A method for detecting faults of base station-wrapped solar panels

Abstract

The present invention discloses a method for detecting faults of base station-wrapped solar panels, which relates to the technical field of fault detection. The present invention obtains power data and environmental parameters of solar panels, adopts high-frequency sampling and sliding window filtering processing, identifies short-term power fluctuation characteristics through time series modeling, combines autoregressive and moving average characteristics, calculates the power change rate, and detects abnormal power fluctuations; in the power anomaly screening stage, based on time series analysis and comparison with historical power data, a light power mapping model is introduced, and a mapping relationship between light and power is established through regression analysis to eliminate misjudgments caused by short-term light changes; in the fault identification stage, infrared thermal imaging is combined to analyze the temperature distribution of the solar panel, calculate the temperature mean and variance, determine whether there is local overheating, and cross-compare the light data with the thermal imaging data to determine whether the power drop is caused by short-term shading or panel damage.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection, and in particular to a method for detecting faults of base station wrapped solar panels. Background Art

[0002] Remote communication base stations and off-grid power supply systems have extremely high requirements for power supply stability. Solar panels have become one of the main energy solutions due to their independence and low maintenance characteristics. To save space, improve structural stability and reduce the impact of the external environment on the panels, solar panels are generally installed in a wrapped manner and distributed along the base station tower or building facade.

[0003] Traditional solar panel fault detection relies on overall power monitoring, which involves detecting changes in the system's output power to determine if there are any anomalies with the panels or electrical connections. When system power falls below a preset threshold, it often indicates that certain panels may be faulty or obstructed. However, in a base station-wrapped layout, this power fluctuation-based monitoring method is susceptible to interference from the external environment. Base station towers are typically equipped with a large number of ancillary equipment, such as antennas, cables, and cooling fans, which cast irregular shadows at different times, affecting the power generation efficiency of some panels. Furthermore, panels exposed to the outdoor environment for long periods of time may be affected by factors such as bird nesting and wind-blown foreign objects, causing dynamic changes in light distribution across different panels. These factors can cause brief power fluctuations but do not cause actual damage to the panels. Traditional power monitoring methods can only detect overall power drops and cannot effectively distinguish between brief shadows and actual faults, resulting in false alarms or missed alarms, and unnecessary maintenance costs. Therefore, a fault detection method for base station-wrapped solar panels is urgently needed to address this issue. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] The present invention provides a base station wrapped solar panel fault detection method to solve the problem that traditional power monitoring methods are difficult to distinguish between local shadows and panel failures, resulting in increased false alarms and missed alarms, affecting the base station power supply stability and maintenance costs.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] An embodiment of the present invention provides a base station wrapped solar panel fault detection method, which includes:

[0008] Step S1, collecting power data of the solar panel, including output power, current, voltage, and environmental data, to form a monitoring data set;

[0009] Step S2, performing high-frequency sampling on the power data collected in step S1, constructing a time series model and identifying short-term power fluctuation characteristics;

[0010] Step S3: Compare the power change trend after processing in step S2 with historical power data and the illumination model to determine whether the power drop meets the short-term occlusion characteristics and screen out abnormal situations;

[0011] Step S4: performing fault identification based on the abnormal conditions screened out in step S3, analyzing the heat distribution of the solar panel in combination with infrared thermal imaging, and distinguishing whether the power drop is caused by a panel failure or light obstruction;

[0012] Step S5: Based on the fault identification information of step S4, a corresponding maintenance instruction is generated, or a maintenance person is notified to perform a maintenance operation.

[0013] As a preferred solution of the base station wrapped solar panel fault detection method described in the present invention, the environmental data includes light intensity, temperature and humidity.

[0014] As a preferred solution of the base station wrapped solar panel fault detection method described in the present invention, in step S2, the power change trend is analyzed to identify short-term power fluctuation characteristics.

[0015] As a preferred solution of the base station wrapped solar panel fault detection method of the present invention, the step of high-frequency sampling of the power data collected in step S1 is as follows:

[0016] Perform high-frequency sampling and data smoothing:

[0017] Assume the power data set is P:

[0018] P={P1,P2,…,P N},

[0019] Where P represents the original power dataset, P i represents the power data at the i-th time point, N represents the total number of data points collected,

[0020] The uniform high-frequency sampling strategy is used for sampling. The sampling time interval is Δt, and the power sequence after sampling is expressed as P t :

[0021]

[0022] Among them, P t represents the power sequence after high-frequency sampling, Indicates the sampling time point t j The power value at , M represents the number of data points after sampling,

[0023] Here, sliding window filtering is used to reduce the impact of high-frequency noise. The window size is set to W, and the smoothed power value is P′ t , the calculation formula is:

[0024]

[0025] Where W represents the sliding window size, P′ t represents the smoothed power series.

