Early warning method, system and product for county building photovoltaic equipment

Abstract

The invention discloses an early warning method, system and product for county building photovoltaic equipment, and belongs to the technical field of equipment early warning, the data of the county building photovoltaic equipment is acquired according to monitoring points, an operation observation vector is obtained, the regional operation and maintenance efficiency is improved, the problem that the false alarm rate is relatively high due to a single data source is avoided, and the early warning efficiency is improved. Meanwhile, the diagnosis precision of the generation power reduction reason is improved, a comparison matrix is constructed, a pre-constructed equipment fault early warning model is trained based on the comparison matrix, equipment real-time monitoring data is compared with equipment historical monitoring data in the same environment, the influence of dynamic environmental factors is fully considered, and the diagnosis accuracy of the generation power reduction reason is improved. And early warning is not limited to a fixed power threshold value, so that the accuracy and the real-time performance of early warning can be remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of equipment early warning technology, specifically relating to an early warning method, system, and product for building photovoltaic equipment in county areas. Background Technology

[0002] With the rapid development of photovoltaic power generation technology, the role of photovoltaic systems in the energy field is becoming increasingly prominent. County-level building photovoltaic equipment refers to the installation of solar photovoltaic modules on the roof, walls or other structures of buildings within a county to realize distributed photovoltaic power generation and provide electricity. It has the characteristics of large number and geographical dispersion. However, the existing county-level building photovoltaic equipment also has corresponding shortcomings, specifically: (1) High false alarm rate leading to low diagnostic accuracy: The existing early warning methods are based only on the electrical data reported by the inverter, including voltage, current and power. When the power generation decreases, it is impossible to effectively distinguish whether the decrease is due to equipment failure, external environmental influence or grid fluctuation, resulting in a high false alarm rate and low diagnostic accuracy; (2) Low operation and maintenance decision efficiency: The existing early warning methods are mostly "single-point" monitoring, which analyzes all equipment in the area in isolation and cannot identify regional and cluster risks, resulting in low operation and maintenance decision efficiency; (3) Failure to consider the impact of dynamic environmental factors: Most of them use fixed power thresholds for alarms and do not fully consider the impact of dynamic environmental factors such as weather changes, resulting in insufficient timeliness and accuracy of early warning.

[0003] Therefore, how to provide an effective technical solution to address the problems of high false alarm rates leading to low diagnostic accuracy, low operational and maintenance decision-making efficiency, and failure to consider the impact of dynamic environmental factors in existing technologies has become an urgent problem to be solved in existing technologies. Summary of the Invention

[0004] The purpose of this invention is to provide an early warning method, system, and product for building photovoltaic equipment in county areas, in order to solve the above-mentioned problems existing in the prior art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides an early warning method for building-integrated photovoltaic (BIPV) equipment in county-level areas, comprising: Data is collected from pre-selected monitoring points for building-integrated photovoltaic (BIPV) equipment in the county to obtain operational observation vectors and historical monitoring data of the county's BIPV equipment. A comparison matrix is ​​constructed based on the operational observation vectors and the historical monitoring data of the county's BIPV equipment in order to build a state model of the county's BIPV equipment based on the comparison matrix. The state model of building photovoltaic equipment in the county is mapped to a probability interval to obtain a mapping model. Basic fault probability allocation factors are selected in the probability interval, and information is fused between the selected basic fault probability allocation factors and the mapping model to obtain fused data. The support of the fused data is calculated to output the posterior probability value of the county-level building photovoltaic equipment status model. The photovoltaic equipment fault early warning index is obtained based on the posterior probability value of the county-level building photovoltaic equipment status model. The pre-constructed equipment fault early warning model is trained based on photovoltaic equipment fault early warning indicators and historical monitoring data of county-level building photovoltaic equipment to obtain the trained equipment fault early warning model. The trained equipment fault early warning model is used to take real-time monitoring data of county-level building photovoltaic equipment as input and output multiple fault indicators to generate fault early warning information based on multiple fault indicators.

[0006] In one possible design, data is collected from county-level building-integrated photovoltaic (BIPV) systems based on pre-selected monitoring points to obtain operational observation vectors. Historical monitoring data for the county-level BIPV systems is then acquired. A comparison matrix is ​​constructed based on the operational observation vectors and the historical monitoring data to build a state model of the county-level BIPV systems, including: Based on pre-selected monitoring points, data is collected from building-integrated photovoltaic (BIPV) equipment in the county at preset times to obtain the observation vector values ​​of the operation of the BIPV equipment. The observation vector values ​​of the operation of the BIPV equipment are combined to obtain the operation observation vector. Acquire historical monitoring data of building-integrated photovoltaic (BIPV) equipment in the county. The historical monitoring data of BIPV equipment in the county includes historical photovoltaic equipment parameters, historical environmental data, equipment geospatial data, and historical environmental image data. Based on the operational observation vector, historical monitoring data under different operating conditions are selected from the historical monitoring data of building photovoltaic equipment in the county, and a comparison matrix is ​​constructed based on the historical monitoring data under different operating conditions. The historical monitoring data of county-level building photovoltaic equipment is upsampled using a preset sliding window and comparison matrix to obtain a smooth data sequence with independent characteristics. The smooth data sequence is then stream-processed using a multivariate nonlinear function to obtain the stream processing result. The preset average activation degree of photovoltaic equipment data is obtained, and the singularities of the running observation vector are examined based on the preset average activation degree of photovoltaic equipment data and the stream processing result to obtain a singularity sequence. Using the autocorrelation function, the partial vector of the state model of county-level building photovoltaic equipment is obtained from the singular point sequence, and the failure loss time of photovoltaic equipment is obtained. Based on the partial vector of the state model of county-level building photovoltaic equipment and the failure loss time of photovoltaic equipment, the state model of county-level building photovoltaic equipment is constructed.

