Quantum gas detection method, device and equipment for power transformation equipment and readable storage medium

Abstract

The invention discloses a power transformation equipment quantum gas detection method, device and equipment and a readable storage medium, and the method comprises the steps: obtaining an error prediction model corresponding to a target gas component, the error prediction model being obtained through training based on a concentration training error of the target gas component detected by a gas concentration detection system at different time points; detecting the concentration information of the target gas component when the internal gas of the power transformation equipment is not introduced; inputting the concentration information into an error prediction model to obtain a concentration prediction error; introducing gas in the power transformation equipment into a gas concentration detection system, and measuring the concentration value of a target gas component; and calculating the content of the target gas component in the internal gas of the power transformation equipment in combination with the concentration value and the concentration prediction error. Therefore, the content estimation deviation caused by long-term correlation noise can be removed from the measured data, the gas component content estimation precision is improved, the anti-interference capability of gas content measurement is improved, and stable operation of power transformation equipment and safety and reliability of a power system are guaranteed.

Description

Technical Field

[0001] The present application relates to the field of power grid technology, and more specifically, to a method, device, equipment and readable storage medium for quantum gas detection in substation equipment. Background Art

[0002] As the dependence on electric energy in various fields increases, the reliability and safety of power systems are becoming more and more critical. As the core component of the power system, the performance of substation equipment is closely related to the stable operation of the power system. Among the many substation equipment performance detection technologies, the internal gas composition detection technology of substation equipment is an indispensable part of the fault diagnosis of substation equipment.

[0003] In order to improve the detection accuracy, the existing technology attempts to introduce quantum technology into the field of gas composition detection inside substation equipment. However, the above technology is greatly affected by noise in actual application, resulting in inaccurate detection of gas composition inside substation equipment, which can easily trigger a chain reaction and affect the safe and stable operation of the power system. It may even cause power system failures, resulting in immeasurable losses. Summary of the invention

[0004] In view of this, the present application provides a method, apparatus, device and readable storage medium for quantum gas detection in substation equipment, which are used to improve the measurement accuracy of existing technology for detecting gas composition inside substation equipment.

[0005] In order to achieve the above objectives, the proposed solution is as follows:

[0006] A quantum gas detection method for power transformation equipment, comprising:

[0007] Obtaining an error prediction model corresponding to the target gas component, wherein the error prediction model is obtained based on concentration training errors of the target gas component detected by the gas concentration detection system at different time points;

[0008] When the gas inside the substation equipment is not introduced, detecting the concentration information of the target gas component in the gas concentration detection system;

[0009] Inputting the concentration information into the error prediction model to obtain the concentration prediction error of the target gas component;

[0010] Introducing the gas inside the power substation into the gas concentration detection system to measure the concentration value of the target gas component;

[0011] The content of the target gas component in the gas inside the substation is calculated based on the concentration value and the concentration prediction error.

[0012] Optionally, obtaining the error prediction model corresponding to the target gas composition includes:

[0013] Build machine learning models for error prediction;

[0014] The concentration training error of the target gas component in the gas concentration detection system is collected at different time points within a continuous cycle without introducing the internal gas of the substation equipment;

[0015] The concentration training errors of the same cycle are combined in sequence according to the time points to form an error training sequence;

[0016] The machine learning model is trained using each error training sequence, and the machine learning model finally obtained by training is used as the error prediction model corresponding to the target gas component.

[0017] Optionally, combining the concentration training errors of the same period in sequence according to time points to form an error training sequence includes:

[0018] N+1 consecutive time points are selected from the same cycle, and the concentration training errors corresponding to the first N selected time points are combined in sequence according to the time points to form training data, and the concentration training error corresponding to the selected N+1th time point is marked on the training data to form an error training sequence.

[0019] Optionally, combining the concentration training errors of the same period in sequence according to time points to form an error training sequence includes:

[0020] N+1 consecutive time points are selected from the same cycle, and the concentration training errors corresponding to the first N selected time points are combined in sequence according to the time points to form training data, and the concentration training error corresponding to the selected N+1th time point is marked on the training data to form an error training sequence.

[0021] Optionally, after the machine learning model is trained using each error training sequence, the method further includes:

[0022] A test set is constructed based on the concentration test error of the target gas component in the gas concentration detection system when the internal gas of the substation equipment is not introduced at different time points;

[0023] The prediction accuracy of the trained machine learning model is evaluated based on the test set. When the prediction accuracy does not meet the preset accuracy condition, return to the step of training the machine learning model using each error training sequence until the prediction accuracy of the trained machine learning model meets the accuracy condition.

