hidden layer is a recurrent unit. The input layer is fully connected to the first hidden layer, the last hidden layer is fully connected to the output layer, and a memory unit is set between the output layer and the input layer.

8. A data analysis system for the turning test results of a hydropower station, characterized in that, Include: An acquisition module for acquiring the state data of the barring gear within a preset time and the real-time state detection information of the barring gear; The first processing module is used to parse the state data of the barring gear within a preset time to obtain an analysis data set of the barring gear result. The analysis data set of the barring gear result includes a first data set, a second data set, and a third data set. The first data set includes rotor end face roundness information and stator end face roundness information. The second data set includes large shaft gas supply pipe displacement information, slip ring displacement information, and water guide assembly displacement information. The third data set is the horizontal information of the mirror plate. The second processing module is used to solve the first data set and the third data set through a preset first correlation model to obtain a first processed data set. The third processing module is used to solve the second data set and the third data set through a preset second correlation model to obtain a second processed data set. The fourth processing module is used to solve the first processed data set and the second processed data set through a preset multi-source data fusion model to obtain a fusion data of the barring gear result. The fifth processing module is used to input the state data of the barring gear within a preset time and the fusion data of the barring gear result into a preset recurrent neural network for training to obtain a barring gear state recognition model. The recognition module is used to input the real-time state detection information of the barring gear into the barring gear state recognition model to obtain adjustment feature information of the barring gear. The adjustment feature information of the barring gear is used to perform maintenance adjustment on the barring gear.

9. The data analysis system for the turning result of a hydropower station according to claim 8, wherein In the second processing module, the rotor end face roundness information includes rotor upper end face roundness information and rotor lower end face roundness information, and the stator end face roundness information includes stator upper end face roundness information and stator lower end face roundness information, including: The first acquisition unit is used to acquire the rotor roundness weight information and stator roundness weight information of the current hydropower station. The first calculation unit is used to construct a first-level multi-parameter matrix information according to the rotor upper end face roundness information, the rotor lower end face roundness information, the stator upper end face roundness information, the stator lower end face roundness information, and the horizontal information of the mirror plate within a preset time. The second calculation unit is used to solve the first-level multi-parameter matrix information through an inner product model to obtain a first-level multi-parameter inner product matrix information. The third calculation unit is used to solve the first-level multi-parameter inner product matrix information through a feature matrix model to obtain a first-level eigenvector matrix information and a first-level diagonal matrix information. The fourth calculation unit is used to solve the rotor roundness weight information, the stator roundness weight information, the first-level eigenvector matrix information, and the first-level diagonal matrix information through a first correlation model to obtain a first processed data set.

10. The data analysis system for the disk turning result of a hydropower station according to claim 8, characterized in that, In the third processing module, it includes: The second acquisition unit is used to acquire the large shaft gas supply pipe displacement weight information, slip ring displacement weight information, and water guide assembly displacement weight information of the current hydropower station. The fifth calculation unit is used to construct a second-level multi-parameter matrix information according to the large shaft gas supply pipe displacement information, the slip ring displacement information, the water guide assembly displacement information, and the horizontal information of the mirror plate within a preset time. The sixth computing unit is configured to solve the second-level multi-parameter matrix information through an inner product model to obtain second-level multi-parameter inner product matrix information; The seventh computing unit is configured to solve the second-level multi-parameter inner product matrix information through an eigenmatrix model to obtain second-level eigenvector matrix information and second-level diagonal matrix information; The eighth computing unit is configured to solve the large shaft gas supply pipe displacement weight information, the slip ring displacement weight information, the water guide assembly displacement weight information, the second-level eigenvector matrix information, and the second-level diagonal matrix information through a second correlation model to obtain a second processed data set.

Images