Box-type substation operation state monitoring method and system facing photovoltaic field area

Abstract

This invention provides a method and system for monitoring the operational status of prefabricated substations in photovoltaic power plants, belonging to the field of substation monitoring technology. The method includes: real-time acquisition of raw state signals during the operation of the prefabricated substation; decomposition of the raw state signals to obtain multiple intrinsic mode function (IMF) components; calculation of the multi-scale permutation entropy of each IMF component; construction of an entropy-energy two-dimensional feature matrix; identification of the operational status of the prefabricated substation; and output of the status monitoring results. This invention solves the technical problem of poor reliability in identifying complex non-stationary operating states of prefabricated substations due to existing technologies relying on simple thresholds or single-scale analysis for status monitoring. By decomposing the state signals and constructing a joint feature matrix of multi-scale permutation entropy and energy, the invention achieves highly sensitive anomaly identification and reliable monitoring of non-stationary operating states of prefabricated substations.

Description

Technical Field

[0001] This invention relates to the field of substation monitoring technology, specifically to a method and system for monitoring the operational status of prefabricated substations in photovoltaic power plants. Background Technology

[0002] In the photovoltaic power generation process, the prefabricated substation, as a key piece of equipment for realizing power collection and voltage boosting transmission, directly affects the power supply reliability and operational safety of the entire photovoltaic field. Due to the intermittent and fluctuating nature of photovoltaic output, prefabricated substations operate in an environment with frequent load fluctuations and complex and changeable operating conditions, making them prone to abnormal operation.

[0003] In existing technologies, a single electrical parameter or physical quantity is typically used as a monitoring feature. Its time-domain statistics or instantaneous amplitude characteristics, such as mean, effective value, or peak value, are then extracted as monitoring indicators. A fixed threshold range is pre-set, and an abnormal state is determined when the real-time monitored value exceeds or falls below this threshold. Alternatively, single-scale signal analysis methods are used, performing single-time-window moving average or single-frequency band filtering analysis on the monitoring features. Feature changes are extracted only within a single time scale or frequency range, and the state is judged based on the trend of this single-scale feature. However, because the operating signals of prefabricated substations exhibit significant nonlinearity, non-stationarity, and multi-source coupling characteristics, a single feature cannot comprehensively reflect the overall operating status of the equipment. Furthermore, simple threshold or single-scale analysis methods are insensitive to weak abnormal changes and are easily affected by noise and operating condition fluctuations, resulting in poor reliability of state identification and the risk of false alarms or missed alarms. This makes it difficult to meet the needs of photovoltaic fields for refined and reliable equipment monitoring.

[0004] In summary, existing technologies suffer from poor reliability in identifying complex and non-stationary operating conditions of prefabricated substations due to their reliance on simple threshold or single-scale analysis for condition monitoring. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for monitoring the operating status of prefabricated substations in photovoltaic fields, in order to solve the technical problem that existing technologies rely on simple threshold or single-scale analysis for status monitoring, resulting in poor reliability in identifying complex and non-stationary operating states of prefabricated substations.

[0006] In view of the above problems, this application provides a method and system for monitoring the operation status of prefabricated substations in photovoltaic fields.

[0007] The first aspect of this application provides a method for monitoring the operational status of prefabricated substations in photovoltaic power plants. This method includes: real-time acquisition of raw state signals during the operation of the prefabricated substation, the raw state signals including electrical and non-electrical parameter signals of the prefabricated substation; decomposition of the raw state signals to obtain multiple intrinsic mode function (IMF) components; calculation of the multi-scale permutation entropy of each IMF component; construction of an entropy-energy two-dimensional feature matrix; identification of the operational status of the prefabricated substation; and output of the status monitoring results.

[0008] Optionally, adaptive white noise of different amplitudes is added multiple times to the original state signal to form multiple noisy signals; empirical mode decomposition is performed on each of the multiple noisy signals to extract their respective first-order intrinsic mode function components, and the average value is calculated as the first-order intrinsic mode function component; the first-order intrinsic mode function component is subtracted from the original state signal to obtain the first-order residual signal; adaptive white noise is added again to the first-order residual signal, and empirical mode decomposition and averaging are repeated to extract the next-order intrinsic mode function component until the stopping condition is met, resulting in the multiple intrinsic mode function components arranged in descending order of frequency.

[0009] Optionally, the stopping condition is that the residual signal becomes a monotonic function or reaches a preset number of decomposition layers.