[0007] In one possible design, a basic fault probability allocation factor is selected within the probability interval, and information fusion is performed between the selected basic fault probability allocation factor and the mapping model to obtain fused data, including: The distance between the selected basic fault probability allocation factors is calculated based on the mapping model to obtain the allocation factor distance; Information fusion is performed on historical monitoring data of building photovoltaic equipment in the county based on the allocation factor distance to obtain fused data.

[0008] In one possible design, the support of the fused data is calculated to output the posterior probability value of the county-level building-based photovoltaic (BPV) equipment state model. Based on the posterior probability value of the county-level BPV equipment state model, photovoltaic equipment fault early warning indicators are obtained, including: The fused data is smoothed based on a preset smoothing coefficient to obtain the processed fused data. A pre-built positive and negative sample classifier is used to classify the historical monitoring data of county-level building photovoltaic equipment to obtain positive and negative samples. Support is calculated on the processed fused data, positive samples, and negative samples to obtain the mutual support between the fused data and the historical monitoring data of county-level building photovoltaic equipment. A classification matrix for negative samples is constructed based on the mutual support between the processed fused data and the historical monitoring data of building photovoltaic equipment in the county. The processed fused data is averaged and weighted based on the classification matrix of negative samples to obtain weighted fused data. The posterior probability value of the state model of building photovoltaic equipment in the county is obtained based on the fused data and the weighted fused data. The cosine distance of the preset fault feature vector is obtained, and the fault warning index of the photovoltaic equipment is obtained based on the cosine distance of the preset fault feature vector and the posterior probability value of the county-level building photovoltaic equipment status model.

[0009] In one possible design, a pre-built equipment fault early warning model is trained based on photovoltaic equipment fault early warning indicators and historical monitoring data of building-integrated photovoltaic equipment in the county, resulting in a trained equipment fault early warning model, including: A self-organizing mapping algorithm is used to construct an equipment fault early warning model based on photovoltaic equipment fault early warning indicators. Based on the equipment fault early warning model, assumptions are made about the distribution of fault data in the historical monitoring data of building photovoltaic equipment in the county. Subsequently, the equipment fault model was trained based on the distribution of fault data and historical monitoring data of building photovoltaic equipment in the county, resulting in a trained equipment fault early warning model.

[0010] In one possible design, the equipment fault model is trained based on the distribution of fault data and historical monitoring data of building-integrated photovoltaic (BIPV) equipment in the county, resulting in a trained equipment fault early warning model, including: Historical monitoring data of building photovoltaic equipment in the county was used as the training set. The training set was labeled based on the distribution of fault data to obtain a training set with label values. Input the training set with labeled values ​​into the device fault model and output the residual prediction value; The loss value is calculated based on the loss function, the predicted residual value, and the label value. The equipment failure model is then updated based on the loss value to obtain the updated equipment failure model. Repeat the above update steps until the preset iteration termination condition is met to obtain the trained equipment fault early warning model.

[0011] In one possible design, the fault indicator is the residual value of the fault data; the step of generating fault early warning information based on multiple fault indicators includes: The average residual values ​​of multiple fault data are averaged to obtain the average residual value; The residual warning interval is obtained based on the preset residual ratio and the average residual value; If the residual value of the fault data is within the residual warning range, it is determined to be a fault warning state and a fault warning message is generated; otherwise, it is determined to be a non-fault warning state.

[0012] Secondly, the present invention provides an early warning system for building-integrated photovoltaic (BIPV) equipment in county-level areas, comprising: The model is used to collect data on county-level building photovoltaic (PV) equipment based on pre-selected monitoring points, obtain operational observation vectors, acquire historical monitoring data of county-level PV equipment, and construct a comparison matrix based on the operational observation vectors and historical monitoring data of county-level PV equipment, so as to construct a state model of county-level PV equipment based on the comparison matrix. The information fusion module is used to map the state model of building photovoltaic equipment in the county to a probability interval to obtain a mapping model. Basic fault probability allocation factors are selected in the probability interval, and information fusion is performed on the selected basic fault probability allocation factors and the mapping model to obtain fused data. The indicator output module is used to calculate the support of the fused data in order to output the posterior probability value of the county-level building photovoltaic equipment status model. Based on the posterior probability value of the county-level building photovoltaic equipment status model, the photovoltaic equipment fault early warning indicator is obtained. The model training module is used to train a pre-built equipment fault early warning model based on photovoltaic equipment fault early warning indicators and historical monitoring data of county-level building photovoltaic equipment to obtain a trained equipment fault early warning model. The trained equipment fault early warning model is used to take real-time monitoring data of county-level building photovoltaic equipment as input and output multiple fault indicators to generate fault early warning information based on multiple fault indicators.

