Adaptive Control System for 10KV Outdoor Current Transformer Based on Optical Measurement

Abstract

The present invention discloses an adaptive control system for optical measurement of 10KV outdoor transformers, which relates to the technical field of power distribution networks. It is used to solve the problems of easy generation of light intensity fluctuations and demodulation distortion, resulting in errors in measurement output. By collecting the output voltage and current of electromagnetic transformers and the light intensity and phase changes of optical transformers, the Fourier-wavelet joint analysis is used to extract the main frequency, harmonics, and mutation characteristics to construct a fusion feature vector; then this feature is compared with historical normal samples, and the demodulation distortion result is output according to the set feature domain judgment mechanism; if there is an abnormality, combined with the supervision signal containing position information, the response of the virtual optical transformer is simulated, and a spatial weighted error compensation function is constructed to achieve multi-point difference fusion and dynamically correct the output of the current optical transformer, improving the stability, anti-interference ability, and measurement accuracy of the optical transformer in the intelligent grid measurement system.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution networks, and more specifically, to an adaptive control system for an optical measurement 10 kV outdoor mutual inductor. Background Art

[0002] With the development of distribution automation and smart grids, as a key measurement device, the accuracy, reliability, and dynamic response ability of 10 kV outdoor mutual inductors have an important impact on power quality monitoring and protection control. In recent years, optical mutual inductors (such as those based on the Faraday effect and Michelson interference principle) have gradually been applied to high-voltage measurement scenarios due to their advantages such as good insulation, high bandwidth, and strong electromagnetic compatibility.

[0003] The existing technology has the following deficiencies:

[0004] Currently, optical mutual inductors are relatively sensitive to factors such as environmental temperature, mechanical vibration, and material stress changes, and are prone to light intensity fluctuations and demodulation distortion, resulting in errors in measurement output and affecting the stability of the long-term stable operation of optical mutual inductors under harsh outdoor conditions. Therefore, an adaptive control system for an optical measurement 10 kV outdoor mutual inductor is proposed.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide an adaptive control system for an optical measurement 10 kV outdoor mutual inductor, which solves the problems proposed in the above background art through Fourier-wavelet joint analysis of optical signals, a demodulation distortion discrimination mechanism, and a spatial error compensation algorithm that integrates electromagnetic mutual inductor supervision signals.

[0007] To achieve the above object, the present invention provides the following technical solution: an adaptive control system for an optical measurement 10 kV outdoor mutual inductor, including an electromagnetic acquisition module, a parallel fusion module, a demodulation analysis module, and a complementary analog module, with signal connections between the modules:

[0008] The electromagnetic acquisition module is used to obtain the analog voltage and current output by the electromagnetic mutual inductor at the boundary of the preset optical mutual inductor measurement area according to the grid signal, denoted as the supervision signal, and collect the light intensity change and phase shift output by the optical mutual inductor;

[0009] The parallel fusion module is used to obtain the output result of the optical mutual inductor, substitute it into Fourier-wavelet joint processing, obtain the main frequency amplitude, harmonic structure, and local characteristics of the mutation point, and construct a fusion feature vector using parallel fusion;

[0010] The demodulation analysis module is used to obtain the fusion feature vector, collect the historical normal demodulation fusion feature vector, set up an optical current transformer feature domain judgment mechanism for evaluation, and determine the demodulation distortion discrimination result;

[0011] The complementary analog module is used to obtain the demodulation distortion result and the supervision signal containing position information, simulate the deployment of the virtual optical current transformer at the position information of the electromagnetic current transformer, and at the same time construct a position error compensation function to determine the optical current transformer correction amount by superimposing the multi-point differences.

[0012] In a preferred embodiment, through the distribution network GIS data and the installation topology map, combined with the layout coordinate point information of the optical current transformer and based on the electrical path and the sensor field of view range, the measurement area corresponding to the optical current transformer is determined;

[0013] For the measurement area corresponding to the optical current transformer, a plurality of electromagnetic current transformers are arranged at its boundary;

[0014] Connect its primary side in parallel to the three-phase bus of the power grid, and connect the secondary side to the analog quantity acquisition terminal, and use the electromagnetic induction principle to perform energy conversion to obtain the analog voltage and current output by the electromagnetic current transformer;

[0015] For the analog voltage and current output by the electromagnetic current transformer, record the position of each electromagnetic current transformer at the same time, and record it as the supervision signal.

[0016] In a preferred embodiment, according to the optical sensing principle, at the target measurement point, through the optical voltage sensor and the optical current sensor, the optical response characteristics are obtained, and the layout coordinate information of the corresponding optical current transformer is recorded to obtain the light intensity change and phase shift output by the optical current transformer.