[0010] Optionally, all local maxima and local minima of each noisy signal are extracted, and upper and lower envelopes are constructed using curve fitting. The mean curves of the upper and lower envelopes are calculated, and the corresponding mean curves are subtracted from the multiple noisy signals to obtain intermediate signals. It is determined whether the intermediate signal satisfies the waveform conditions of the intrinsic mode function. If it does, it is output as the first-order intrinsic mode function component corresponding to each noisy signal. If it does not, the filtering process is repeated for the intermediate signal until the waveform conditions of the intrinsic mode function are met.

[0011] Optionally, for any component among the multiple intrinsic mode function components, coarsening is performed according to multiple different scale factors to generate a multi-scale coarsened sequence; the phase space of the multi-scale coarsened sequence is reconstructed at each scale to reconstruct the one-dimensional sequence into a multi-dimensional sequence; the data in each dimension after reconstruction are sorted by size, and the probability of each sorting pattern is calculated; the permutation entropy value at each scale is calculated based on the probability of each sorting pattern; the permutation entropy value at each scale is averaged to obtain the multi-scale permutation entropy value of any component.

[0012] Optionally, the sum of squares of the component sequences of the plurality of intrinsic mode function components is calculated as the energy value; the energy value is paired with the multi-scale permutation entropy values ​​corresponding to the plurality of intrinsic mode function components to establish the entropy-energy two-dimensional feature matrix.

[0013] Optionally, when the prefabricated substation is in a healthy state, a normal state reference feature matrix is ​​collected; the deviation index between the entropy-energy two-dimensional feature matrix and the normal state reference feature matrix is ​​calculated; when the deviation index exceeds a preset alarm threshold, the prefabricated substation is determined to be in an abnormal operating state and an early warning signal is issued.

[0014] A second aspect of this application provides a monitoring system for the operation status of a prefabricated substation in a photovoltaic (PV) field. The system includes: a signal acquisition module for real-time acquisition of raw state signals during the operation of the prefabricated substation, the raw state signals including electrical and non-electrical parameter signals of the prefabricated substation; a signal decomposition module for decomposing the raw state signals to obtain multiple intrinsic mode function (IMF) components; and a state identification module for calculating the multi-scale permutation entropy of the multiple IMF components, constructing an entropy-energy two-dimensional feature matrix, identifying the operation status of the prefabricated substation, and outputting the state monitoring results.

[0015] One or more technical solutions provided in this application have at least the following technical effects or advantages: The method provided in this application acquires raw state signals during the real-time operation of a prefabricated substation. These raw state signals include electrical and non-electrical parameter signals. The raw state signals are decomposed to obtain multiple intrinsic mode function (IMF) components. The multi-scale permutation entropy of each IMF component is calculated, and an entropy-energy two-dimensional feature matrix is ​​constructed to identify the operating state of the prefabricated substation and output the state monitoring results. This achieves the technical effect of highly sensitive anomaly identification and reliable monitoring of non-stationary operating states of prefabricated substations by decomposing the state signals and constructing a joint feature matrix of multi-scale permutation entropy and energy.

[0016] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0018] Figure 1 A flowchart illustrating the method for monitoring the operational status of prefabricated substations in photovoltaic power plants provided in this application.

[0019] Figure 2 A schematic diagram of the operation status monitoring system for a prefabricated substation in a photovoltaic field provided in this application.

[0020] Explanation of reference numerals in the attached diagram: Signal acquisition module 11, Signal decomposition module 12, Status recognition module 13. Detailed Implementation

[0021] This application provides a method and system for monitoring the operational status of prefabricated substations in photovoltaic power plants. It addresses the technical problem of poor reliability in identifying complex, non-stationary operating states of prefabricated substations due to existing technologies relying on simple thresholds or single-scale analysis. The method achieves highly sensitive anomaly identification and reliable monitoring of non-stationary operating states of prefabricated substations by decomposing the state signals and constructing multi-scale permutation entropy and energy joint features.

[0022] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all of them.

[0023] Example 1, as Figure 1 As shown, this application provides a method for monitoring the operational status of a prefabricated substation in a photovoltaic power plant area. The method includes: The raw status signals of the prefabricated substation during operation are collected in real time. The raw status signals include electrical parameter signals and non-electrical parameter signals of the prefabricated substation.