[0013] Thirdly, the present invention provides a computer device comprising a memory, a processor, and a transceiver connected in sequence and communication, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the early warning method for building photovoltaic equipment in county areas as described in the first aspect above.

[0014] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, perform the early warning method for building-integrated photovoltaic (BIPV) equipment in a county as described in the first aspect above.

[0015] Fifthly, the present invention provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the early warning method for building photovoltaic equipment in a county as described in the first aspect above.

[0016] The beneficial effects of this invention are as follows: This invention discloses an early warning method, system, device, and product for building-integrated photovoltaic (BIPV) equipment in county-level areas. First, data is collected from BIPV equipment at pre-selected monitoring points to obtain operational observation vectors and historical monitoring data. A comparison matrix is ​​constructed based on the operational observation vectors and historical monitoring data. A state model of the BIPV equipment is then built based on the comparison matrix. This state model is mapped to a probability interval to obtain a mapping model. Basic fault probability allocation factors are selected within the probability interval, and information is fused between the selected basic fault probability allocation factors and the mapping model to obtain fused data. The support of the fused data is calculated to output the posterior probability value of the BIPV equipment state model. A fault warning index for the photovoltaic equipment is obtained based on the posterior probability value. The pre-constructed equipment fault warning model is trained based on the fault warning index and historical monitoring data of the BIPV equipment to obtain a trained equipment fault warning model. The trained equipment fault warning model takes real-time monitoring data of the BIPV equipment as input and outputs multiple fault indicators to generate fault warning information. This invention collects data from county-level building-integrated photovoltaic (BIPV) equipment at monitoring points. The collected data covers county-level BIPV equipment within a specific region, improving regional operation and maintenance efficiency. It also addresses the issue of high false alarm rates caused by a single data source, enhancing the diagnostic accuracy of power generation decline. By constructing a comparison matrix to establish a dynamic baseline, real-time equipment data is compared with equipment data under the same environment. This fully considers the impact of dynamic environmental factors and is not limited to fixed power thresholds for early warning, significantly improving the accuracy and real-time performance of early warnings, making it easy to apply and promote. Attached Figure Description

[0017] Figure 1 A flowchart of an early warning method for building photovoltaic equipment in a county-level area provided by an embodiment of the present invention; Figure 2 This is a module block diagram of an early warning system for building photovoltaic equipment in a county, provided in an embodiment of the present invention. Figure 3 A structural diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0019] It should be understood that although the terms first, second, etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit, without departing from the scope of the exemplary embodiments of the invention.

[0020] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist simultaneously. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.

[0021] Example: like Figure 1 As shown, the first aspect of this embodiment provides an early warning method for building-integrated photovoltaic (BIPV) equipment in county-level areas. This method can be executed, but is not limited to, by a computer device or virtual machine with certain computing resources, such as a personal computer or smartphone, or by a virtual machine. The early warning method includes, but is not limited to, the following steps: S1. Data is collected from the county-level building photovoltaic equipment based on pre-selected monitoring points to obtain the operation observation vector, acquire the historical monitoring data of the county-level building photovoltaic equipment, and construct a comparison matrix based on the operation observation vector and the historical monitoring data of the county-level building photovoltaic equipment so as to construct a state model of the county-level building photovoltaic equipment based on the comparison matrix; It should be noted that the county-level building photovoltaic equipment in this embodiment refers to small and medium-sized distributed solar power generation systems that are widely and dispersedly installed on various buildings within a county-level administrative region, consisting of multiple photovoltaic devices installed on buildings. For pre-selected monitoring points, for example, common fault occurrence points can be selected, such as the DC side of the inverter, the AC side of the inverter, and meteorological stations. The data collected in this embodiment includes, but is not limited to, photovoltaic equipment parameters, real-time equipment operation data, environmental meteorological data, and equipment spatial static data. Among them, photovoltaic equipment parameters include, but are not limited to, the output voltage and current of each photovoltaic string; real-time equipment operation data includes, but is not limited to, the real-time power generation, conversion efficiency, and internal temperature of the inverter; environmental meteorological data includes, but is not limited to, real-time and forecast meteorological data of the location of each device, specifically including solar irradiance, ambient temperature, wind speed, and precipitation; and equipment spatial static data includes, but is not limited to, the latitude and longitude, altitude, installation tilt angle, and equipment static profile of each device. The equipment static profile includes the device model and historical maintenance records. Furthermore, it also includes environmental image data. According to a predetermined interval, aerial photography is conducted using drones or satellite remote sensing to obtain high-resolution images of each photovoltaic device.