[0026] As a preferred solution of the base station wrapped solar panel fault detection method of the present invention, the step of constructing the time series model is:

[0027] Based on the smoothed power data P′ t , construct the autoregressive moving average model ARMA, the model expression is:

[0028]

[0029] Among them, P′ t represents the prediction power at time t, P represents the order of the autoregressive term, α i represents the autoregressive coefficient, where i = 1, 2, ..., P, Q represents the order of the moving average term, β j represents the moving average coefficient, where j = 1, 2, ..., Q, ∈ t represents the noise term at time t, ∈ t-j represents the noise term at time tj.

[0030] As a preferred solution of the base station wrapped solar panel fault detection method of the present invention, the step of analyzing the power change trend and identifying the short-term power fluctuation characteristics is as follows:

[0031] Calculate the short-time power change rate using the following formula:

[0032] ΔP t =P′ t -P′ t-1 ,

[0033] Set the fluctuation threshold Θ P , identify abnormal fluctuations:

[0034] If |ΔP t |>Θ P , it is determined to be a short-term power fluctuation.

[0035] Where ΔP t represents the rate of change of power at time t,

[0036] Θ PIndicates the short-term power fluctuation judgment threshold.

[0037] As a preferred solution of the base station wrapped solar panel fault detection method described in the present invention, the step of comparing the power change trend after processing in step S2 with the historical power data and the illumination model to determine whether the power drop meets the short-term shading characteristics includes:

[0038] The time series trend P′ t Compared with historical power data P H For comparison:

[0039] D t =P′ t -P H ,

[0040] Among them, D t Indicates the power drop at time t, P H Indicates the historical average power value of the corresponding time period.

[0041] If satisfied:

[0042] |D t |>Θ D , then it is determined that the power drop is abnormal, where

[0043] Θ D Indicates the power drop judgment threshold.

[0044] As a preferred solution of the base station wrapped solar panel fault detection method of the present invention, the step of determining whether the power drop meets the short-term shading characteristics also includes:

[0045] Analyze with the illumination model:

[0046] Assume that the light intensity data set is: I={I1,I2,…,I L},

[0047] Among them, I represents the light intensity dataset, I l represents the light intensity at the lth time point, L represents the total number of sampling points of light intensity data,

[0048] The light power mapping model F(I) is used to calculate the light impact, and the formula is:

[0049] P H =F(I),

[0050] Where F(I) is represented by a polynomial regression model, which is expressed as:

[0051] F(I)=c0+c1I+c2I 2 +…+cK I K ,

[0052] Among them, c k is the regression coefficient, K is the polynomial order,

[0053] Calculate the power deviation after illumination adjustment using the following formula:

[0054] D′ t =P′ t -F(I t ),

[0055] If |D′ t |>Θ D , it is marked as an abnormal situation.

[0056] As a preferred solution of the base station wrapped solar panel fault detection method of the present invention, the step of analyzing the heat distribution of the solar panel by combining infrared thermal imaging is as follows:

[0057] Assume that the infrared imaging data matrix is:

[0058] T={T ij},i=1,2,…,M,j=1,2,…,N,

[0059] Calculate the mean temperature

[0060]

[0061] Calculating temperature variance

[0062]

[0063] Among them, T ij Represents the temperature at position (i, j) on the solar panel, M represents the number of rows of the panel, and N represents the number of columns of the panel. represents the mean panel temperature, represents the variance of the panel temperature distribution,

[0064] like It is determined that there is local overheating of the panel and there may be a fault.

[0065] Θ T Indicates the temperature variance threshold for fault judgment.

[0066] As a preferred solution of the base station wrapped solar panel fault detection method of the present invention, in step S4, the cause of the fault is comprehensively determined:

[0067] Comparing illumination data with infrared thermal imaging data:

[0068] If |D′ t |>Θ D and

[0069] Then the auxiliary judgment power drop is caused by light blocking,

[0070] If |D′ t |>Θ D and

[0071] The auxiliary determines that the power drop is caused by a panel failure.

[0072] The beneficial effects of the present invention are as follows: the present invention obtains the power data and environmental parameters of the solar panel, adopts high-frequency sampling and sliding window filtering processing to reduce high-frequency noise and improve data stability; then, through time series modeling, combined with autoregressive and moving average characteristics, it identifies the characteristics of short-term power fluctuations, calculates the power change rate, and detects abnormal power fluctuations; in the power anomaly screening stage, based on time series analysis and comparison with historical power data, it introduces a light power mapping model, and uses regression analysis to establish a mapping relationship between light and power, eliminating misjudgments caused by short-term light changes and improving the accuracy of identifying abnormal power drops.