[0017] In a preferred embodiment, the output result of the optical current transformer includes the light intensity change and the phase shift, which are converted into a time-domain signal in the form of voltage and current to form the time-domain signal of the optical current transformer;

[0018] The time-domain signal is processed in parallel in the Fourier channel and the wavelet channel;

[0019] In the Fourier channel, apply the discrete Fourier transform to the time-domain signal, extract the frequency-domain spectrum, and define the fundamental frequency component as the main frequency amplitude to obtain the frequency-domain characteristics;

[0020] The wavelet channel adopts a multi-scale wavelet packet decomposition strategy to perform multi-layer processing on the signal, extract the approximation and detail components at each scale, and construct a time-frequency local energy map to obtain the time-domain characteristics;

[0021] Fuse the frequency-domain and time-domain characteristics extracted from the Fourier and wavelet channels, construct a fusion feature vector, and standardize each component in the fusion feature vector.

[0022] In a preferred embodiment, by continuously collecting the demodulation output under normal operating conditions, a historical normal demodulation fusion feature vector is obtained;

[0023] The optical current transformer feature domain judgment mechanism is discriminated based on a support vector machine, and the specific steps are as follows:

[0024] S1: Take the currently collected fusion feature vector as the analysis feature input, and take the fusion feature vector under the normal demodulation state collected historically as the comparison feature input;

[0025] S2: Select a kernel function and calculate the kernel function result according to the analysis feature;

[0026] S3: Set an adjustment coefficient to converge the kernel function result and calculate the demodulation distortion discrimination threshold;

[0027] S4: Discriminate the demodulation accuracy of the optical current transformer;

[0028] S5: Output the demodulation distortion discrimination result.

[0029] In a preferred embodiment, in step S4, by setting an evaluation module in the support vector machine model, the accuracy of the output result of the support vector machine model is evaluated to determine the analysis difference.

[0030] In a preferred embodiment, A1: Construct a data set, arrange it according to the feature sequence, and combine the discrimination results into a discrimination data set;

[0031] A2: Set evaluation rules;

[0032] A3: Perform cross-validation and calculate the predicted value and the true value;

[0033] A4: Evaluate the accuracy, perform a ratio operation on the predicted value and the true value, and take the ratio as the accuracy;

[0034] A5: Determine the analysis difference, set an expected interval, and compare whether the accuracy is within the set expected interval to determine its accuracy level;

[0035] A6: Select a mode improvement model and use different mode improvement models according to the accuracy level.

[0036] In a preferred embodiment, the demodulation distortion result includes the light intensity change and phase shift output by the optical current transformer, the demodulated analog voltage and current, and their corresponding distortion state labels;

[0037] By simulating the output behavior of a virtual optical current transformer at the electromagnetic current transformer layout point, a response model of the electromagnetic current transformer coordinate point is constructed;

[0038] Construct the field strength estimation function of the entire area by using the multi-point sampling data provided by the electromagnetic mutual inductor, substitute the interpolated electric field into the optical mutual inductor response model, and obtain the simulated virtual output.

[0039] In a preferred embodiment, at all virtual points, that is, the deployment points of the electromagnetic mutual inductor, compare the predicted optical response values calculated with the output of the currently actually deployed optical mutual inductor, extract the systematic offset law, and construct the spatial position error compensation function;

[0040] Combine the virtual optical mutual inductor output values at the corresponding positions of the electromagnetic mutual inductor, calculate the difference between the virtual optical mutual inductor output value and the electromagnetic mutual inductor output value, and obtain the response difference between the electromagnetic mutual inductor and the virtual optical mutual inductor;

[0041] Based on the deployment position of the actual optical mutual inductor, generate the spatial fusion weight based on the distance relationship between each error point and the target position.

[0042] In a preferred embodiment, weight and superimpose all the response differences between the electromagnetic mutual inductor and the virtual optical mutual inductor according to the spatial fusion weight, and calculate the error compensation vector at the current position, including the voltage compensation value and the current compensation value;

[0043] Superimpose the voltage compensation value and the current compensation value on the original output of the current optical mutual inductor in real time to form the correction amount of the optical mutual inductor.

[0044] The technical effects and advantages of the present invention:

[0045] 1. By collecting the output voltage and current of the electromagnetic mutual inductor and the light intensity and phase changes of the optical mutual inductor, the present invention uses the Fourier-wavelet joint analysis to extract the main frequency, harmonics and mutation characteristics to construct the fusion feature vector; then compare this feature with the historical normal samples, and output the demodulation distortion result according to the set feature domain judgment mechanism; if there is an abnormality, combine the supervision signal containing position information, simulate the virtual optical mutual inductor response, and construct the spatial weighted error compensation function to achieve multi-point difference fusion and dynamically correct the output of the current optical mutual inductor, effectively identifying the possible demodulation distortion problems of the optical mutual inductor in a complex electromagnetic environment, and realizing the on-line dynamic compensation of its voltage and current output without affecting the structure of the optical sensing link, improving the stability, anti-interference ability and measurement accuracy of the optical mutual inductor in the intelligent power grid measurement system. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is the process flow chart of the adaptive control system of the 10KV outdoor mutual inductor based on optical measurement of the present invention.