[0022] Further, in step S1, data is collected from the county-level building-integrated photovoltaic (BIPV) equipment based on pre-selected monitoring points to obtain operational observation vectors and historical monitoring data of the county-level BIPV equipment. A comparison matrix is ​​constructed based on the operational observation vectors and the historical monitoring data of the county-level BIPV equipment to build a state model of the county-level BIPV equipment, including: S11. Based on pre-selected monitoring points, data is collected from the county-level building photovoltaic equipment at preset times to obtain the observation vector value of the county-level building photovoltaic equipment operation. The observation vector value of the county-level building photovoltaic equipment operation is combined to obtain the operation observation vector. S12. Obtain historical monitoring data of building-in photovoltaic equipment in the county, including historical photovoltaic equipment parameters, historical environmental data, equipment geospatial data, and historical environmental image data. S13. Based on the operational observation vector, select historical monitoring data under different operating conditions from the historical monitoring data of building photovoltaic equipment in the county, and construct a comparison matrix based on the historical monitoring data under different operating conditions; It should be noted that the historical monitoring data for different operating conditions refers to the specific operating status and mode of building-integrated photovoltaic (BIPV) equipment in the county under various combinations of internal and external factors. Specifically, it may include external environmental conditions, internal load conditions, equipment health conditions, and grid interaction conditions, such as peak conditions on sunny days, low load conditions on cloudy days, and transitional conditions between morning and evening. By filtering the historical monitoring data of building-integrated photovoltaic equipment in the county under different operating conditions, historical monitoring data under different operating conditions can be obtained.

[0023] In practice, monitoring points are selected based on historical experience and used as relevant operating parameters for county-level building photovoltaic (BPV) equipment. At time t, data is collected from the BPV equipment at the monitoring points to obtain the operating observation vector. Based on the operating observation vector, representative historical monitoring data under different operating conditions are selected from the historical monitoring data of the BPV equipment. A comparison matrix is ​​constructed based on the historical monitoring data under the same operating conditions. The resulting comparison matrix is ​​equivalent to a historical case database of the current status of the county-level BPV equipment. This avoids using fixed standard thresholds to compare the status of the county-level BPV equipment under all conditions in the future, and is used for subsequent comparison with real-time monitoring data.

[0024] S14. Use a preset sliding window and comparison matrix to upsample the historical monitoring data of county-level building photovoltaic equipment to obtain a smooth data sequence with independent characteristics. Use a multivariate nonlinear function to stream process the smooth data sequence to obtain the streaming processing result. Obtain the preset average activation degree of photovoltaic equipment data. Use the preset average activation degree of photovoltaic equipment data and streaming processing result to check the singular points of the running observation vector and obtain the singular point sequence. It should be noted that a sliding window is used to perform sliding analysis and upsampling on the comparison matrix and historical monitoring data. Upsampling refers to the process of increasing the sampling rate or frequency of the signal or data, inserting additional sample points to improve the granularity or resolution of the data, thereby increasing the data density. The specific upsampling process includes zero insertion and interpolation filtering operations. Specifically, zero values ​​are inserted between adjacent data points in the historical monitoring data to increase the number of historical monitoring data points. Then, an interpolation filter is used to smooth the inserted zero values ​​to restore the continuity of the historical monitoring data, ultimately obtaining a smooth data sequence with filtered random noise. Subsequently, multivariate nonlinear... The function performs stream processing on a smooth data sequence, and the resulting stream processing result is the expected model result, specifically indicating how the photovoltaic equipment should operate under the current conditions. This process is continuous. The stream processing result is compared with the preset average activation degree of the photovoltaic equipment data and the operating observation vector. Here, the preset average activation degree of the photovoltaic equipment data is a manually set benchmark threshold or range for the photovoltaic equipment during normal operation. If the deviation between the operating observation vector and the stream processing result exceeds the average activation degree of the photovoltaic equipment data, it is marked as a singular point. All the obtained singular points are arranged in chronological order to obtain a singular point sequence, thereby improving the accuracy and timeliness of fault warning.

[0025] S15. Using the autocorrelation function, obtain the partial vector of the county-level building photovoltaic equipment state model based on the singular point sequence, obtain the photovoltaic equipment failure loss time, and construct the county-level building photovoltaic equipment state model based on the partial vector of the county-level building photovoltaic equipment state model and the photovoltaic equipment failure loss time.

[0026] It should be noted that the autocorrelation coefficient is used to analyze the regularity of the occurrence of singular points in the singular point sequence, thereby obtaining the partial vector. The partial vector reflects the systematic trend and degree of the photovoltaic equipment's state deviating from the ideal state. At the same time, the partial vector is combined with the photovoltaic equipment's failure loss time to construct a county-level building photovoltaic equipment state model. The county-level building photovoltaic equipment state model is used to determine whether the county-level building photovoltaic equipment is healthy, and can also quantify the degree of its unhealthiness and predict future failure risks and potential losses.