[0024] Optionally, when the internal gas of the power transformation equipment is not introduced, detecting the concentration information of the target gas component in the gas concentration detection system includes:

[0025] In combination with non-orthogonal quantum measurement technology, when the internal gas of the substation equipment is not introduced, the concentration information of the target gas component in the gas concentration detection system is detected.

[0026] Optionally, introducing the gas inside the power transformation equipment into the gas concentration detection system to measure the concentration value of the target gas component includes:

[0027] The gas inside the power conversion equipment is introduced into the gas concentration detection system, and the concentration value of the target gas component is measured in combination with non-orthogonal quantum measurement technology.

[0028] A quantum gas detection device for power transformation equipment, comprising:

[0029] An acquisition module, used to acquire an error prediction model corresponding to a target gas component, wherein the error prediction model is obtained based on concentration training errors of the target gas component detected by a gas concentration detection system at different time points;

[0030] A detection module, used for detecting the concentration information of the target gas component in the gas concentration detection system when the gas inside the substation is not introduced;

[0031] A prediction module, used for inputting the concentration information into the error prediction model to obtain a concentration prediction error of the target gas component;

[0032] A measuring module, used for introducing the gas inside the power substation into the gas concentration detection system to measure the concentration value of the target gas component;

[0033] A calculation module is used to calculate the content of the target gas component in the gas inside the substation by combining the concentration value and the concentration prediction error.

[0034] A quantum gas detection device for power substation equipment, comprising a memory and a processor;

[0035] The memory is used to store programs;

[0036] The processor is used to execute the program to implement the various steps of the above-mentioned substation equipment quantum gas detection method.

[0037] A readable storage medium stores a computer program, which, when executed by a processor, implements each step of the above-mentioned method for detecting quantum gas in substation equipment.

[0038] It can be seen from the above technical scheme that the quantum gas detection method for substation equipment provided by the present application can obtain an error prediction model corresponding to the target gas component, and the error prediction model is obtained based on the concentration training error of the target gas component detected by the gas concentration detection system at different time points; when the internal gas of the substation equipment is not introduced, the concentration information of the target gas component in the gas concentration detection system is detected; the concentration information is input into the error prediction model to obtain the concentration prediction error of the target gas component; based on this, the present application can use the trained error prediction model corresponding to the specific gas component and the concentration information when the internal gas of the substation equipment is not introduced to generate the concentration prediction error, and quantify and digitize the subsequent concentration error; then , the gas inside the substation equipment can be introduced into the gas concentration detection system to measure the concentration value of the target gas component; based on this, the present application can measure the concentration value containing errors; then, the concentration value and the concentration prediction error can be combined to calculate the content of the target gas component in the gas inside the substation equipment; based on this, the present application can remove the quantified concentration prediction error from the measured concentration value, reduce the measurement error caused by different factors such as noise and the gas detection device itself, improve the reliability and accuracy of the target gas component content measurement, help to accurately evaluate the gas condition inside the substation equipment, improve the accuracy and anti-interference ability of gas concentration estimation, and provide a reliable gas content basis for substation equipment fault detection and diagnosis. It can be seen that the present application can remove the content estimation deviation caused by long-term correlation noise in the measurement data, improve the accuracy of gas component content estimation, and improve the anti-interference ability of gas content measurement, thereby ensuring the stable operation of substation equipment and the safety and reliability of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0040] Figure 1 A flow chart of a method for detecting quantum gas in a power substation disclosed in an embodiment of the present application;

[0041] Figure 2 A system architecture diagram of a gas concentration detection system disclosed in an embodiment of the present application;

[0042] Figure 3 This is a structural block diagram of a quantum gas detection device for power substation equipment disclosed in an embodiment of the present application;

[0043] Figure 4 This is a hardware structure block diagram of a quantum gas detection device for power transformation equipment disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0045] The embodiment of the present application provides a method for detecting quantum gas in a substation. The method can be applied to various substation detection systems or gas composition detection systems, and can also be applied to various computer terminals or smart terminals. The execution subject can be a processor or server of a computer terminal or a smart terminal. The method flow chart of the method for detecting quantum gas in a substation is as follows: Figure 1 As shown, specifically including:

[0046] Step S1, obtaining an error prediction model corresponding to the target gas composition.

[0047] Specifically, the error prediction model can be obtained based on the concentration training error of the target gas component detected by the gas concentration detection system at different time points.