[0024] Specifically, by installing voltage transformers and current transformers on the high-voltage and low-voltage sides of the prefabricated substation respectively, the bus voltage and circuit current are synchronously sampled to obtain electrical parameter signals during the operation of the prefabricated substation. At the same time, vibration sensors and temperature sensors, such as triaxial accelerometers and PT100 platinum resistance thermometers, are installed on the transformer body, the enclosure structure and key heat dissipation parts to collect non-electrical parameter signals during the operation of the prefabricated substation.

[0025] After the electrical and non-electrical parameter signals of the prefabricated substation are uniformly timestamped and synchronized, an original state signal is formed. The original state signal includes electrical parameter signals and non-electrical parameter signals. The electrical parameter signals include at least the voltage and current signals of the prefabricated substation, and the non-electrical parameter signals include at least one of the vibration and temperature signals of the prefabricated substation.

[0026] The original state signal is decomposed to obtain multiple intrinsic mode function components.

[0027] Furthermore, the original state signal is decomposed to obtain multiple intrinsic mode function (IMF) components, including: adding adaptive white noise of different amplitudes to the original state signal multiple times to form multiple noisy signals; performing empirical mode decomposition on each of the multiple noisy signals to extract their respective first-order IMF components, calculating the average value as the first-order IMF component; subtracting the first-order IMF component from the original state signal to obtain a first-order residual signal; adding adaptive white noise again to the first-order residual signal, repeating the empirical mode decomposition and averaging process, extracting the next-order IMF component, until a stopping condition is met, resulting in the multiple IMF components arranged in descending order of frequency.

[0028] Furthermore, the stopping condition is that the residual signal becomes a monotonic function or reaches a preset number of decomposition layers.

[0029] Specifically, the original state signal is denoted as x(t). Adaptive white noise of different amplitudes is superimposed multiple times onto x(t). The adaptive white noise is defined as the standard deviation of the white noise added each time being set to 0.1 to 0.3 times the standard deviation of the original state signal, thus generating multiple sets of noisy signals xi(t) = x(t) + ni(t), where ni(t) represents the i-th added white noise sequence. The spectral uniformity of the white noise is used to suppress mode aliasing. Then, Empirical Mode Decomposition (EMD) is performed on each noisy signal to obtain its respective first-order intrinsic mode function (IMF1(t)). The first-order IMF1(t) refers to the highest-frequency and fastest-changing oscillating component in the original state signal, typically containing high-frequency dynamic information generated during equipment operation. The average value of the multiple first-order IMF1(t) is calculated and used as the first-order IMF1(t).

[0030] Subtracting the first-order intrinsic mode function component IMF1(t) from the original state signal x(t) yields the first-order residual signal r1(t) = x(t) - IMF1(t), which represents the low-frequency components after removing the highest-frequency oscillation. The above process is repeated for the first-order residual signal r1(t), superimposing adaptive white noise of different amplitudes to generate a new noisy set. Empirical mode decomposition and average value calculation are performed on each noisy signal in the new noisy set to extract the next-order intrinsic mode function component IMF2(t). Subtracting IMF2(t) from r1(t) yields r2(t), and so on, layer by layer, until a stopping condition is reached. The stopping condition is when the residual signal becomes a monotonic function, i.e., the number of extreme points of the residual signal is less than two, at which point there are no extractable oscillatory components, or the number of decomposition layers reaches a preset upper limit, typically set to 8-12 layers, to limit computational complexity and avoid over-decomposition. The loop terminates once any condition is met, and a set of multiple IMF components arranged from high to low frequency is used to reflect the dynamic changes of the original state signal at different time scales.

[0031] For example, for a current signal x(t) collected from a prefabricated substation, the set size is set to 100 iterations. White noise with a standard deviation of 0.2 times that of the original state signal is added each time, generating 100 sets of noisy signals. After EMD decomposition of each set of noisy signals, 100 first-order intrinsic mode functions (IMF) components are obtained. The average of these components is calculated to obtain IMF1(t), which mainly corresponds to high-frequency oscillation components above approximately 200Hz. Then, IMF1(t) is subtracted from x(t) to obtain the residual signal r1(t). White noise is added again and the decomposition is repeated to obtain IMF2(t), which corresponds to mid-frequency fluctuation components of approximately 50-200Hz. The process is then iterated to obtain IMF3(t), which is the low-frequency load change component. The decomposition stops when the residual signal of the fourth layer shows an approximately monotonically decreasing trend and has no obvious local extrema. Finally, three components, IMF1, IMF2, and IMF3, are obtained for subsequent state feature extraction and anomaly identification analysis.