[0047] Figure 2This is a schematic diagram of the modules of the adaptive control system for optically measuring 10KV outdoor transformers according to the present invention. Specific embodiments

[0048] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0049] Embodiment 1

[0050] The adaptive control system for optically measuring 10KV outdoor transformers, as Figures 1 to 2 shown, includes an electromagnetic acquisition module, a demodulation analysis module, a compensation simulation module, and a calibration control module, and the signals between the modules are connected:

[0051] The electromagnetic acquisition module is used to obtain the analog voltage and current output by the electromagnetic transformer at the boundary of the preset optical transformer measurement area according to the grid signal, denoted as the supervision signal, and collect the light intensity change and phase shift output by the optical transformer;

[0052] The grid signal refers to the actual power system signal that drives the operation of the entire transformer system and serves as the measurement basis reference quantity. In a 10kV medium-voltage distribution system, it is the power frequency AC signal during the transmission from the bus or feeder to the load, which includes three-phase AC voltage signals and three-phase AC current signals, etc., and will not be elaborated here;

[0053] Through the distribution network GIS data and the installation topology map, combined with the layout coordinate point information of the optical transformer (GPS positioning or relative coordinate system), and based on the electrical path and the sensor field of view, the measurement area corresponding to the optical transformer is determined;

[0054] For the measurement area corresponding to the optical transformer, a plurality of electromagnetic transformers are set at its boundary;

[0055] As can be seen from the above, the optical transformer is based on an optical voltage sensor (Faraday magneto-optical effect) and an optical current sensor (Pockels effect), uses no windings, optical fiber transmission, and optical elements, has high precision, a wide linear range, outputs light intensity change and phase shift, and restores physical quantities through demodulation calculation to obtain analog voltage and current;

[0056] The electromagnetic transformer is relatively traditional, and the secondary winding directly outputs analog voltage and current. Based on linear ratio conversion, the effective value, phase, etc. are directly measured, and will not be elaborated here;

[0057] The acquisition logic of the analog voltage and current output by the electromagnetic transformer is to connect its primary side in parallel to the three-phase bus of the power grid, and the secondary side is connected to the analog quantity acquisition terminal. The electromagnetic induction principle is used for energy conversion to obtain the analog voltage and current output by the electromagnetic transformer;

[0058] Specifically, the analog voltage and current output by the electromagnetic transformer are expressed as:

[0059] ;

[0060] In the formula, is the true signal of the primary side bus, is the voltage and current transformation ratio of the transformer, and is the measured secondary side analog signal, that is, the analog voltage and current output by the electromagnetic transformer;

[0061] For the analog voltage and current output by the electromagnetic transformer, the position of each electromagnetic transformer is recorded simultaneously and denoted as the supervision signal;

[0062] Specifically, the signal acquisition link is that the electromagnetic transformer is connected to the primary side of the power grid, and the analog quantity output by the secondary side is sampled through the acquisition terminal. For each position of the electromagnetic transformer, its output voltage, current and their spatial coordinate information are recorded;

[0063] The acquisition logic of the light intensity change and phase shift output by the optical transformer is based on the optical sensing principle. At the target measurement point, the optical voltage sensor and the optical current sensor are used to obtain the optical response characteristics, and the layout coordinate information of the corresponding optical transformer is recorded to obtain the light intensity change and phase shift output by the optical transformer;

[0064] It should be noted that the optical voltage sensor determines the rotation angle of the light polarization direction under the magnetic field caused by the measured voltage through the Faraday magneto-optical effect, so as to correspondingly reflect the voltage change. The optical current sensor determines the change in the refractive index in the nonlinear optical crystal under the action of the measured electric field strength through the Pockels effect, and then correspondingly reflects the current change, which will not be elaborated here;

[0065] Among them, the light intensity change amount reflects the change trend of the light energy caused by the electric field or magnetic field, and indirectly corresponds to the amplitude change of the voltage and current. The phase shift amount is the phase difference of the interference fringes, which is used to describe the periodic response caused by the measured current or voltage signal;

[0066] Furthermore, the Faraday effect is applied to the optical voltage sensor: when there is a voltage across the measured conductor, a corresponding electric field and induced magnetic field will be formed in the surrounding space. This magnetic field acts on the coated optical fiber or magneto-optical crystal, causing the rotation of the polarization plane in the light propagation path (the rotation angle is proportional to the voltage), and demodulating the polarization change in the output signal to restore the measured voltage;