[0048] The error prediction model can be used to analyze the variation of the target gas composition measurement error over time in the gas concentration detection system.

[0049] The target gas component may be any gas related to a failure of a power transformation equipment.

[0050] Among them, the error prediction model can be generated in many ways.

[0051] For example, a linear regression model, a polynomial regression model, an exponential model or a logarithmic model can be established, and the least squares method can be used to adjust the parameters of the linear regression model, the polynomial regression model, the exponential model or the logarithmic model based on the concentration training error of the target gas component detected by the gas concentration detection system at different time points.

[0052] Different gas components may correspond to different error prediction models, or may correspond to the same error prediction model.

[0053] The same gas concentration detection system can correspond to error prediction models of different gas components.

[0054] The same gas composition can correspond to multiple error prediction models.

[0055] Step S2: When the internal gas of the power transformation equipment is not introduced, the concentration information of the target gas component in the gas concentration detection system is detected.

[0056] Specifically, the gas concentration detection system can be used as a detection system for detecting the content of any gas component in the gas inside the substation equipment.

[0057] The concentration information of the target gas component in the gas concentration detection system can be detected without introducing any gas inside the substation equipment, wherein the concentration information can be composed of multiple continuous concentration values ​​or can be the concentration value at the current moment.

[0058] Step S3: input the concentration information into the error prediction model to obtain the concentration prediction error of the target gas component.

[0059] Specifically, the concentration information can be input into the error prediction model, and the error prediction model can be used to analyze the change of the detection error of the target gas component in the gas concentration detection system over time, and the concentration prediction error can be output based on the change of the event.

[0060] Among them, a rolling prediction method can be used to input concentration information into the error prediction model to obtain the concentration prediction error of the target gas component.

[0061] Specifically, the concentration information may include the concentration error of the target gas component detected by the gas concentration detection system at each collection point in the time period before the current moment when the internal gas of the substation equipment is not introduced; based on the concentration error of each collection point in the previous time period, combined with the error prediction model, the error value is predicted; based on the concentration error of each collection point and the latest predicted error value, combined with the error prediction model, the concentration prediction error of the target gas component is obtained.

[0062] Step S4: Introduce the gas inside the power transformation equipment into the gas concentration detection system to measure the concentration value of the target gas component.

[0063] Specifically, the gas inside the power substation equipment can be input into the gas concentration detection system to measure the concentration value of the target gas component in the gas inside the power substation equipment.

[0064] Step S5: Calculate the content of the target gas component in the gas inside the substation by combining the concentration value and the concentration prediction error.

[0065] Specifically, the concentration prediction error may be subtracted from the concentration value to calculate the content of the target gas component in the gas inside the substation equipment.

[0066] It can be seen from the above technical scheme that the quantum gas detection method for substation equipment provided by the present application can obtain an error prediction model corresponding to the target gas component, and the error prediction model is obtained based on the concentration training error of the target gas component detected by the gas concentration detection system at different time points; when the internal gas of the substation equipment is not introduced, the concentration information of the target gas component in the gas concentration detection system is detected; the concentration information is input into the error prediction model to obtain the concentration prediction error of the target gas component; based on this, the present application can use the trained error prediction model corresponding to the specific gas component and the concentration information when the internal gas of the substation equipment is not introduced to generate the concentration prediction error, and quantify and digitize the subsequent concentration error; then , the gas inside the substation equipment can be introduced into the gas concentration detection system to measure the concentration value of the target gas component; based on this, the present application can measure the concentration value containing errors; then, the concentration value and the concentration prediction error can be combined to calculate the content of the target gas component in the gas inside the substation equipment; based on this, the present application can remove the quantified concentration prediction error from the measured concentration value, reduce the measurement error caused by different factors such as noise and the gas detection device itself, improve the reliability and accuracy of the target gas component content measurement, help to accurately evaluate the gas condition inside the substation equipment, improve the accuracy and anti-interference ability of gas concentration estimation, and provide a reliable gas content basis for substation equipment fault detection and diagnosis. It can be seen that the present application can remove the content estimation deviation caused by long-term correlation noise in the measurement data, improve the accuracy of gas component content estimation, and improve the anti-interference ability of gas content measurement, thereby ensuring the stable operation of substation equipment and the safety and reliability of the power system.

[0067] In some embodiments of the present application, the process of step S1, obtaining the error prediction model corresponding to the target gas composition, is described in detail, and the steps are as follows:

[0068] S10. Build a machine learning model for error prediction.