[0032] By decomposing the complex original state signal with strong nonlinearity and non-stationary characteristics during the operation of a prefabricated substation into multiple eigenmode function components with clear frequency levels, the effective separation of different physical mechanism components in the signal is achieved, thereby improving the effectiveness and accuracy of monitoring the operating status of the prefabricated substation.

[0033] Furthermore, empirical mode decomposition is performed on the multiple noisy signals to extract their respective first-order intrinsic mode function components. This includes: extracting all local maxima and local minima of each noisy signal; constructing upper and lower envelopes using curve fitting; calculating the mean curves of the upper and lower envelopes; subtracting the corresponding mean curves from the multiple noisy signals to obtain intermediate signals; determining whether the intermediate signals satisfy the waveform conditions of the intrinsic mode functions; if they do, they are output as the first-order intrinsic mode function components corresponding to each noisy signal; if they do not, the filtering process is repeated on the intermediate signals until the waveform conditions of the intrinsic mode functions are satisfied.

[0034] Specifically, standard empirical mode decomposition (EMD) is performed independently for each noisy signal. First, any noisy signal xi(t) traverses all sampling points along the time axis. By comparing the numerical values ​​of the current sampling point with those of its neighbors, all local maxima and local minima in the noisy signal are identified. Local maxima are sampling points whose values ​​are greater than those of their two adjacent sampling points, and local minima are sampling points whose values ​​are less than those of their two adjacent sampling points. Simultaneously, a mirror expansion method is used to treat the two endpoints of the sampling points as extreme points to control envelope divergence.

[0035] Using all local maxima as control nodes, a cubic spline interpolation continuous curve fitting method is employed to construct the upper envelope. Specifically, time is used as the independent variable, and the values ​​of each local maxima are used as the dependent variable. A cubic polynomial function is constructed between every two adjacent maxima to ensure that the interpolation curve not only passes through all maxima but also exhibits continuity in its first and second derivatives at the maxima, forming the upper envelope eu(t). Similarly, using all local minima as control nodes, a lower envelope el(t) is constructed. The upper envelope represents the continuous boundary of local signal peaks, and the lower envelope represents the continuous boundary of local signal valleys; both reflect the local trend changes of the current signal.

[0036] The mean curve m(t) of the upper envelope eu(t) and lower envelope el(t) is calculated using m(t) = (eu(t) + el(t)) / 2, where the mean curve m(t) represents the local low-frequency trend component in the current signal. Then, the mean curve is subtracted from the current noisy signal to obtain the intermediate signal h(t): h(t) = xi(t) - m(t). This intermediate signal is the candidate signal of the oscillation component obtained after stripping the local trend term from the original signal. The waveform conditions of the intrinsic mode function (IMF) are then determined for the intermediate signal h(t). These IMF waveform conditions are waveform constraints used in empirical mode decomposition to determine whether a signal can be used as an IMF output. Specifically, these constraints include: first, the difference between the number of zero-crossing points and the number of local extrema in the intermediate signal does not exceed 1; second, over the entire time range, the mean of the upper envelope formed by local maxima and the lower envelope formed by local minima should be close to zero, meaning the signal oscillates around a zero mean. When the intermediate signal satisfies both of the above conditions, it indicates that it has a single-scale oscillation characteristic and can be used as the first-order intrinsic mode function component output of the noisy signal.

[0037] If the intermediate signal does not satisfy the intrinsic mode function (IMF) condition, the current intermediate signal h(t) is taken as the new signal to be processed, and the extreme point extraction, envelope fitting, mean curve calculation, and trend stripping processes are repeated, i.e., the sieving operation continues. Through iterative sieving operations, the local trend components in the intermediate signal are gradually eliminated, and its oscillation characteristics gradually stabilize. When the preset stopping criterion is reached, such as reaching the maximum number of iterations of 50, or when the IMF condition is satisfied, the final result is output as the first-order IMF component of the current noisy signal. By repeatedly performing the above empirical mode decomposition process on multiple noisy signals, their respective first-order IMF components are obtained.

[0038] By performing empirical mode decomposition on the noisy signal, the most representative high-frequency oscillation components are accurately extracted from the noisy signal, so that the obtained intrinsic mode functions meet the requirements of local zero mean and single-scale oscillation. This ensures that the subsequently obtained IMF components have clear frequency levels and physical meanings, providing stable and reliable input data for the calculation of multi-scale permutation entropy and energy characteristics. At the same time, it reduces the impact of mode mixing and trend interference on the state identification results, thereby improving the accuracy and reliability of the operation status monitoring of the prefabricated substation.