[0067] Furthermore, the Pockels effect is applied to an optical current sensor. The electric field generated by the current flowing through the conductor acts on the nonlinear optical crystal, causing a change in its refractive index tensor, resulting in a change in the propagation light phase delay or interference fringes (the phase difference is proportional to the current). Demodulating this phase change can restore the measured current;

[0068] Among them, the calculation formulas for the light intensity change and phase shift output by the optical transformer are:

[0069] ;

[0070] In the formula, is a time-varying voltage signal, is the Faraday magneto-optical effect coefficient, is the light propagation path length, is the initial incident light intensity, is the light intensity change output by the optical transformer;

[0071] The above is applicable to the optical voltage sensor, and the voltage change trend is reflected by affecting the interference intensity through the light polarization rotation angle;

[0072] ;

[0073] In the formula, is the light wave wavelength, is the path length of the light propagating in the crystal, is the refractive index change caused by the electric field, is the phase shift output by the optical transformer;

[0074] Furthermore, , where, is the refractive index constant of the crystal without the action of the electric field, is the Pockels effect coefficient, is the time-varying electric field intensity, which is proportional to the current;

[0075] The above two formulas respectively reflect the response of the optical transformer to voltage and current signals from two dimensions of energy change and interference change;

[0076] The parallel fusion module is used to obtain the output result of the optical transformer, substitute it into the Fourier-wavelet joint processing, obtain the main frequency amplitude, harmonic structure and local characteristics of the mutation point, and construct a fusion feature vector by parallel fusion;

[0077] The output result of the optical transformer includes light intensity change and phase shift, which are converted into a time-domain signal in the form of voltage and current to form the time-domain signal of the optical transformer;

[0078] The time-domain signal is processed in parallel in the Fourier channel and the wavelet channel respectively:

[0079] In the Fourier channel, the discrete Fourier transform is applied to the time-domain signal to extract the frequency-domain spectrum. The fundamental frequency component is defined as the main frequency amplitude, and the amplitude of the m-th harmonic is (where is the spectral value corresponding to the m-th harmonic frequency). It is efficiently implemented using the fast Fourier transform:

[0080] ;

[0081] where N is the number of points of the fast Fourier transform, and the main frequency amplitude and the amplitudes of several previous harmonics are extracted to form the frequency-domain feature vector;

[0082] It should be noted that the fast Fourier transform is based on the divide-and-conquer strategy, which reduces the computational complexity from to, making it suitable for spectrum analysis in real-time or big data scenarios and ensuring efficient acquisition of frequency-domain information. This will not be elaborated here;

[0083] The wavelet channel adopts the multi-scale wavelet packet decomposition strategy to perform multi-layer processing on the signal, extract the approximation and detail components at each scale, and construct the time-frequency local energy map; the multi-layer wavelet packet decomposition is completed using the orthogonal wavelet basis, and the formula is as follows:

[0084] ;

[0085] In the formula, represents the -th wavelet packet coefficient in the -th layer of decomposition;

[0086] The energy of each sub-band is calculated by the following formula:

[0087] ; [[ID=�6]]

[0088] where is the sample point index;

[0089] By selecting the wavelet basis to perform multi-scale decomposition on the signal, a set of detail coefficients at different scales is obtained. Based on the local maxima and threshold processing of the detail coefficients, the mutation points and their local characteristic quantities in the signal are identified. The local statistical quantity of the wavelet coefficients is used as the time-domain mutation feature. The feature quantities extracted by the Fourier channel and the wavelet channel clearly define the spectral structure and mutation structure of the signal;

[0090] It should be noted that the local maxima and threshold processing are the key means to judge the abnormal signal structure in mutation point detection. Its basic logic is that mutation points often show a sharp increase in the modulus of local coefficients in the wavelet domain, that is, there are modulus maximum points in a local area at a certain scale, while the wavelet coefficients in the normal and stable regions change relatively smoothly; <@

[0091] Furthermore, kurtosis and skewness analysis can be performed on the signals reconstructed at each scale to assist in determining whether abnormal disturbances are caused by factors such as transient interference, harmonic distortion, or demodulation distortion;

[0092] It should be noted that the wavelet basis function selected during the wavelet packet decomposition has a crucial impact on the feature retention ability. To improve the ability to retain transient features such as spikes, edges, and short-time oscillations in the signal, the specific selection can be optimized according to the sampling signal bandwidth and resolution requirements, which will not be elaborated here;

[0093] During this process, a wavelet basis with good transient response ability can be selected to retain non-stationary features such as mutation edges, spike distortions, and short-time disturbances in the voltage / current signal;