[0069] Specifically, the parameters of the machine learning model can be initialized and adjusted.

[0070] Among them, a long short-term memory network LSTM can be constructed, and the number of LSTM layers, the value of time_step, the length of the hidden state h, the number of epochs of model training, etc. can be set to form a machine learning model for error prediction.

[0071] LSTM is a special recurrent neural network (RNN) that can be used to process sequence problems. The structural cell state, input gate, forget gate and output gate contained in LSTM can solve the gradient disappearance problem. Therefore, LSTM is suitable for solving this long-term time series data prediction problem.

[0072] There is a structure in LSTM that needs to be executed repeatedly. The number of repetitions is determined by the parameter time_step. time_stemp=N indicates that it needs to be repeated N times, and the size of time_stemp can be set, which is an adjustable parameter. LSTM can pass information from the previous time_step to the next time_step. In each time_step data processing process, the data processing method is the same, and the calculation process is as follows:

[0073]

[0074] in, and Represents the hidden state of the k-th sample at the j-1th time_step and the jth time_step. Usually the hidden state is a vector whose dimension can be set and is an adjustable parameter.

[0075] Input for the jth time_step; , , , is the weight matrix of the network; , , , is the bias vector of the network; , , Represent the forget gate, input gate and output gate vectors respectively; , .

[0076] After the above process is executed N times, we will get the output of the LSTM network , and the length of the hidden state is usually set not to 1. Therefore, a fully connected layer is applied to output the predicted value. If a multi-layer LSTM network is to be established, the fully connected layer should be connected to the last time_step output of the last layer to obtain the predicted value.

[0077] Before training the network, we need to initialize some parameters. To distinguish them from the network parameters W, these parameters are usually called hyperparameters. In the above network, the hyperparameters include the number of LSTM layers, the value of time_step, the length of the hidden state h, the value of N, the number of epochs of model training, and the time_step matrix in the data preprocessing stage. The number of rows N is the same.

[0078] S11. Collect the concentration training error of the target gas component in the gas concentration detection system at different time points within a continuous cycle without introducing the internal gas of the substation equipment.

[0079] Specifically, the concentration training error of the target gas component in the gas concentration detection system when the internal gas of the substation equipment is not introduced can be collected at every hour for multiple consecutive days.

[0080] S12, combining the concentration training errors of the same cycle in sequence according to the time points to form an error training sequence.

[0081] Specifically, the training error of each concentration can be measured multiple times to obtain the average value, data filtering, data normalization, outlier detection and processing, and converted into a data format suitable for processing by a machine learning model.

[0082] Multiple concentration training errors on the same day can be combined in sequence according to the hourly time to form multiple error training sequences.

[0083] For example, the training errors of each concentration in the same cycle can be converted into a training sample matrix and a predicted target value matrix. Multiple continuous elements are selected from the training sample matrix as error training sequences and input into the machine learning network. The elements in the predicted target value matrix of the training sample matrix are used for comparison with the output of the machine learning model.

[0084] S13. Use each error training sequence to train the machine learning model, and use the machine learning model finally obtained by training as the error prediction model corresponding to the target gas component.

[0085] Specifically, each error training sequence can be used in turn to iteratively train the machine learning model to obtain an error prediction model corresponding to the target gas composition.

[0086] It can be seen from the above technical solution that this embodiment provides an optional method for obtaining an error prediction model, which can reduce the fitting phenomenon, improve the generalization ability, and improve the anti-interference ability.

[0087] In some embodiments of the present application, step S12, the process of combining the concentration training errors of the same cycle in sequence according to time points to form an error training sequence is described in detail, and the steps are as follows:

[0088] S120, selecting N+1 consecutive time points from the same cycle, and combining the concentration training errors corresponding to the first N selected time points in sequence according to the time points to form training data, and marking the concentration training error corresponding to the selected N+1th time point on the training data to form an error training sequence.

[0089] Specifically, the concentration training errors corresponding to N+1 consecutive time points in the same cycle can be selected, and the first N selected concentration training errors are selected and combined in sequence according to the time points to form training data, and the selected N+1th concentration training error is used as the annotation label of the training data to form an error training sequence.

[0090] It can be seen from the above technical solution that this embodiment provides an optional method for combining the concentration training errors of the same cycle in sequence according to time points to form an error training sequence. Through the above method, multiple error training sequences can be constructed based on the concentration training errors of the same cycle, thereby improving the training reliability of the machine learning model.