[0039] The multi-scale permutation entropy of the multiple intrinsic mode function components is calculated respectively, and a two-dimensional feature matrix of entropy value-energy is constructed to identify the operating status of the box-type substation and output the status monitoring results.

[0040] Furthermore, the multi-scale permutation entropy of the multiple intrinsic mode function components is calculated, including: for any component among the multiple intrinsic mode function components, coarsening is performed according to multiple different scale factors to generate a multi-scale coarse-grained sequence; the phase space of the multi-scale coarse-grained sequence is reconstructed at each scale to reconstruct the one-dimensional sequence into a multi-dimensional sequence; the data in each dimension after reconstruction are sorted by size, and the probability of each sorting pattern is calculated; the permutation entropy value at each scale is calculated based on the probability of each sorting pattern; the permutation entropy values ​​at each scale are averaged to obtain the multi-scale permutation entropy value of any component.

[0041] Specifically, for any IMF component among multiple intrinsic mode function components, multiple scale factors τ are set according to the signal sampling rate and fault characteristic frequency band, τ=1, 2, 3...N. For example, for a vibration signal sampled at 2560Hz, the maximum scale factor τ is set to 10~15. The scale factor is used to characterize the observation granularity of the time series; the larger the scale, the stronger the time smoothing processing of the signal. At each scale τ, the original IMF component is coarsened: the original IMF component is divided into non-overlapping segments according to the length τ, and the arithmetic mean of the sampled values ​​in each segment is taken to form a new coarse-grained sequence y(j). By using a sliding interval averaging method, the time structure features of different scales in the original IMF are extracted respectively. Small scale factors, such as τ=1, retain the micro-fluctuation details of the original IMF, while large scale factors, such as τ=10, reflect the macro-change trend of the IMF within a longer time window, thereby reducing the impact of high-frequency fluctuations on the complexity calculation. After coarsening of all scale factors, a multi-scale coarse-grained sequence of the IMF component is obtained.

[0042] Phase space reconstruction is performed on each scale sequence in the multi-scale coarse-grained sequence. Phase space reconstruction refers to constructing a multi-dimensional sequence from a one-dimensional time series using an embedding dimension m and a time delay λ. The embedding dimension m is the dimension of the phase space reconstruction, typically taken as 3~7, and the time delay λ is preferably 1. For example, for a coarse-grained sequence y(j) at any scale τ, a phase space vector is constructed: Y(j) = [y(j), y(j+λ), ..., y(j+(m-1)λ)], where j is the sliding time index. This allows multiple phase space vectors Y(j) to be formed by sliding a window with a step size of 1 point-by-point on the original sequence, resulting in a vector set {Y(1), Y(2), ..., Y(N)}, used to map the dynamic evolution relationship of the time series to a multi-dimensional space. After obtaining the phase space vector set, the data in each dimension after reconstruction are arranged in ascending order to form a corresponding sorting pattern. For example, when m=3, Y(1)=[2.5,4.5,7], since 2.5<4.5<7, the corresponding permutation pattern is (1,2,3); Y(2)=[4.5,7,8], also with the permutation pattern (1,2,3); Y(3)=[7,8,6], with the corresponding permutation pattern (3,1,2). The relative frequency of all permutation patterns at this scale τ is statistically analyzed, i.e., the proportion pi of each permutation pattern's occurrence to the total number of windows is calculated to obtain the probability pi of each permutation pattern's occurrence. The permutation entropy value is then calculated based on pi: H(τ)=-∑pilog(pi), used to characterize the uncertainty or complexity of the signal at this scale. Then, the permutation entropy of the IMF component at all scales τ is averaged to obtain the multi-scale permutation entropy value of the IMF component, used to comprehensively characterize the overall complexity level of the component at different time scales.

[0043] By combining multi-scale coarsening with phase space reconstruction, the dynamic uncertainty changes of the operating signals of prefabricated substations at different time scales can be obtained, thereby enhancing the sensitivity to early weak faults, random disturbances and operating status evolution trends, and improving the sensitivity, accuracy and reliability of prefabricated substation operating status monitoring.

[0044] Furthermore, constructing an entropy-energy two-dimensional feature matrix includes: calculating the sum of squares of the component sequences for each of the multiple intrinsic mode function components as energy values; pairing the energy values ​​with the multi-scale permutation entropy values ​​corresponding to the multiple intrinsic mode function components to establish the entropy-energy two-dimensional feature matrix.