[0094] Fuse the frequency-domain and time-domain features extracted from the Fourier and wavelet channels to construct a fused feature vector, which is specifically defined as follows:

[0095] ;

[0096] In the formula, is the frequency-domain feature extracted from the Fourier channel, is the time-domain feature extracted from the wavelet channel;

[0097] Among them, parallel fusion means directly combining the two features into a feature vector without weighting or other operations, retaining the independence of their respective features, enabling the model to utilize the information in both the frequency domain and the time domain simultaneously, which will not be elaborated here;

[0098] Standardize each component in the fused feature vector. Among them, the methods of standardization include but are not limited to standard linear transformation based on interval scaling, Z-Score standardization method based on statistics, or normalization method based on non-linear mapping function. The application methods of standardization will not be elaborated here;

[0099] Among them, the fused feature vector after preprocessing comprehensively characterizes the spectrum and mutation characteristics of the output signal of the optical current transformer, which will not be elaborated here;

[0100] The demodulation analysis module is used to obtain the fused feature vector, collect the historical normal demodulation fused feature vector, set up an optical current transformer feature domain judgment mechanism for evaluation, and determine the demodulation distortion discrimination result;

[0101] The acquisition logic of the historical normal demodulation fused feature vector is to obtain the historical normal demodulation fused feature vector by continuously collecting the demodulation output under the normal operating state of demodulation;

[0102] It should be noted that the feature extraction algorithm is consistent with the current demodulation processing flow to ensure the consistency of feature dimensions and distributions. By performing sliding window analysis and statistical processing on the fusion feature vectors within multiple consecutive demodulation cycles in this stage, the system generates a set of representative feature distribution models, which can be used as a reference for the historical normal state and will not be elaborated here;

[0103] The feature domain judgment mechanism of the optical current transformer is discriminated based on the support vector machine, and the specific steps are as follows:

[0104] S1: Take the currently collected fusion feature vector as the analysis feature input, and take the fusion feature vector in the normal demodulation state collected historically as the comparison feature input;

[0105] Among them, the fusion feature vector is composed of the parallel fusion of the Fourier channel frequency domain feature and the wavelet channel time domain feature, and the comparison feature input is used as a reference baseline;

[0106] S2: Select the Sigmoid function as the kernel function to perform feature transformation on the input features. Through the Sigmoid kernel function formula, the specific expression is:

[0107] ;

[0108] In the formula, is the result of the kernel function after feature transformation, is the hyperbolic tangent function, is a hyperparameter used to control the width of the kernel function, is a constant used to control the offset of the kernel function, and are the two input features in this example, namely the current analysis feature and the historical comparison feature;

[0109] S3: Set the adjustment coefficient k, and converge the result of the kernel function according to the adjustment coefficient. Its convergence formula can be:

[0110] [[ID=3८]]

[0111] In the formula, is the adjustment coefficient, is the result of the kernel function after convergence, and the result of the kernel function after convergence is used as the demodulation distortion discrimination threshold;

[0112] S4: According to the corresponding data of each fusion feature vector in the analysis feature, calculate the result of the kernel function after feature transformation; compare the calculated result of the kernel function with the demodulation distortion discrimination threshold. If the calculated result of the kernel function is higher than the discrimination threshold, mark the corresponding fusion feature vector as 0, otherwise, mark it as 1;

[0113] S5: Output the demodulation distortion discrimination result. When the fused feature vector is marked as 0, the demodulation distortion discrimination result is normal demodulation, indicating that the output signal of the current optical current transformer does not show distortion. When the fused feature vector is marked as 1, the demodulation distortion discrimination result is demodulation distortion, indicating that the output signal of the current optical current transformer has distortion characteristics, and output the corresponding demodulation distortion discrimination result;

[0114] By using a support vector machine to comprehensively fuse the frequency domain features and time domain mutation features in the feature vector, the demodulation state of the output signal of the optical current transformer is discriminated, which greatly improves the accuracy and robustness of the demodulation distortion detection; By setting the adjustment coefficient method, it is convenient to optimize and adjust the sensitivity and generalization ability of the discrimination model in a targeted manner in the future;

[0115] It should be noted that the role of the Sigmoid function is to convert the input features into data that are easy to analyze and process. The selection of the kernel function and loss function in the above steps is not unique, and the hyperparameters and constants can be set according to actual situations such as demodulation discrimination requirements. For example, set the hyperparameter to 0.8, the constant to 0.2, and preset the adjustment coefficient to 1, etc. The preset adjustment coefficient is analyzed and adjusted by step S4, which will not be elaborated here;

[0116] In S4, an evaluation module is set in the support vector machine model to evaluate the accuracy of the output result of the support vector machine model, determine the analysis difference, and select different modes to improve the model according to the accuracy. The specific steps are as follows:

[0117] A1: Correspondingly match the discrimination output result of the support vector machine model for the current fused feature vector with the control fused feature result in the historical demodulation state, arrange them according to the feature sequence, form a discrimination data set with the matched discrimination results, and form a control data set with the corresponding historical results for subsequent error analysis;

[0118] A2: Set the evaluation rule, compare the discrimination data set and the control data set according to the feature sequence. If the discrimination value and the control value of the current sample are both 1 (both judged as distorted), it is marked as true distortion, corresponding to low positioning accuracy. If the discrimination value and the control value of the current sample are both 0 (both judged as normal), it is marked as true normal, corresponding to high positioning accuracy. If the discrimination value is 1 and the control value is 0, it is marked as false distortion, corresponding to false high accuracy. If the discrimination value is 0 and the control value is 1, it is marked as false normal, corresponding to false low accuracy;

[0119] A3: Cross-validation. Select k-fold cross-validation. Randomly divide the discriminant dataset and the control dataset into k subsets of equal size, and correspond them one by one according to the feature sequence. Each time, select one subset as the test set, and the remaining k - 1 subsets as the training set. Count the true distorted and true normal samples in the test set and sum them up. Divide the sum result by the total number of labels in the test set to obtain the test ratio. Similarly, obtain the training ratio for the training set. Repeat cross-validation, and calculate the average values of the obtained test ratio and training ratio respectively to get the predicted value and the true value.

[0120] A4: Evaluate the accuracy. Calculate the ratio of the predicted value to the true value, and use this ratio as the accuracy. It should be noted that the closer the predicted value is to the true value, the more accurate the prediction result is, that is, the smaller the ratio of the predicted value to the true value is, the more accurate the prediction result is.

[0121] A5: Determine the analysis difference. Set the expected interval, and compare whether the accuracy is within the set expected interval. When the accuracy is within the expected interval, it is judged that the model accuracy is high; when the accuracy is outside the expected interval, it is judged that the model accuracy is low.

[0122] A6: Select a mode to improve the model. If it is judged that the model accuracy is high, then save the construction mechanism of the support vector machine model, that is, save the construction process of the support vector machine model this time; if it is judged that the model accuracy is low, then set the correction parameter , and correct the adjustment coefficient according to the correction parameter:

[0123]

[0124] Among them, is the uncorrected adjustment coefficient, the uncorrected adjustment coefficient. The adjustment coefficient is corrected by the randomly generated correction parameter . Repeat the running process of the support vector machine model until the accuracy calculated in the evaluation module is within the expected interval, then the correction ends, and the corrected adjustment coefficient is retained.

[0125] It should be noted that the expected interval in the evaluation module can be set according to the expected result. For example, set the expected interval to [0.7, 1.5]. The random generation of the correction parameter can also delimit the interval. For example, randomly generate the correction parameter within the interval of [0.5, 1.8], so as to reduce the operation time and improve the model running efficiency.

[0126] The complementary analog module is used to obtain the demodulation distortion result and the supervision signal containing position information, simulate the deployment of the virtual optical current transformer at the position of the electromagnetic current transformer, and construct a position error compensation function to determine the correction amount of the optical current transformer by superimposing the multi-point differences.

[0127] The demodulation distortion result includes the analog voltage, current and their corresponding distortion state labels obtained by demodulating the light intensity change and phase shift output by the optical current transformer.

[0128] By simulating the output behavior of the virtual optical current transformer at the installation point of the electromagnetic current transformer, an error compensation model is constructed to realize the correction of the optical current transformer.

[0129] It can be understood that since it is impossible to deploy the actual optical current transformer at the positions of all electromagnetic current transformers, it is necessary to "construct" the expected output behavior of the optical current transformer at these positions in the simulation space, that is, to establish a response model of a "virtual optical current transformer" at the coordinate points of the electromagnetic current transformer.

[0130] Construct the optical current transformer response model, and assume that the output response of the optical current transformer satisfies the following physical model:

[0131] ;

[0132] In the formula, is the optical voltage output at time t and position P, is the electric field strength at time t and position P, is the noise component, , and are the response coefficients of the current transformer obtained by calibration;

[0133] Use the multi-point sampling data provided by the electromagnetic current transformer to construct the field strength estimation function of the entire area:

[0134] ;

[0135] In the formula, is the real-time electric field data of the th electromagnetic current transformer, is the interpolation weight of the th point pair to the position ;

[0136] ;

[0137] Among them, usually take , is the stability constant;

[0138] Substitute the interpolated electric field into the optical current transformer response model to obtain the simulated virtual output:

[0139] ;

[0140] Similarly, the current output simulation can be obtained:

[0141] ;

[0142] function represents the non - linear or near - linear response function of the optical current transformer, which can be obtained from historical calibration;