[0091] In some embodiments of the present application, the process of training the machine learning model using each error training sequence in step S13 is described in detail, and the steps are as follows:

[0092] S130. Input each error training sequence into the machine learning model in turn to obtain an error prediction value of the machine learning model.

[0093] Specifically, the latest machine learning model can be used to parse each error training sequence and generate error prediction values.

[0094] S131, comparing the error prediction value with the annotation label of the corresponding error training sequence, and calculating the loss value.

[0095] Specifically, the distance between the error prediction value and the annotation label of the corresponding error training sequence can be calculated as the loss value.

[0096] S132. Based on the loss value, adjust the parameters of the machine learning model.

[0097] Specifically, the hyperparameters of the machine learning model can be adjusted according to each loss value until multiple consecutive loss values ​​are less than a preset loss threshold.

[0098] It can be seen from the above technical solution that this embodiment provides an optional method for training a machine learning model. Through the above method, multiple error training sequences of different periods can be integrated to iteratively train the machine learning model, thereby improving the prediction reliability and prediction accuracy of the machine learning model, thereby improving the reliability of the present application in measuring the target gas component content in the internal gas of the substation.

[0099] In some embodiments of the present application, considering that the trained machine learning model can be tested from time to time to evaluate the training quality of the machine learning model and adjust the training direction in time to accelerate the training process, the process of testing the trained machine learning model can be added after the machine learning model is trained using each error training sequence. Next, the process will be described in detail, and the steps are as follows:

[0100] S133. Construct a test set based on the concentration test error of the target gas component in the gas concentration detection system when the internal gas of the substation is not introduced at different time points.

[0101] Specifically, the content of the target gas component in the gas concentration detection system when the gas inside the substation is not introduced can be collected at multiple time points over multiple days as the concentration test error;

[0102] The concentration test errors corresponding to N+1 consecutive time points in the same cycle can be selected, and the first N selected concentration test errors are selected and combined in sequence according to the time points to form test data, and the selected N+1th concentration test error is used as the annotation label of the test data to form an error test sequence.

[0103] The error test sequences are combined to form a test set.

[0104] The ratio between the total number of error test sequences and the total number of error training sequences may be 2:8.

[0105] S134. Evaluate the prediction accuracy of the trained machine learning model based on the test set. When the prediction accuracy does not meet the preset accuracy condition, return to the step of training the machine learning model using each error training sequence until the prediction accuracy of the trained machine learning model meets the accuracy condition.

[0106] Specifically, each error test sequence in the test set can be input into the trained machine learning model to obtain the prediction error corresponding to each error test sequence;

[0107] Compare each prediction error with the annotated label of its corresponding error test sequence to calculate the prediction accuracy of the trained machine learning model; wherein, based on the annotated label of each prediction error and its corresponding error test sequence, the mean square error of the test set can be calculated as the prediction accuracy of the trained machine learning model;

[0108] The mean square error calculation expression is as follows:

[0109]

[0110] Where M is the total number of error test sequences in the test set, k is the number of error test sequences, , is the prediction error, For labeling, the smaller the MSE, the higher the prediction accuracy.

[0111] When the prediction accuracy is less than the preset accuracy threshold, return to the step of training the machine learning model using each error training sequence.

[0112] When the prediction accuracy is greater than or equal to a preset accuracy threshold, the latest machine learning model is used as the error prediction model corresponding to the target gas component.

[0113] It can be seen from the above technical solution that compared with the previous embodiment, this embodiment adds an optional method of testing the machine learning model. Through the above method, the training progress of the machine learning model can be timely understood, the training direction of the machine learning model can be adjusted in time, and the training process of the machine learning model can be accelerated.

[0114] In some embodiments of the present application, step S2, the process of detecting the concentration information of the target gas component in the gas concentration detection system when the internal gas of the substation is not introduced, is described in detail, and the steps are as follows:

[0115] S20. In combination with non-orthogonal quantum measurement technology, when internal gas of the substation equipment is not introduced, the concentration information of the target gas component in the gas concentration detection system is detected.

[0116] Specifically, the gas concentration detection system can be as follows Figure 2 shown.

[0117] The initial polarization state of the signal light is modulated into a specific state through a pre-selection process, and the light beam is decomposed into two paths through a polarization beam splitter. After passing through a gas pool, a small light intensity loss is introduced due to gas absorption. After being recombined, a post-selection process is performed, that is, a specific polarization state is used for polarization detection, and the spectrum mean of the optical signal is measured. The offset is obtained by subtracting the initial light spectrum mean from the offset, and finally, the weak value amplification effect formula is used to obtain the concentration information of the target gas component in the gas concentration detection system when the internal gas of the substation equipment is not introduced.