[0045] Furthermore, the operating status of the prefabricated substation is identified, and the status monitoring results are output, including: when the prefabricated substation is in a healthy state, collecting the normal state reference feature matrix; calculating the deviation index between the entropy-energy two-dimensional feature matrix and the normal state reference feature matrix; when the deviation index exceeds a preset alarm threshold, determining that the prefabricated substation is in an abnormal operating state and issuing an early warning signal.

[0046] Specifically, the sum of squares of the component sequences for each of the multiple intrinsic mode function (IMF) components is calculated as the energy value. That is, the energy value is the cumulative sum of the squared amplitudes of the component over the entire time interval, used to characterize the signal strength carried by the IMF component. A larger energy value indicates more intense dynamic activity or more significant anomalous disturbances in that frequency band. The energy value Ek is paired with the multi-scale entropy values ​​Hk corresponding to each of the multiple IMF components to construct a two-dimensional feature vector Fk=(Hk,Ek), where Fk represents the coordinates of the k-th IMF component in the feature space, Hk is the multi-scale entropy value of the k-th IMF component, and Ek is the energy value of the k-th IMF component. Arranging the feature vectors of all IMF components in component order forms the entropy-energy two-dimensional feature matrix.

[0047] When the prefabricated substation is in a healthy operating state, such as when the equipment is fault-free, the load is stable, and it has passed the historical stable operation stage after acceptance and commissioning, the original state signals under multiple time windows are collected, and the signal decomposition, IMF extraction, multi-scale entropy calculation and energy calculation are completed according to the aforementioned method, so as to construct the benchmark feature matrix F0 under the healthy state. The benchmark feature matrix is ​​a reference set of entropy value-energy distribution of multiple IMF components under the healthy operating condition, which is used to reflect the normal dynamic behavior mode of the prefabricated substation.

[0048] The deviation index between the entropy-energy two-dimensional feature matrix and the normal state baseline feature matrix is ​​calculated using the Euclidean distance form of the eigenvectors corresponding to each IMF component. Where Hk and Ek represent the multi-scale permutation entropy and energy value of the k-th IMF component in the current state, respectively; Hk0 and Ek0 represent the characteristic values ​​of the corresponding IMF component in the healthy baseline state; and K is the number of IMF components. The deviation index reflects the overall degree of deviation of the current operating state from the healthy state.

[0049] Alarm thresholds are preset based on historical health data, such as by using the mean plus three standard deviations or empirical quantiles. The deviation index is compared with the preset alarm threshold. When the deviation index exceeds the preset alarm threshold, the current operating status of the prefabricated substation is determined to be abnormal, and a warning signal is output. When the deviation index is less than the preset alarm threshold, the equipment is determined to be in normal healthy operating condition, and monitoring continues.

[0050] By constructing a health benchmark feature space and quantifying deviations, a standardized and quantifiable comparative analysis of the operating status of prefabricated substations can be achieved. This avoids the instability caused by relying solely on a single feature or manual threshold judgment, improves the sensitivity and accuracy of abnormal operating status monitoring, and enables early warning of early faults.

[0051] In summary, the method for monitoring the operational status of prefabricated substations in photovoltaic power plants provided in this application has the following technical advantages: By acquiring raw state signals during the operation of prefabricated substations in real time, including electrical and non-electrical parameter signals, multi-source sensing of equipment power transmission status, mechanical vibration status, and thermal operation status is achieved. The raw state signals are decomposed to obtain multiple intrinsic mode function components, thus adaptively decomposing complex nonlinear and non-stationary mixed signals into inherent oscillation components at different frequency scales. This achieves effective decoupling and noise suppression of signals from different physical mechanisms. The multi-scale permutation entropy of each intrinsic mode function component is calculated, and an entropy-energy two-dimensional feature matrix is ​​constructed by combining the energy characteristics of each component. This jointly reflects the equipment operating status in both complexity and energy intensity dimensions, enhancing the discriminative power of feature expression. Finally, based on the two-dimensional feature matrix, the operating status of the prefabricated substation is identified, and the status monitoring results are output. This enables refined quantitative assessment of equipment operating status and early identification of abnormal states. Compared to traditional single electrical quantity or single-scale feature analysis methods, this improves the sensitivity and reliability of monitoring early weak faults, non-stationary disturbances, and abnormal states.