[0027] S2. Map the state model of building photovoltaic equipment in the county to a probability interval to obtain the mapping model. Select the basic fault probability allocation factor in the probability interval and fuse the selected basic fault probability allocation factor and the mapping model to obtain the fused data. In a preferred embodiment, a basic fault probability allocation factor is selected within the probability interval, and information fusion is performed on the selected basic fault probability allocation factor and the mapping model to obtain fused data, including: S21. Calculate the distance between the selected basic fault probability allocation factors based on the mapping model to obtain the allocation factor distance; S22. Information fusion is performed on historical monitoring data of building photovoltaic equipment in the county based on the allocation factor distance to obtain fused data.

[0028] It should be noted that the number of basic fault probability allocation factors selected is two, and the calculation expression for the distance between the allocation factors is as follows: In the formula, For the distance of the allocation factor, For mapping models, As the basic failure probability allocation factor, Assigning another basic failure probability factor, The difference between the basic failure probability allocation factors. The rate of change between the basic fault probability allocation factors; the specific process of information fusion based on the allocation factor distance for historical monitoring data of county-level building photovoltaic equipment includes: extracting the overall operating parameters and state change characteristics of the historical monitoring data of county-level building photovoltaic equipment, and fusing the historical monitoring data of county-level building photovoltaic equipment based on the allocation factor distance, the overall operating parameters and state change characteristics of the historical monitoring data of county-level building photovoltaic equipment, to obtain the fused data.

[0029] S3. Calculate the support of the fused data in order to output the posterior probability value of the county-level building photovoltaic equipment status model, and obtain the photovoltaic equipment fault early warning index based on the posterior probability value of the county-level building photovoltaic equipment status model. Furthermore, the support of the fused data is calculated to output the posterior probability value of the county-level building-based photovoltaic (BPV) equipment state model. Based on the posterior probability value of the county-level BPV equipment state model, photovoltaic equipment fault early warning indicators are obtained, including: S31. Smooth the fused data based on the preset smoothing coefficient to obtain the processed fused data. Use a pre-built positive and negative sample classifier to classify the historical monitoring data of county-level building photovoltaic equipment to obtain positive and negative samples. Calculate the support of the processed fused data, positive samples, and negative samples to obtain the mutual support between the fused data and the historical monitoring data of county-level building photovoltaic equipment. S32. Construct a classification matrix for negative samples based on the mutual support between the processed fused data and the historical monitoring data of building photovoltaic equipment in the county; S33. The processed fused data is averaged and weighted based on the classification matrix of the negative samples to obtain weighted fused data. The posterior probability value of the state model of building photovoltaic equipment in the county is obtained based on the fused data and the weighted fused data. S34. Obtain the cosine distance of the preset fault feature vector, and obtain the photovoltaic equipment fault early warning index based on the cosine distance of the preset fault feature vector and the posterior probability value of the county building photovoltaic equipment state model.

[0030] It should be noted that the positive and negative sample classifier is trained using a large training set labeled with both positive and negative samples to obtain a classifier capable of classifying positive and negative samples. In this embodiment, the positive and negative samples represent data from the historical monitoring data of county-level building-based photovoltaic equipment in a normal state and data from the historical monitoring data of county-level building-based photovoltaic equipment in an abnormal state, respectively. The mutual support between the processed fused data, positive samples, and negative samples is calculated to calibrate the fused data and the historical monitoring data of county-level building-based photovoltaic equipment. The expression for calculating the mutual support is: In the formula, For mutual support, For the merged data, The smoothing coefficient is used for smoothing. The number of positive samples. The number of negative samples is represented here. The classification matrix of the negative samples constructed here is equivalent to a fault feature template library. By using the fault feature template library, the processed fused data is averaged and weighted to obtain data that better reflects the probability of the fault, i.e., weighted fused data. The posterior probability value is obtained based on the weighted fused data. The expression for calculating the posterior probability value is as follows: In the formula, This is the posterior probability value. To weighted merge data, The coefficients of variation for the fused data are the sample dispersion coefficients. The mixing coefficient is the fusion coefficient of the fused data. In this embodiment, the cosine distance of the preset fault feature vector is the cosine distance between the fault feature vector and the standard feature vector, which is used to measure the magnitude of the fault. Furthermore, the expression for the photovoltaic equipment fault early warning index is: In the formula, As a fault early warning indicator for photovoltaic equipment, To predetermine the cosine distance of the fault feature vector, It is an integrable function. This is done through inner product operations to obtain photovoltaic equipment fault warning indicators, which can then be used for subsequent fault warnings.

[0031] S4. The pre-constructed equipment fault early warning model is trained based on photovoltaic equipment fault early warning indicators and historical monitoring data of county-level building photovoltaic equipment to obtain the trained equipment fault early warning model. The trained equipment fault early warning model is used to take real-time monitoring data of county-level building photovoltaic equipment as input and output multiple fault indicators to generate fault early warning information based on multiple fault indicators.