[0143] It should be noted that the simulated virtual optical transformer is deployed at the location of the electromagnetic transformer. By constructing a spatial field strength distribution function based on the data of multiple - point electromagnetic transformers and calling the response simulation model of the optical transformer, the predicted values of voltage and current outputs at the target point are obtained, thus completing the fitting modeling of the "virtual" optical output at this point;

[0144] Compare the predicted values of the optical responses calculated at all virtual points (i.e., the deployment points of the electromagnetic transformers) with the outputs of the currently actually deployed optical transformers, extract the systematic offset law, and construct a spatial position error compensation function;

[0145] Furthermore, to achieve dynamic correction of the output error of the optical transformer, a position error compensation function based on spatial distribution error is constructed. By weighted fusion of the response differences between the electromagnetic transformers and the virtual optical transformers deployed at multiple points, the required output correction amount at the current deployment point of the optical transformer is calculated;

[0146] Among them, the acquisition logic of the response difference between the electromagnetic transformer and the virtual optical transformer is to combine the output value of the virtual optical transformer at the corresponding position of the electromagnetic transformer, calculate the difference between the output value of the virtual optical transformer and the output value of the electromagnetic transformer, and obtain the response difference between the electromagnetic transformer and the virtual optical transformer;

[0147] Among them, the voltage and current sampling results at the positions of multiple electromagnetic transformers are obtained, and the output values of the virtual optical transformers at the corresponding positions are simulated by combining the optical transformer response model to form the response difference between the electromagnetic transformer and the virtual optical transformer, which will not be elaborated here;

[0148] It should be noted that both the output value of the virtual optical transformer and the output value of the electromagnetic transformer are the simulated voltage and simulated current of their outputs;

[0149] Based on the deployment position of the actual optical transformer, a spatial fusion weight is generated according to the distance relationship between each error point and the target position;

[0150] The response differences between all electromagnetic transformers and virtual optical transformers are weighted and superimposed according to the spatial fusion weights to calculate the error compensation vector at the current position, including the voltage compensation value and the current compensation value;

[0151] It should be noted that weighted superposition is common knowledge for the experimenters and will not be elaborated here;

[0152] The voltage compensation value and the current compensation value are superimposed on the original output of the current optical transformer in real time to form the optical transformer correction amount;

[0153] The above formulas are all dimensionless and take their numerical calculations. The formulas are obtained by collecting a large amount of data for software simulation to obtain a formula closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0154] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired or wireless (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server or data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0155] It should be understood that in various embodiments of the present application, the magnitudes of the serial numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0156] Those of ordinary skill in the art will realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. A professional technician can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of this application.

[0157] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0158] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be electrical, mechanical, or other forms.

[0159] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0160] In addition, the functional units in the various embodiments of this application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit.

[0161] When the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0162] As described above, the above are only specific embodiments of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered by the protection scope of this application. Therefore, the protection scope of this application shall be subject to the protection scope of the claims.

Claims

1. An adaptive control system for optically measuring 10KV outdoor instrument transformers, characterized in that: It includes an electromagnetic acquisition module, a parallel fusion module, a demodulation analysis module, and a complementary analog module. The signals are connected between the modules: The electromagnetic acquisition module is used to obtain the analog voltage and current output by the electromagnetic transformer at the boundary of the preset optical transformer measurement area according to the grid signal, which is recorded as the supervision signal, and collect the light intensity change and phase shift output by the optical transformer. The parallel fusion module is used to obtain the output result of the optical transformer, substitute it into the joint Fourier-wavelet processing, obtain the main frequency amplitude, harmonic structure, and local characteristics of the mutation point, and construct a fusion feature vector by parallel fusion. The demodulation analysis module is used to obtain the fusion feature vector, collect the historical normal demodulation fusion feature vector, set up an optical transformer feature domain judgment mechanism for evaluation, determine the demodulation distortion discrimination result, and obtain the demodulation distortion result. The complementary analog module is used to obtain the demodulation distortion result and the supervision signal containing position information, simulate the deployment of the virtual optical transformer at the position information of the electromagnetic transformer, and at the same time construct a position error compensation function to superimpose the multi-point differences to determine the correction amount of the optical transformer. At all virtual points, that is, the deployment points of the electromagnetic transformers, compare the predicted optical response values calculated with the output of the currently actual deployed optical transformer to extract the systematic offset law and construct a spatial position error compensation function. Combine the output values of the virtual optical transformers at the corresponding positions of the electromagnetic transformers, calculate the difference between the output values of the virtual optical transformers and the output values of the electromagnetic transformers to obtain the response difference between the electromagnetic transformers and the virtual optical transformers. Weight and superimpose all the response differences between the electromagnetic transformers and the virtual optical transformers according to the spatial fusion weight, and calculate the error compensation vector at the current position, including the voltage compensation value and the current compensation value. Superimpose the voltage compensation value and the current compensation value on the original output of the current optical transformer in real time to form the correction amount of the optical transformer.