[0052] Example 2, based on the same inventive concept as the method for monitoring the operating status of a prefabricated substation for photovoltaic power plants in the foregoing examples, such as... Figure 2 As shown, this application provides a prefabricated substation operation status monitoring system for photovoltaic power plants, wherein the prefabricated substation operation status monitoring system for photovoltaic power plants includes: The signal acquisition module 11 is used to acquire the original state signals during the operation of the prefabricated substation in real time. The original state signals include electrical parameter signals and non-electrical parameter signals of the prefabricated substation. The signal decomposition module 12 is used to decompose the original state signals to obtain multiple intrinsic mode function components. The state identification module 13 is used to calculate the multi-scale permutation entropy of the multiple intrinsic mode function components, construct an entropy-energy two-dimensional feature matrix, identify the operating state of the prefabricated substation, and output the state monitoring results.

[0053] Furthermore, the signal decomposition module 12 is also used to: add adaptive white noise of different amplitudes to the original state signal multiple times to form multiple noisy signals; perform empirical mode decomposition on the multiple noisy signals respectively, extract the first-order intrinsic mode function components of each, calculate the average value as the first-order intrinsic mode function component; subtract the first-order intrinsic mode function component from the original state signal to obtain the first-order residual signal; add adaptive white noise to the first-order residual signal again, repeat the empirical mode decomposition and averaging process, extract the next-order intrinsic mode function component, until the stopping condition is reached, and obtain the multiple intrinsic mode function components arranged in descending order of frequency.

[0054] Furthermore, the signal decomposition module 12 is also used to: the stopping condition is that the residual signal becomes a monotonic function or reaches a preset number of decomposition layers.

[0055] Furthermore, the signal decomposition module 12 is also used to: extract all local maxima and local minima of each noisy signal, construct upper and lower envelopes using curve fitting; calculate the mean curves of the upper and lower envelopes, and subtract the corresponding mean curves from the multiple noisy signals to obtain intermediate signals; determine whether the intermediate signals satisfy the waveform conditions of the intrinsic mode functions; if they satisfy, they are output as the first-order intrinsic mode function components corresponding to each noisy signal; if they do not satisfy, the filtering process is repeated on the intermediate signals until the waveform conditions of the intrinsic mode functions are satisfied.

[0056] Furthermore, the state recognition module 13 is also used to: coarse-grainedly process any component among the plurality of intrinsic mode function components according to multiple different scale factors to generate a multi-scale coarse-grained sequence; reconstruct the phase space of the multi-scale coarse-grained sequence at each scale to reconstruct the one-dimensional sequence into a multi-dimensional sequence; sort the data in each dimension after reconstruction according to size and count the probability of each sorting pattern; calculate the permutation entropy value at each scale according to the probability of each sorting pattern; and average the permutation entropy values ​​at each scale to obtain the multi-scale permutation entropy value of any component.

[0057] Furthermore, the state recognition module 13 is also used to: calculate the sum of squares of the component sequences for each of the multiple intrinsic mode function components as energy values; pair the energy values ​​with the multi-scale permutation entropy values ​​corresponding to the multiple intrinsic mode function components to establish the entropy-energy two-dimensional feature matrix.

[0058] Furthermore, the status identification module 13 is also used to: collect the normal state reference feature matrix when the prefabricated substation is in a healthy state; calculate the deviation index between the entropy-energy two-dimensional feature matrix and the normal state reference feature matrix; and determine that the prefabricated substation is in an abnormal operating state and issue a warning signal when the deviation index exceeds a preset alarm threshold.

[0059] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The method and specific examples for monitoring the operation status of a prefabricated substation for photovoltaic fields in the foregoing embodiment 1 are also applicable to the prefabricated substation operation status monitoring system for photovoltaic fields in this embodiment. Through the foregoing detailed description of the method for monitoring the operation status of a prefabricated substation for photovoltaic fields, those skilled in the art can clearly understand the prefabricated substation operation status monitoring system for photovoltaic fields in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0060] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0061] Obviously, those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for monitoring the operational status of prefabricated substations in photovoltaic power plants, characterized in that, include: The raw status signals of the prefabricated substation during operation are collected in real time. The raw status signals include electrical parameter signals and non-electrical parameter signals of the prefabricated substation. The original state signal is decomposed to obtain multiple intrinsic mode function components; The multi-scale permutation entropy of the multiple intrinsic mode function components is calculated respectively, and a two-dimensional feature matrix of entropy value-energy is constructed to identify the operating status of the box-type substation and output the status monitoring results.