[0032] In a preferred embodiment, a pre-built equipment fault early warning model is trained based on photovoltaic equipment fault early warning indicators and historical monitoring data of building-integrated photovoltaic equipment in a county, resulting in a trained equipment fault early warning model, including: S41. Using the self-organizing map algorithm, construct an equipment fault early warning model based on photovoltaic equipment fault early warning indicators; S42. Assumptions are made regarding the distribution of fault data in historical monitoring data of building photovoltaic equipment in the county based on the equipment fault early warning model; S43. Subsequently, the equipment fault model is trained based on the distribution of fault data and historical monitoring data of building photovoltaic equipment in the county, resulting in a trained equipment fault early warning model.

[0033] The Self-organizing Maps (SOM) algorithm is an unsupervised learning neural network method. Its core principle is that neurons in the network compete with each other to match input data and adjust weights to reflect the data distribution.

[0034] Furthermore, in step S43, the equipment fault model is trained based on the distribution of fault data and historical monitoring data of building-integrated photovoltaic equipment in the county, resulting in a trained equipment fault early warning model, including: S43.1. Use the historical monitoring data of building photovoltaic equipment in the county as the training set, and label the training set based on the distribution of fault data to obtain a training set with label values; S43.2. Input the training set with label values ​​into the device fault model and output the residual prediction values; S43.3. Calculate the loss value based on the loss function, residual prediction value and label value, update the equipment failure model according to the loss value, and obtain the updated equipment failure model; S43.4. Repeat the above update steps until the preset iteration termination condition is reached to obtain the trained equipment fault early warning model.

[0035] It should be noted that the preset iteration termination conditions include, but are not limited to, reaching the preset number of iterations or the obtained loss value being less than the preset error threshold.

[0036] In a preferred embodiment, the fault indicator is the residual value of the fault data; the step of generating fault warning information based on multiple fault indicators includes: S44. Calculate the average value of the residuals of multiple fault data to obtain the average residual value; S45. Obtain the residual warning interval based on the preset residual ratio and the average residual value; S46. If the residual value of the fault data is within the residual warning range, it is determined to be a fault warning state and a fault warning message is generated; otherwise, it is determined to be a non-fault warning state.

[0037] In specific implementation, the preset residual ratios in this embodiment are 80% and 120%. Therefore, the residual warning interval is [80% of the average residual value and 120% of the average residual value]. If the residual value of the fault data is within this residual warning interval, it is determined to be a fault warning state; otherwise, it is determined to be a non-fault warning state.

[0038] Based on the above-disclosed content, this embodiment provides an early warning method for building-integrated photovoltaic (BIPV) equipment in county-level areas. It collects multimodal data of BIPV equipment in county-level areas and constructs a state model of the BIPV equipment based on this data. This model is then compared with real-time monitoring data to reduce warning errors. Simultaneously, it calculates the expected state of BIPV equipment in county-level areas under different environments and compares this expected state with the real-time monitoring state to improve the accuracy and timeliness of fault warnings. This avoids using a single threshold for fault warnings and is suitable for the photovoltaic equipment early warning industry, facilitating application and promotion.

[0039] like Figure 2 As shown, the second aspect of this embodiment provides an early warning system for building-integrated photovoltaic (BIPV) equipment in county-level areas, comprising: The model is used to collect data on county-level building photovoltaic (PV) equipment based on pre-selected monitoring points, obtain operational observation vectors, acquire historical monitoring data of county-level PV equipment, and construct a comparison matrix based on the operational observation vectors and historical monitoring data of county-level PV equipment, so as to construct a state model of county-level PV equipment based on the comparison matrix. The information fusion module is used to map the state model of building photovoltaic equipment in the county to a probability interval to obtain a mapping model. Basic fault probability allocation factors are selected in the probability interval, and information fusion is performed on the selected basic fault probability allocation factors and the mapping model to obtain fused data. The indicator output module is used to calculate the support of the fused data in order to output the posterior probability value of the county-level building photovoltaic equipment status model. Based on the posterior probability value of the county-level building photovoltaic equipment status model, the photovoltaic equipment fault early warning indicator is obtained. The model training module is used to train a pre-built equipment fault early warning model based on photovoltaic equipment fault early warning indicators and historical monitoring data of county-level building photovoltaic equipment to obtain a trained equipment fault early warning model. The trained equipment fault early warning model is used to take real-time monitoring data of county-level building photovoltaic equipment as input and output multiple fault indicators to generate fault early warning information based on multiple fault indicators.

[0040] The working process, working details and technical effects of the early warning system for building photovoltaic equipment in county areas provided in the second aspect of this embodiment can be found in the early warning method for building photovoltaic equipment in county areas described in the first aspect, and will not be repeated here.

[0041] like Figure 3 As shown, the third aspect of this embodiment provides a computer device, including a memory, a processor, and a transceiver connected in sequence for communication. The memory stores a computer program, the transceiver sends and receives messages, and the processor reads the computer program to execute the early warning method for county-level building photovoltaic equipment as described in the first aspect. Specifically, the memory may include, but is not limited to, random-access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or first-in-last-out (FILO) memory, etc.; the processor may include, but is not limited to, an STM32F105 series microprocessor. Furthermore, the computer device may also include, but is not limited to, a power module, a display screen, and other necessary components.