[0073] In the fault identification stage, the present invention combines infrared thermal imaging to analyze the temperature distribution of the solar panel, calculates the temperature mean and variance, determines whether there is local overheating, and cross-compares the illumination data with the thermal imaging data to determine whether the power drop is caused by short-term shading or panel damage, thereby reducing false alarms or missed alarms caused by environmental factors.

[0074] The present invention generates maintenance instructions or notifies maintenance personnel to perform maintenance through a maintenance scheduling strategy after the fault is confirmed, and optimizes power output in combination with a dynamic power compensation strategy to improve the power supply stability of the system.

[0075] In summary, the present invention improves detection accuracy, reduces maintenance workload, and enhances the adaptability of the detection method, making the base station solar power supply system more efficient and stable. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0077] Figure 1Schematic diagram of the flow of the base station wrapped solar panel fault detection method of the present invention. DETAILED DESCRIPTION

[0078] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0079] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0080] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0081] Example 1, with reference to Figure 1 This embodiment provides a base station wrapped solar panel fault detection method, comprising the following steps:

[0082] Step S1, collecting power data of the solar panel, including output power, current, voltage, and environmental data, to form a monitoring data set;

[0083] Environmental data includes light intensity, temperature, and humidity;

[0084] Step S2, performing high-frequency sampling on the power data collected in step S1, constructing a time series model and identifying short-term power fluctuation characteristics;

[0085] In step S2, the power variation trend is analyzed to identify short-term power fluctuation characteristics;

[0086] The steps of high-frequency sampling the power data collected in step S1 are as follows:

[0087] Perform high-frequency sampling and data smoothing:

[0088] Assume the power data set is P:

[0089] P={P1,P2,…,P N},

[0090] Where P represents the original power dataset, P i represents the power data at the i-th time point, N represents the total number of data points collected,

[0091] The uniform high-frequency sampling strategy is used for sampling. The sampling time interval is Δt, and the power sequence after sampling is expressed as P t :

[0092]

[0093] Among them, P t represents the power sequence after high-frequency sampling, Indicates the sampling time point t j The power value at , M represents the number of data points after sampling,

[0094] Here, sliding window filtering is used to reduce the impact of high-frequency noise. The window size is set to W, and the smoothed power value is P′ t , the calculation formula is:

[0095]

[0096] Where W represents the sliding window size, P′ t represents the smoothed power series;

[0097] The steps to build a time series model are:

[0098] Based on the smoothed power data P′ t , construct the autoregressive moving average model ARMA, the model expression is:

[0099]

[0100] Among them, P′ t represents the prediction power at time t, P represents the order of the autoregressive term, α i represents the autoregressive coefficient, where i = 1, 2, ..., P, Q represents the order of the moving average term, β j represents the moving average coefficient, where j = 1, 2, ..., Q, ∈ t represents the noise term at time t, ∈ t-j represents the noise term at time tj;

[0101] The steps to analyze the power change trend and identify the characteristics of short-term power fluctuations are as follows:

[0102] Calculate the short-time power change rate using the following formula:

[0103] ΔP t =P′ t -P′ t-1 ,

[0104] Set the fluctuation threshold Θ P , identify abnormal fluctuations:

[0105] If |ΔPt |>Θ P , it is determined to be a short-term power fluctuation.

[0106] Where ΔP t represents the rate of change of power at time t,

[0107] Θ P Indicates the short-term power fluctuation judgment threshold;

[0108] Specifically, in step S2, high-frequency sampling is performed to capture the changes in solar panel power, and sliding window filtering is used to reduce high-frequency noise, improve data stability, and enhance the accuracy of time series modeling;

[0109] The ARMA model is used to model power data and identify power trends over time. The ARMA model combines the characteristics of autoregression and moving average to effectively capture short-term dynamic changes and predict future power values. Finally, the short-term power change rate is calculated, and a fluctuation threshold is set to identify abnormal fluctuation characteristics.

[0110] In step S2, the sharp change of short-term power is accurately detected;

[0111] Step S3: Compare the power change trend after processing in step S2 with historical power data and the illumination model to determine whether the power drop meets the short-term occlusion characteristics and screen out abnormal situations;

[0112] Based on the comparison of the power change trend after processing in step S2 with the historical power data and the illumination model, the step of determining whether the power drop meets the short-term occlusion characteristics includes:

[0113] The time series trend P′ t Compared with historical power data P H For comparison:

[0114] D t =P′ t -P H ,

[0115] Among them, D t Indicates the power drop at time t, P H Indicates the historical average power value of the corresponding time period.