[0118] It can be seen from the above technical solution that this embodiment provides an optional method for detecting the concentration information of the target gas component in the gas concentration detection system when the internal gas of the substation is not introduced. Through the above method, quantum technology can be combined with the measurement of gas component content to improve the measurement reliability of concentration information.

[0119] In some embodiments of the present application, the process of step S4, introducing the gas inside the power substation into the gas concentration detection system and measuring the concentration value of the target gas component is described in detail, and the steps are as follows:

[0120] S40, introducing the gas inside the power conversion equipment into the gas concentration detection system, and measuring the concentration value of the target gas component in combination with non-orthogonal quantum measurement technology.

[0121] Specifically, the quantum signal source is used as the light comb light source, and the pre-selected state is prepared through a polarizer. After the light beam passes through the pre-selection, the signal is divided into two identical beams through a beam splitter, one of which enters the gas pool as the signal light to interact with the gas inside the substation equipment introduced into the gas concentration detection system, and the other beam is used as the reference light and does not pass through the gas pool. The two beams are reunited for post-selection. Finally, we use detectors to detect signals and collect data.

[0122] Among them, the front and back selections meet the conditions:

[0123]

[0124] In the formula, represents the eigenstate of the gas concentration detection system, Represents a weak value, and They represent the pre-selection state and post-selection state of the gas concentration detection system, respectively. represents the Hilbert state vector.

[0125] The spectrum state of the light emitted by the light source is recorded as , where ω is the optical frequency, is the wave function related to the optical frequency. The light beam is modulated into the initial polarization state after the preselection ÷ , where H represents horizontal polarization and V represents vertical polarization. The initial state of the input light evolves as follows:

[0126]

[0127] The signal beam then passes through the gas pool. Since the gas absorbs light beams in a specific wavelength band, the absorption is very small and is usually not easy to measure without post-selection. This process is equivalent to the light undergoing unitary evolution. ,in, After evolution, the initial state evolves to:

[0128]

[0129] Then the beam passes through post-selection, and the post-selection state is ÷ , ε is a very small deflection angle, then the optical spectrum state eventually evolves to:

[0130]

[0131] The initial spectrum of light is , and the spectrum distribution of the light received by the detector is It can be seen that when the spectral state changes, the light signal received by the detector will have a relatively large offset compared to the light emitted by the light source. This offset is theoretically obtained through the weak value calculation formula. After substituting the pre-selection and post-selection, the formula for the weak value amplification effect is as follows:

[0132]

[0133] in, It indicates the deviation of the center frequency of the light received by the detector compared to the initial light. express The imaginary part of Represents the variance of the initial light frequency. After calculation, the light absorption estimate can be obtained:

[0134]

[0135] After obtaining the light absorption estimate, the concentration value c of the target gas component in the gas inside the substation can be calculated according to the Lambert-Beer law:

[0136]

[0137] in, is the gas absorption, is the molar absorption coefficient, c is the gas concentration, and l is the path length of light through the absorbing substance.

[0138] Next, we will combine Figure 3 The substation equipment quantum gas detection device provided in the present application is introduced in detail. The substation equipment quantum gas detection device provided below can be compared with the substation equipment quantum gas detection method provided above.

[0139] See also Figure 3 It can be found that the quantum gas detection device for substation equipment may include:

[0140] An acquisition module 10 is used to acquire an error prediction model corresponding to a target gas component, wherein the error prediction model is obtained based on concentration training errors of the target gas component detected by a gas concentration detection system at different time points;

[0141] The detection module 20 is used to detect the concentration information of the target gas component in the gas concentration detection system when the gas inside the substation is not introduced;

[0142] A prediction module 30, used for inputting the concentration information into the error prediction model to obtain a concentration prediction error of the target gas component;

[0143] The measuring module 40 is used to introduce the gas inside the power conversion equipment into the gas concentration detection system to measure the concentration value of the target gas component;

[0144] The calculation module 50 is used to calculate the content of the target gas component in the gas inside the substation by combining the concentration value and the concentration prediction error.