2. The adaptive control system for optically measuring a 10 kV outdoor instrument transformer according to claim 1, characterized in that: Through the distribution network GIS data and the installation topology diagram, combine the layout coordinate point information of the optical transformer and based on the electrical path and the sensor field of view range, determine the measurement area corresponding to the optical transformer; For the measurement area corresponding to the optical transformer, set multiple electromagnetic transformers at its boundary; Connect its primary side in parallel to the three-phase bus of the power grid, and connect the secondary side to the analog quantity acquisition end, and use the electromagnetic induction principle for energy conversion to obtain the analog voltage and current output by the electromagnetic transformer; For the analog voltage and current output by the electromagnetic transformer, record the position of each electromagnetic transformer at the same time and record it as the supervision signal.

3. The adaptive control system for optically measuring a 10 kV outdoor mutual inductor according to claim 2, characterized in that: According to the optical sensing principle, at the target measurement point, obtain the optical response characteristics through the optical voltage sensor and the optical current sensor, and record the layout coordinate information of the corresponding optical transformer to obtain the light intensity change and phase shift output by the optical transformer.

4. The adaptive control system for optically measuring 10KV outdoor instrument transformers according to claim 3, characterized in that: The output result of the optical transformer includes the light intensity change and the phase shift, which are converted into a time-domain signal in the form of voltage and current to form the time-domain signal of the optical transformer; Parallel process the time-domain signal in the Fourier channel and the wavelet channel respectively; In the Fourier channel, apply the discrete Fourier transform to the time-domain signal, extract the frequency-domain spectrum, and define the fundamental frequency component as the main frequency amplitude to obtain the frequency-domain characteristics; The wavelet channel adopts a multi-scale wavelet packet decomposition strategy to perform multi-layer processing on the signal, extract the approximation and detail components at each scale, construct a time-frequency local energy spectrum, and obtain time-domain features; Fuse the frequency-domain and time-domain features extracted from the Fourier and wavelet channels, construct a fusion feature vector, and standardize each component in the fusion feature vector.

5. The adaptive control system for optically measuring a 10 kV outdoor instrument transformer according to claim 4, characterized in that: By continuously collecting the demodulation output under the normal operating state, obtain the historical normal demodulation fusion feature vector; The optical current transformer feature domain judgment mechanism is based on a support vector machine for discrimination, and the specific steps are as follows: S1: Take the currently collected fusion feature vector as the analysis feature input, and take the fusion feature vector under the normal demodulation state collected historically as the comparison feature input; S2: Select a kernel function to calculate the kernel function result according to the analysis feature; S3: Set an adjustment coefficient to converge the kernel function result and calculate the demodulation distortion discrimination threshold; S4: Discriminate the demodulation accuracy of the optical current transformer; S5: Output the demodulation distortion discrimination result.

6. The adaptive control system for optically measuring 10KV outdoor instrument transformers according to claim 5, characterized in that: In step S4, by setting an evaluation module in the support vector machine model, evaluate the accuracy of the output result of the support vector machine model to determine the analysis difference.

7. The adaptive control system for optically measuring 10KV outdoor instrument transformers according to claim 6, characterized in that: A1: Construct a data set, arrange it according to the feature sequence, and merge the discrimination results into a discrimination data set; A2: Set evaluation rules; A3: Perform cross-validation and calculate the predicted value and the true value; A4: Evaluate the accuracy, perform a ratio operation on the predicted value and the true value, and use the ratio as the accuracy; A5: Determine the analysis difference, set an expected interval, and compare whether the accuracy is within the set expected interval to determine its accuracy; A6: Select a mode improvement model and use different mode improvement models according to the accuracy.

8. The adaptive control system for optically measuring a 10 kV outdoor mutual inductor according to claim 1, characterized in that: The demodulation distortion result includes the analog voltage and current after demodulating the optical intensity change and phase shift output by the optical current transformer, and their corresponding distortion state labels; By simulating the output behavior of the virtual optical current transformer at the electromagnetic current transformer layout point, construct a response model for the electromagnetic current transformer coordinate point; Use the multi-point sampling data provided by the electromagnetic current transformer to construct a field strength estimation function for the entire area, substitute the interpolated electric field into the optical current transformer response model to obtain a simulated virtual output.

9. The adaptive control system based on an optical measurement 10KV outdoor current transformer according to claim 8, characterized in that: Based on the distance relationship between each error point and the target position according to the deployment position of the actual optical current transformer, generate a spatial fusion weight.

Images