[0116] If satisfied:

[0117] |D t |>Θ D , then it is determined that the power drop is abnormal, where

[0118] Θ D Indicates the power drop judgment threshold,

[0119] The step of determining whether the power drop meets the short-term shading characteristics also includes:

[0120] Analyze with the illumination model:

[0121] Assume that the light intensity data set is: I={I1,I2,…,I L},

[0122] Among them, I represents the light intensity dataset, I l represents the light intensity at the lth time point, L represents the total number of sampling points of light intensity data,

[0123] The light power mapping model F(I) is used to calculate the light impact, and the formula is:

[0124] P H =F(I),

[0125] Where F(I) is represented by a polynomial regression model, which is expressed as:

[0126] F(I)=c0+c1I+c2I 2 +…+c K I K ,

[0127] Among them, c k is the regression coefficient, K is the polynomial order,

[0128] Calculate the power deviation after illumination adjustment using the following formula:

[0129] D′ t =P′ t -F(I t ),

[0130] If |D′ t |>Θ D , it is marked as an abnormal situation,

[0131] Specifically, in step S3, the deviation between the current power data and the historical power data is calculated to identify abnormal power drops. Combined with the illumination model, a regression analysis is performed on the illumination intensity to establish a mapping relationship between illumination and power, thereby eliminating power drops caused by changes in ambient illumination and making the screening results more accurate.

[0132] Here, power deviation is further calculated using data corrected by the illumination power model, and anomalies are determined based on thresholds. This method effectively avoids false alarms caused by short-term environmental factors and improves the accuracy of power anomaly detection. Overall, this method achieves intelligent screening of power drops and can distinguish between environmental factors and systemic anomalies.

[0133] Step S4: performing fault identification based on the abnormal conditions screened out in step S3, analyzing the heat distribution of the solar panel in combination with infrared thermal imaging, and distinguishing whether the power drop is caused by a panel failure or light obstruction;

[0134] The steps to analyze the heat distribution of solar panels using infrared thermal imaging are:

[0135] Assume that the infrared imaging data matrix is:

[0136] T={T ij},i=1,2,…,M,j=1,2,…,N,

[0137] Calculate the mean temperature

[0138]

[0139] Calculating temperature variance

[0140]

[0141] Among them, T ij Represents the temperature at position (i, j) on the solar panel, M represents the number of rows of the panel, and N represents the number of columns of the panel. represents the mean panel temperature, represents the variance of the panel temperature distribution,

[0142] like It is determined that there is local overheating of the panel and there may be a fault.

[0143] Θ T represents the temperature variance threshold for fault determination;

[0144] In step S4, the cause of the fault is comprehensively determined:

[0145] Comparing illumination data with infrared thermal imaging data:

[0146] If |D′ t |>Θ D and

[0147] Then the auxiliary judgment power drop is caused by light blocking,

[0148] If |D′ t |>Θ D and

[0149] Then the auxiliary determines that the power drop is caused by a panel failure;

[0150] In step S4, infrared thermal imaging technology is used to analyze the temperature distribution of the solar panel. The temperature mean and variance are used to calculate the local hotspot area to identify possible fault points. The abnormal power drop points are combined with the temperature distribution to accurately distinguish between shading and panel failure.

[0151] Specifically, if the power drops significantly but the thermal imaging temperature variance is small, it is likely caused by a short-term obstruction. If the power drops significantly and the temperature variance is large, it indicates a possible panel fault, such as local overheating or circuit abnormality. In addition, the analysis method in this step enhances the ability to interpret power anomalies. Maintenance personnel can make accurate diagnoses based on the dual characteristics of power and temperature, thereby improving maintenance efficiency and reducing unnecessary maintenance costs. It provides an efficient means of detecting panel health and can assist in fault tracing and decision-making.

[0152] Step S5: Based on the fault identification information of step S4, a corresponding maintenance instruction is generated, or a maintenance person is notified to perform a maintenance operation.

