A Fault Detection Method and System for Photovoltaic Energy Storage Equipment

Abstract

The present invention relates to the technical field of data processing, and more specifically, the present invention relates to a fault detection method and system for a photovoltaic energy storage device, including: collecting the operating parameters of the light energy storage device, where the operating parameters include light intensity, power generation, current, and voltage; obtaining the light intensity at the current moment, and selecting the moment with the same light intensity as it from the historical operating parameters, and constructing the light intensity at this moment and the previous N moments into a light intensity sequence. The present invention obtains the standard operating parameters of the light intensity according to the light intensity of any photovoltaic energy storage device at the current moment, taking into account the influence of the light intensity on the voltage and current, and then judges whether the photovoltaic energy storage device has a fault according to the difference value between the standard operating parameters and the actual operating parameters, solving the problem that the voltage and current data of the photovoltaic energy storage device are greatly affected by the light intensity, resulting in inaccurate fault detection.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing. More specifically, the present invention relates to a fault detection method and system for photovoltaic energy storage devices. Background Art

[0002] As the proportion of photovoltaic energy storage systems in the global energy structure continues to rise, in order to meet the power supply requirements of large-scale and high stability, multiple photovoltaic energy storage devices need to be deployed to generate electricity collaboratively in many application scenarios. However, due to differences in the installation angles and geographical locations of each device, the light intensities received by different devices at the same moment show significant differences. For example, the undulating mountain terrain causes some devices to be in shaded areas for a long time, and the devices in urban building complexes have different light exposure durations and intensities due to different orientations. This non-uniformity of light conditions makes the operating parameters such as voltage and current output by multiple photovoltaic energy storage devices fluctuate greatly even under normal operating conditions.

[0003] Currently, the fault detection methods for photovoltaic energy storage devices based on voltage and current are mainly divided into the threshold judgment method and the intelligent detection method based on data analysis. The threshold judgment method determines a fault when the real-time data exceeds the threshold by setting the normal ranges of voltage and current. However, in the scenario of multiple devices operating collaboratively, it is difficult for a fixed threshold to take into account the normal operating fluctuations of devices under different light intensities, and false alarms and missed detections are likely to occur. For example, when the light intensity changes drastically in the early morning or evening, or when the light is uneven due to different installation conditions, the output voltage and current of the photovoltaic system will have large fluctuations within the normal range and are often misjudged as faults. Therefore, in actual operation, the voltage and current data of photovoltaic energy storage devices are greatly affected by the light intensity, and the data distribution is complex and changeable, resulting in low accuracy of fault detection for photovoltaic energy storage devices. Summary of the Invention

[0004] The present invention provides a fault detection method and system for photovoltaic energy storage devices, aiming to solve the problem in the related art that the voltage and current data of photovoltaic energy storage devices are greatly affected by the light intensity, and the data distribution is complex and changeable, resulting in low accuracy of fault detection for photovoltaic energy storage devices.

[0005] In a first aspect, the present invention provides a fault detection method for a photovoltaic energy storage device, comprising: collecting operation parameters of the light energy storage device, where the operation parameters include light intensity and power generation; obtaining the light intensity at the current moment, selecting from the historical operation parameters the moments with the same light intensity as it, and constructing a light intensity sequence with the light intensity at this moment and the previous N moments, so as to obtain all light intensity sequences under the same light intensity as the current moment; correcting all light intensity sequences to obtain a plurality of normal light intensity sequences, where correcting the light intensity sequence includes: after segmenting each light intensity sequence by using ordered sample clustering, comparing the light intensity variance and variance threshold in each data segment to obtain the abnormal fluctuation segments and stable segments in each light intensity sequence, and determining whether the abnormal fluctuation segments are valid bands or invalid bands, where the valid bands or invalid bands reflect whether they will affect the total power generation of the light intensity sequence, and correcting the light intensity corresponding to the invalid band in any light intensity sequence to the light intensity mean value of any adjacent stable segment; clustering all normal light intensity sequences to obtain a plurality of clustering clusters, obtaining the average value of the corresponding operation parameters of each normal light intensity sequence in the clustering cluster to which the current moment light intensity sequence belongs as the standard operation parameter under this light intensity, and calculating the fault degree at the current moment based on the standard operation parameter and the actual operation parameter under the current moment light intensity, so as to detect the photovoltaic energy storage device. By constructing a light intensity sequence based on historical data and performing similarity analysis, various types of normal operation states are identified through clustering, so as to accurately match the difference between the current operation state and the standard state under the same light intensity condition. This method has the advantages of strong adaptability and high detection accuracy, improving the accuracy and real-time performance of fault detection.

[0006] Further, detecting the photovoltaic energy storage device includes: if the fault degree of the photovoltaic energy storage device at the current moment is greater than the fault threshold, determining that the photovoltaic energy storage device has a fault, and then giving an alarm to warn the staff.

[0007] Further, calculating the fault degree at the current moment includes: calculating the fault degree at the current moment according to at least one of the operation parameters.

[0008] Further, calculating the fault degree at the current moment according to at least one of the operation parameters includes: selecting any one of the operation parameters as the target parameter, calculating the difference between the actual target parameter at the current moment light intensity and the standard target parameter under the same light intensity, and taking the normalized difference as the fault degree at the current moment. Based only on a single parameter, the algorithm complexity is low and the calculation speed is fast, which is suitable for scenarios with high real-time requirements.