[0042] The working process, working details and technical effects of the aforementioned computer equipment provided in the third aspect of this embodiment can be found in the early warning method for building photovoltaic equipment in county areas described in the first aspect, and will not be repeated here.

[0043] The fourth aspect of this embodiment provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, and when the instructions are executed on a computer, perform the early warning method for building-integrated photovoltaic equipment in county areas as described in the first aspect. The computer-readable storage medium refers to a carrier for storing data, and may include, but is not limited to, computer-readable storage media such as floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or Memory Sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0044] The working process, working details and technical effects of the aforementioned computer-readable storage medium provided in the fourth aspect of this embodiment can be found in the early warning method for building photovoltaic equipment in county areas as described in the first aspect, and will not be repeated here.

[0045] The fifth aspect of this embodiment provides a computer program product, including a computer program or instructions, which, when executed by a computer, are used to implement the early warning method for building photovoltaic equipment in county areas as described in the first aspect.

[0046] The working process, working details and technical effects of the aforementioned computer program product provided in this embodiment can be found in the early warning method for building photovoltaic equipment in county areas as described in the first aspect, and will not be repeated here.

[0047] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An early warning method for building-integrated photovoltaic (BIPV) equipment in county-level areas, characterized in that, include: Data is collected from pre-selected monitoring points for building-integrated photovoltaic (BIPV) equipment in the county to obtain operational observation vectors and historical monitoring data of the county-level BIPV equipment. A comparison matrix is ​​constructed based on the operational observation vectors and the historical monitoring data of the county-level BIPV equipment in order to build a state model of the county-level BIPV equipment based on the comparison matrix. The state model of building photovoltaic equipment in the county is mapped to a probability interval to obtain the mapping model. Basic fault probability allocation factors are selected in the probability interval, and the selected basic fault probability allocation factors and the mapping model are fused to obtain the fused data. The support of the fused data is calculated to output the posterior probability value of the county-level building photovoltaic equipment status model. The photovoltaic equipment fault early warning index is obtained based on the posterior probability value of the county-level building photovoltaic equipment status model. The pre-constructed equipment fault early warning model is trained based on photovoltaic equipment fault early warning indicators and historical monitoring data of county-level building photovoltaic equipment to obtain the trained equipment fault early warning model. The trained equipment fault early warning model is used to take real-time monitoring data of county-level building photovoltaic equipment as input and output multiple fault indicators to generate fault early warning information based on multiple fault indicators.

2. The early warning method for building-integrated photovoltaic (BIPV) equipment in a county-level area according to claim 1, characterized in that, Data is collected from pre-selected monitoring points for county-level building-integrated photovoltaic (BIPV) equipment to obtain operational observation vectors. Historical monitoring data for BIPV equipment in the county is acquired. A comparison matrix is ​​constructed based on the operational observation vectors and the historical monitoring data of BIPV equipment in the county. This comparison matrix is ​​then used to build a state model of the county-level BIPV equipment, including: Based on pre-selected monitoring points, data is collected from building-integrated photovoltaic (BIPV) equipment in the county at preset times to obtain the observation vector values ​​of the operation of the BIPV equipment. The observation vector values ​​of the operation of the BIPV equipment are combined to obtain the operation observation vector. Obtain historical monitoring data of building-integrated photovoltaic (BIPV) equipment in the county. The historical monitoring data of BIPV equipment in the county includes historical photovoltaic equipment parameters, historical environmental data, equipment geospatial data, and historical environmental image data. Based on the operational observation vector, historical monitoring data under different operating conditions are selected from the historical monitoring data of building photovoltaic equipment in the county, and a comparison matrix is ​​constructed based on the historical monitoring data under different operating conditions. The historical monitoring data of county-level building photovoltaic equipment is upsampled using a preset sliding window and comparison matrix to obtain a smooth data sequence with independent characteristics. The smooth data sequence is then stream-processed using a multivariate nonlinear function to obtain the stream processing result. The preset average activation degree of photovoltaic equipment data is obtained, and the singularities of the running observation vector are examined based on the preset average activation degree of photovoltaic equipment data and the stream processing result to obtain a singularity sequence. Using the autocorrelation function, the partial vector of the state model of county-level building photovoltaic equipment is obtained from the singular point sequence, and the failure loss time of photovoltaic equipment is obtained. Based on the partial vector of the state model of county-level building photovoltaic equipment and the failure loss time of photovoltaic equipment, the state model of county-level building photovoltaic equipment is constructed.

3. The early warning method for building-integrated photovoltaic (BIPV) equipment in a county-level area according to claim 1, characterized in that, A basic fault probability allocation factor is selected within the probability interval, and information fusion is performed between the selected basic fault probability allocation factor and the mapping model to obtain fused data, including: The distance between the selected basic fault probability allocation factors is calculated based on the mapping model to obtain the allocation factor distance; Information fusion is performed on historical monitoring data of building photovoltaic equipment in the county based on the allocation factor distance to obtain fused data.