2. The method for monitoring the operating status of a prefabricated substation for photovoltaic power plants as described in claim 1, characterized in that, The original state signal is decomposed to obtain multiple intrinsic mode function components, including: Multiple adaptive white noises of different amplitudes are added to the original state signal to form multiple noisy signals; Empirical mode decomposition is performed on the multiple noisy signals respectively to extract their respective first-order intrinsic mode function components, and the average value is calculated as the first-order intrinsic mode function component. Subtracting the first-order intrinsic mode function component from the original state signal yields the first-order residual signal; Adaptive white noise is added again to the first-order residual signal, and empirical mode decomposition and averaging are repeated to extract the next-order intrinsic mode function components until the stopping condition is met, resulting in the plurality of intrinsic mode function components arranged in descending order of frequency.

3. The method for monitoring the operating status of a prefabricated substation for photovoltaic power plants as described in claim 2, characterized in that, The stopping condition is that the residual signal becomes a monotonic function or reaches a preset number of decomposition layers.

4. The method for monitoring the operating status of a prefabricated substation for photovoltaic power plants as described in claim 2, characterized in that, Empirical mode decomposition (EMD) is performed on each of the multiple noisy signals to extract their respective first-order intrinsic mode function (IMF) components, including: Extract all local maxima and local minima of each noisy signal, and construct upper and lower envelopes using curve fitting. Calculate the mean curves of the upper envelope and the lower envelope, and subtract the corresponding mean curves from the multiple noisy signals to obtain the intermediate signal; Determine whether the intermediate signal satisfies the waveform conditions of the intrinsic mode function. If it does, then output it as the first-order intrinsic mode function component corresponding to each noise signal. If the conditions are not met, the filtering process is repeated for the intermediate signal until the waveform conditions of the intrinsic mode function are met.

5. The method for monitoring the operating status of a prefabricated substation for photovoltaic power plants as described in claim 4, characterized in that, Calculate the multi-scale permutation entropy of the multiple intrinsic mode function components, including: For any component among the multiple intrinsic mode function components, coarsening is performed according to multiple different scale factors to generate a multi-scale coarsened sequence. The phase space of the multi-scale coarse-grained sequence is reconstructed at each scale, and the one-dimensional sequence is reconstructed into a multi-dimensional sequence. The data in each dimension after reconstruction are sorted by size, and the probability of each sorting pattern is counted. Calculate the permutation entropy value at each scale based on the probability of each sorting pattern occurring. The permutation entropy value at each scale is averaged to obtain the multi-scale permutation entropy value of any component.

6. The method for monitoring the operating status of a prefabricated substation for photovoltaic power plants as described in claim 5, characterized in that, Constructing an entropy-energy two-dimensional feature matrix includes: For each of the multiple intrinsic mode function components, the sum of squares of the component sequences is calculated as the energy value; The energy value is paired with the multi-scale entropy values ​​corresponding to the multiple intrinsic mode function components to establish the entropy-energy two-dimensional feature matrix.

7. The method for monitoring the operating status of a prefabricated substation for photovoltaic power plants as described in claim 6, characterized in that, The operating status of the prefabricated substation is identified, and the status monitoring results are output, including: When the prefabricated substation is in a healthy state, the normal state baseline feature matrix is ​​collected. Calculate the deviation index between the entropy-energy two-dimensional feature matrix and the normal state baseline feature matrix; When the deviation index exceeds the preset alarm threshold, the prefabricated substation is determined to be in an abnormal operating state and an early warning signal is issued.

8. A prefabricated substation operation status monitoring system for photovoltaic power plants, characterized in that, The steps for implementing the method for monitoring the operation status of a prefabricated substation for photovoltaic power plants according to any one of claims 1 to 7, wherein the prefabricated substation operation status monitoring system for photovoltaic power plants comprises: The signal acquisition module is used to acquire raw status signals during the operation of the prefabricated substation in real time. The raw status signals include electrical parameter signals and non-electrical parameter signals of the prefabricated substation. The signal decomposition module is used to decompose the original state signal to obtain multiple intrinsic mode function components; The status identification module is used to calculate the multi-scale permutation entropy of the multiple intrinsic mode function components, construct an entropy-energy two-dimensional feature matrix, identify the operating status of the box-type substation, and output the status monitoring results.

Images