[0145] Further, the acquisition module 10 may include:

[0146] A machine learning model building unit, used for building a machine learning model for error prediction;

[0147] A concentration training error acquisition unit is used to acquire the concentration training error of the target gas component in the gas concentration detection system when the internal gas of the substation equipment is not introduced at different time points in a continuous cycle;

[0148] The error training sequence combination unit is used to combine the concentration training errors of the same period in sequence according to time points to form an error training sequence;

[0149] The machine learning model training unit is used to train the machine learning model using each error training sequence, and use the machine learning model finally obtained by training as the error prediction model corresponding to the target gas component.

[0150] Further, the error training sequence combination unit may include:

[0151] The time point selection component is used to select N+1 consecutive time points from the same cycle, and combine the concentration training errors corresponding to the first N selected time points in sequence according to the time points to form training data, and mark the concentration training error corresponding to the selected N+1th time point on the training data to form an error training sequence.

[0152] Furthermore, the machine learning model training unit may include:

[0153] A first machine learning model training component is used to input each error training sequence into the machine learning model in turn to obtain an error prediction value of the machine learning model;

[0154] A second machine learning model training component is used to compare the error prediction value with the annotation label of the corresponding error training sequence to calculate the loss value;

[0155] The third machine learning model training component is used to adjust the parameters of the machine learning model based on the loss value.

[0156] Furthermore, the machine learning model training unit may also include:

[0157] A test set building component, used to build a test set based on the concentration test error of the target gas component in the gas concentration detection system when the internal gas of the substation equipment is not introduced at different time points;

[0158] A prediction accuracy evaluation component is used to evaluate the prediction accuracy of the trained machine learning model based on the test set, and when the prediction accuracy does not meet the preset accuracy condition, call the first machine learning model training component until the prediction accuracy of the trained machine learning model meets the accuracy condition.

[0159] Further, the detection module 20 may include:

[0160] The concentration information detection unit is used to detect the concentration information of the target gas component in the gas concentration detection system in combination with the non-orthogonal quantum measurement technology when the internal gas of the substation equipment is not introduced.

[0161] Further, the measurement module 40 may include:

[0162] The concentration value measuring unit is used to introduce the gas inside the substation equipment into the gas concentration detection system, and measure the concentration value of the target gas component in combination with non-orthogonal quantum measurement technology.

[0163] The substation equipment quantum gas detection device provided in the embodiment of the present application can be applied to substation equipment quantum gas detection equipment, such as PC terminals, cloud platforms, servers and server clusters. Figure 4The hardware structure diagram of the quantum gas detection device for power substation equipment is shown. Figure 4 ,The hardware structure of the power conversion equipment quantum gas detection device may include: at least one processor 1, at least one communication interface 2, at least one memory 3 and at least one communication bus 4;

[0164] In the embodiment of the present application, the number of the processor 1, the communication interface 2, the memory 3, and the communication bus 4 is at least one, and the processor 1, the communication interface 2, and the memory 3 communicate with each other through the communication bus 4;

[0165] The processor 1 may be a central processing unit CPU, or an application-specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention, etc.;

[0166] The memory 3 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), etc., such as at least one disk memory;

[0167] The memory stores a program, and the processor can call the program stored in the memory, wherein the program is used to:

[0168] Obtaining an error prediction model corresponding to the target gas component, wherein the error prediction model is obtained based on concentration training errors of the target gas component detected by the gas concentration detection system at different time points;

[0169] When the gas inside the substation equipment is not introduced, detecting the concentration information of the target gas component in the gas concentration detection system;

[0170] Inputting the concentration information into the error prediction model to obtain the concentration prediction error of the target gas component;

[0171] Introducing the gas inside the power substation into the gas concentration detection system to measure the concentration value of the target gas component;

[0172] The content of the target gas component in the gas inside the substation is calculated based on the concentration value and the concentration prediction error.

[0173] Optionally, the detailed functions and extended functions of the program may refer to the above description.

[0174] The embodiment of the present application further provides a readable storage medium, which may store a program suitable for execution by a processor, wherein the program is used to:

[0175] Obtaining an error prediction model corresponding to the target gas component, wherein the error prediction model is obtained based on concentration training errors of the target gas component detected by the gas concentration detection system at different time points;

[0176] When the gas inside the substation equipment is not introduced, detecting the concentration information of the target gas component in the gas concentration detection system;

[0177] Inputting the concentration information into the error prediction model to obtain the concentration prediction error of the target gas component;

[0178] Introducing the gas inside the power substation into the gas concentration detection system to measure the concentration value of the target gas component;

[0179] The content of the target gas component in the gas inside the substation is calculated based on the concentration value and the concentration prediction error.