[0153] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for detecting faults in a base station-wrapped solar panel, characterized by: include, Step S1, collecting power data of the solar panel, including output power, current, voltage, and environmental data, to form a monitoring data set; Step S2, performing high-frequency sampling on the power data collected in step S1, constructing a time series model and identifying short-term power fluctuation characteristics; Step S3: Compare the power change trend after processing in step S2 with historical power data and the illumination model to determine whether the power drop meets the short-term occlusion characteristics and screen out abnormal situations; Step S4: performing fault identification based on the abnormal conditions screened out in step S3, analyzing the heat distribution of the solar panel in combination with infrared thermal imaging, and distinguishing whether the power drop is caused by a panel failure or light obstruction; Step S5: Based on the fault identification information of step S4, a corresponding maintenance instruction is generated, or a maintenance person is notified to perform a maintenance operation; In step S2, the power variation trend is analyzed to identify the short-term power fluctuation characteristics, and the steps are as follows: Calculate the short-time power change rate using the following formula: ΔP t =P′ t -P′ t-1 , Set the fluctuation threshold Θ P , identify abnormal fluctuations: If |ΔP t |>Θ P , it is determined to be a short-term power fluctuation. Where ΔP t represents the rate of change of power at time t, Θ P Indicates the short-term power fluctuation judgment threshold; Step S3 includes: The time series trend P' t Compared with historical power data P H For comparison: D t =P′ t -P H , Among them, D t Indicates the power drop at time t, P H Indicates the historical average power value of the corresponding time period. If satisfied: |D t |>Θ D , then it is determined that the power drop is abnormal, where Θ D Indicates the power drop judgment threshold; Analyze with the illumination model: Assume that the light intensity data set is: I={I1,I2,…,I L }, Among them, I represents the light intensity dataset, I l represents the light intensity at the lth time point, L represents the total number of sampling points of light intensity data, The light power mapping model F(I) is used to calculate the light impact, and the formula is: P H =F(I), Where F(I) is represented by a polynomial regression model, which is expressed as: F(I)=c0+c1I+c2I 2 +…+c k I K , Among them, c k is the regression coefficient, K is the polynomial order, Calculate the power deviation after illumination adjustment using the following formula: D′ t =P′ t -F(I t ), If |D' t |>Θ D , it is marked as an abnormal situation.

2. The base station wrapped solar panel fault detection method according to claim 1, characterized in that: The environmental data includes light intensity, temperature and humidity.

3. The base station wrapped solar panel fault detection method according to claim 2, characterized in that: In step S2, the power variation trend is analyzed to identify short-term power fluctuation characteristics.

4. The base station wrapped solar panel fault detection method according to claim 3, characterized in that: The step of high-frequency sampling the power data collected in step S1 is as follows: Perform high-frequency sampling and data smoothing: Assume the power data set is P: P={P1,P2,…,P N }, Where P represents the original power dataset, P i represents the power data at the i-th time point, N represents the total number of data points collected, The uniform high-frequency sampling strategy is used for sampling. The sampling time interval is Δt, and the power sequence after sampling is expressed as P t : Among them, P t represents the power sequence after high-frequency sampling, Indicates the sampling time point t j The power value at , M represents the number of data points after sampling, Sliding window filtering is used, the window size is set to W, and the smoothed power value is P' t , the calculation formula is: Where W represents the sliding window size, P' t represents the smoothed power series.

5. The base station wrapped solar panel fault detection method according to claim 4, characterized in that: The steps of constructing the time series model are: Based on the smoothed power data P' t , construct the autoregressive moving average model ARMA, the model expression is: Among them, P' t represents the prediction power at time t, P represents the order of the autoregressive term, α i represents the autoregressive coefficient, where i = 1, 2, ..., P, Q represents the order of the moving average term, β j represents the moving average coefficient, where j = 1, 2, ..., Q, ∈ t represents the noise term at time t, ∈ t-j represents the noise term at time tj.

6. The base station wrapped solar panel fault detection method according to claim 5, characterized in that: The steps of analyzing the heat distribution of the solar panel by combining infrared thermal imaging are as follows: Assume that the infrared imaging data matrix is: T={T ij },i=1,2,…,M,j=1,2,…,N, Calculate the mean temperature Calculating temperature variance Among them, T ij Represents the temperature at position (i, j) on the solar panel, M represents the number of rows of the panel, and N represents the number of columns of the panel. represents the mean panel temperature, represents the variance of the panel temperature distribution, like It is determined that there is local overheating of the panel, and there is a fault, among which, Θ T Indicates the temperature variance threshold for fault judgment.

7. The base station wrapped solar panel fault detection method according to claim 6, characterized in that: In step S4, the cause of the fault is comprehensively determined: Comparing illumination data with infrared thermal imaging data: If |D' t |>Θ D and Then the auxiliary judgment power drop is caused by light blocking, If |D' t |>Θ D and The auxiliary determines that the power drop is caused by a panel failure.

Images