4. The early warning method for building-integrated photovoltaic (BIPV) equipment in a county-level area according to claim 1, characterized in that, The support of the fused data is calculated to output the posterior probability value of the county-level building photovoltaic (PV) equipment state model. Based on the posterior probability value of the county-level PV equipment state model, PV equipment fault early warning indicators are obtained, including: The fused data is smoothed based on a preset smoothing coefficient to obtain the processed fused data. A pre-built positive and negative sample classifier is used to classify the historical monitoring data of county-level building photovoltaic equipment to obtain positive and negative samples. Support is calculated on the processed fused data, positive samples, and negative samples to obtain the mutual support between the fused data and the historical monitoring data of county-level building photovoltaic equipment. A classification matrix for negative samples is constructed based on the mutual support between the processed fused data and the historical monitoring data of building photovoltaic equipment in the county. The processed fused data is averaged and weighted based on the classification matrix of negative samples to obtain weighted fused data. The posterior probability value of the state model of building photovoltaic equipment in the county is obtained based on the fused data and the weighted fused data. The cosine distance of the preset fault feature vector is obtained, and the fault warning index of the photovoltaic equipment is obtained based on the cosine distance of the preset fault feature vector and the posterior probability value of the county-level building photovoltaic equipment status model.

5. The early warning method for building-integrated photovoltaic (BIPV) equipment in a county-level area according to claim 1, characterized in that, The pre-built equipment fault early warning model is trained based on photovoltaic equipment fault early warning indicators and historical monitoring data of building-integrated photovoltaic equipment in the county, resulting in the trained equipment fault early warning model, including: A self-organizing mapping algorithm is used to construct an equipment fault early warning model based on photovoltaic equipment fault early warning indicators. Based on the equipment fault early warning model, assumptions are made about the distribution of fault data in the historical monitoring data of building photovoltaic equipment in the county. Subsequently, the equipment fault model was trained based on the distribution of fault data and historical monitoring data of building photovoltaic equipment in the county, resulting in a trained equipment fault early warning model.

6. The early warning method for building-integrated photovoltaic (BIPV) equipment in a county-level area according to claim 5, characterized in that, The equipment fault model was trained based on the distribution of fault data and historical monitoring data of building-integrated photovoltaic (BIPV) equipment in the county, resulting in a trained equipment fault early warning model, including: Historical monitoring data of building photovoltaic equipment in the county was used as the training set. The training set was labeled based on the distribution of fault data to obtain a training set with label values. Input the training set with labeled values ​​into the device fault model and output the residual prediction value; The loss value is calculated based on the loss function, the predicted residual value, and the label value. The equipment failure model is then updated based on the loss value to obtain the updated equipment failure model. Repeat the above update steps until the preset iteration termination condition is met to obtain the trained equipment fault early warning model.

7. The early warning method for building-integrated photovoltaic (BIPV) equipment in a county-level area according to claim 1, characterized in that, The fault index is the residual value of the fault data; The generation of fault warning information based on multiple fault indicators includes: The average residual values ​​of multiple fault data are averaged to obtain the average residual value; The residual warning interval is obtained based on the preset residual ratio and the average residual value; If the residual value of the fault data is within the residual warning range, it is determined to be a fault warning state and a fault warning message is generated; otherwise, it is determined to be a non-fault warning state.

8. An early warning system for building-integrated photovoltaic (BIPV) equipment in a county, for implementing the method according to any one of claims 1 to 7, characterized in that, include: The model is used to collect data on county-level building photovoltaic (PV) equipment based on pre-selected monitoring points, obtain operational observation vectors, acquire historical monitoring data of county-level PV equipment, and construct a comparison matrix based on the operational observation vectors and historical monitoring data of county-level PV equipment, so as to construct a state model of county-level PV equipment based on the comparison matrix. The information fusion module is used to map the state model of building photovoltaic equipment in the county to a probability interval to obtain a mapping model. Basic fault probability allocation factors are selected in the probability interval, and information fusion is performed on the selected basic fault probability allocation factors and the mapping model to obtain fused data. The indicator output module is used to calculate the support of the fused data in order to output the posterior probability value of the county-level building photovoltaic equipment status model. Based on the posterior probability value of the county-level building photovoltaic equipment status model, the photovoltaic equipment fault early warning indicator is obtained. The model training module is used to train a pre-built equipment fault early warning model based on photovoltaic equipment fault early warning indicators and historical monitoring data of county-level building photovoltaic equipment to obtain a trained equipment fault early warning model. The trained equipment fault early warning model is used to take real-time monitoring data of county-level building photovoltaic equipment as input and output multiple fault indicators to generate fault early warning information based on multiple fault indicators.

9. A computer device, characterized in that, The device includes a memory, a processor, and a transceiver that are sequentially and communicatively connected. The memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the early warning method for building photovoltaic equipment in a county as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or the instructions are executed by the computer, they implement the early warning method for building photovoltaic equipment in county areas as described in any one of claims 1 to 7.

Images