[0180] Optionally, the detailed functions and extended functions of the program may refer to the above description.

[0181] Finally, 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. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0182] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0183] The above description of the disclosed embodiments enables professionals and technicians in the field to implement or use the present application. Various modifications to these embodiments will be apparent to professionals and technicians in the field, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. The various embodiments of the present application can be combined with each other. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting quantum gas in power substation equipment, characterized in that: include: Obtaining an error prediction model corresponding to the target gas component, wherein the error prediction model is obtained based on concentration training errors of the target gas component detected by the gas concentration detection system at different time points; When the gas inside the substation equipment is not introduced, detecting the concentration information of the target gas component in the gas concentration detection system; Inputting the concentration information into the error prediction model to obtain the concentration prediction error of the target gas component; Introducing the gas inside the power substation into the gas concentration detection system to measure the concentration value of the target gas component; The content of the target gas component in the gas inside the substation is calculated based on the concentration value and the concentration prediction error.

2. The method for detecting quantum gas in power transformation equipment according to claim 1, characterized in that: The step of obtaining the error prediction model corresponding to the target gas composition includes: Build machine learning models for error prediction; The concentration training error of the target gas component in the gas concentration detection system is collected at different time points within a continuous cycle without introducing the internal gas of the substation equipment; The concentration training errors of the same cycle are combined in sequence according to the time points to form an error training sequence; The machine learning model is trained using each error training sequence, and the machine learning model finally obtained by training is used as the error prediction model corresponding to the target gas component.

3. The method for detecting quantum gas in power substation equipment according to claim 2, characterized in that: The step of combining the concentration training errors of the same cycle in sequence according to time points to form an error training sequence includes: N+1 consecutive time points are selected from the same cycle, and the concentration training errors corresponding to the first N selected time points are combined in sequence according to the time points to form training data, and the concentration training error corresponding to the selected N+1th time point is marked on the training data to form an error training sequence.

4. The method for detecting quantum gas in power transformation equipment according to claim 3, characterized in that: The step of training the machine learning model using each error training sequence includes: Inputting each error training sequence into the machine learning model in turn to obtain an error prediction value of the machine learning model; Compare the error prediction value with the label of the corresponding error training sequence to calculate the loss value; Based on the loss value, parameters of the machine learning model are adjusted.

5. The method for detecting quantum gas in power transformation equipment according to claim 2, characterized in that: After the machine learning model is trained using each error training sequence, the method further includes: A test set is constructed based on the concentration test error of the target gas component in the gas concentration detection system when the internal gas of the substation equipment is not introduced at different time points; The prediction accuracy of the trained machine learning model is evaluated based on the test set. When the prediction accuracy does not meet the preset accuracy condition, return to the step of training the machine learning model using each error training sequence until the prediction accuracy of the trained machine learning model meets the accuracy condition.

6. The method for detecting quantum gas in power transformation equipment according to claim 1, characterized in that: When the internal gas of the power transformation equipment is not introduced, detecting the concentration information of the target gas component in the gas concentration detection system includes: In combination with non-orthogonal quantum measurement technology, when the internal gas of the substation equipment is not introduced, the concentration information of the target gas component in the gas concentration detection system is detected.

7. The method for detecting quantum gas in power substation equipment according to claim 6, characterized in that: The step of introducing the gas inside the power conversion equipment into the gas concentration detection system and measuring the concentration value of the target gas component comprises: The gas inside the power conversion equipment is introduced into the gas concentration detection system, and the concentration value of the target gas component is measured in combination with non-orthogonal quantum measurement technology.

8. A quantum gas detection device for power substation equipment, characterized in that: include: An acquisition module, used to acquire an error prediction model corresponding to a target gas component, wherein the error prediction model is obtained based on concentration training errors of the target gas component detected by a gas concentration detection system at different time points; A detection module, used for detecting the concentration information of the target gas component in the gas concentration detection system when the gas inside the substation is not introduced; A prediction module, used for inputting the concentration information into the error prediction model to obtain a concentration prediction error of the target gas component; A measuring module, used for introducing the gas inside the power substation into the gas concentration detection system to measure the concentration value of the target gas component; A calculation module is used to calculate the content of the target gas component in the gas inside the substation by combining the concentration value and the concentration prediction error.

9. A quantum gas detection device for power substation equipment, characterized in that: including memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the power conversion equipment quantum gas detection method as described in any one of claims 1-7.

10. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the method for detecting quantum gas in a power conversion device as described in any one of claims 1 to 7 is implemented.

Images