[0009] Further, calculating the fault degree at the current moment according to at least one of the operating parameters includes: selecting any two of the operating parameters as the first target parameter and the second target parameter, respectively calculating the first difference between the actual first target parameter at the current moment's light intensity and the standard first target parameter at the same light intensity, and the second difference between the actual second target parameter and the standard second target parameter, and taking the mean of the normalized first difference and second difference as the fault degree at the current moment. By fusing the two differences with the mean, the risk of a single parameter being affected by noise or abnormal fluctuations is reduced, and the stability of the fault degree estimation is improved.

[0010] Further, determining whether an abnormal fluctuation segment is a valid band or an invalid band includes: determining whether this abnormal fluctuation segment will affect the total power generation. If it has an impact, this abnormal fluctuation segment is regarded as a valid band; if it has no impact, this abnormal fluctuation segment is regarded as an invalid band. When analyzing a large amount of power generation data, this method can quickly screen out valuable information, that is, those abnormal fluctuation segments that affect the total power generation, eliminate the interference of irrelevant data, and improve the efficiency and accuracy of data analysis.

[0011] Further, obtaining the cluster to which the light intensity sequence at the current moment belongs includes: calculating the similarity between the light intensity sequence at the current moment and the standard light intensity sequence of any cluster, where the standard light intensity sequence of each cluster is the mean of all normal light intensity sequences in this cluster, and determining the one with the highest similarity as the cluster to which the light intensity sequence at the current moment belongs.

[0012] Further, calculating the similarity between the light intensity sequence at the current moment and the normal initial sequence of any type, and the calculation formula is: ; in the formula, represents the similarity between the light intensity sequence at the current moment and the standard light intensity sequence of the th type, represents the light intensity value at the th moment in the current light intensity sequence, represents the th type of standard light intensity sequence at the th moment of the light intensity value, represents the number of moments in the light intensity sequence, represents the standard normalization function. The formula takes into account the values of the light intensity sequence at multiple moments, comprehensively considering the changes in the light intensity sequence at different time points, rather than just the light intensity value at a single moment. This can more comprehensively capture the characteristics of the light intensity sequence, including the change trend and fluctuation of the light intensity, so as to more accurately evaluate its similarity with the normal initial sequence.

[0013] Further, cluster all normal light intensity sequences, including: clustering all normal light intensity sequences using the K-means clustering algorithm.

[0014] In a second aspect of the present invention, there is also provided a fault detection system for a photovoltaic energy storage device, including a processor and a memory. The memory stores a computer program, and the processor executes the computer program to implement the fault detection method for a photovoltaic energy storage device described in any one of the above.

[0015] Beneficial effects: According to the current light intensity of any photovoltaic energy storage device, obtain the standard operating parameters of this light intensity, taking into account the influence of light intensity on voltage and current, and then judge whether the photovoltaic energy storage device has a fault according to the difference value between the standard operating parameters and the actual operating parameters, solving the problem that the voltage and current data of the photovoltaic energy storage device are greatly affected by the light intensity factor, resulting in inaccurate fault detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flowchart schematically showing the calculation of the fault degree according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] The following will describe in detail the specific embodiments of the present invention with reference to the accompanying drawings.

[0018] Step S101: Collect the operating parameters of the light energy storage device.

[0019] In one embodiment, in a photovoltaic power generation group, obtain the operating parameters of each photovoltaic energy storage device, where the operating parameters include light intensity, power generation, current, and voltage; it should be noted that the light intensity refers to the solar power received per unit area, and a light sensor (such as a photoresistor, a silicon photocell, or a professional photometer) can be used to measure the light intensity. A voltage sensor and a current sensor are used to collect the voltage and current of the photovoltaic energy storage device respectively, and an electric energy meter or a power meter is used to measure the electric energy output by the photovoltaic system.

[0020] S102: Construct a light intensity sequence under the same light intensity.

[0021] In one embodiment, when the light intensity changes drastically in the early morning or evening, or when the light is uneven due to different installation conditions, the output voltage and current of the photovoltaic system will experience large fluctuations within the normal range, which are often misjudged as faults. Therefore, in actual operation, the voltage and current data of the photovoltaic energy storage device are greatly affected by environmental factors (especially the difference in light intensity). Therefore, it is necessary to take into account the normal operation fluctuations of the voltage and current data of the device under the same light intensity and calculate the standard operation parameters under this light intensity, and then judge whether there is a fault in the photovoltaic energy storage device according to the difference between the actual operation parameters at the current moment and the standard operation parameters under this light intensity.

[0022] Specifically, obtain the light intensity at the current moment, and select the moment with the same light intensity as the current moment from the historical operation parameters as the target moment, and the operation parameters (operation parameters include light intensity, power generation, current and voltage) are collected at each target moment. For any target moment, construct the light intensity sequence between the target moment and the previous N moments to obtain all the light intensity sequences under the same light intensity as the current moment. In this embodiment, the value of N is 5. In other embodiments, the value of N can be 4 or 6, etc., which can be adjusted according to the actual situation.

[0023] Exemplarily, the light intensity at the current moment is , if there are two moments with the same light intensity as the current moment in the historical operation parameters, which are and respectively, then the light intensity collected at all moments between and the previous is used as the light intensity sequence, and the light intensity collected at all moments between the previous is also the light intensity sequence. Thus, all the light intensity sequences under any light intensity can be obtained.

[0024] S103: Correct all the light intensity sequences to obtain the normal light intensity sequences.

[0025] In one embodiment, the light intensity in some light intensity sequences will change, but the duration of the change is short, resulting in no impact on the total power generation of the photovoltaic device in the end. Therefore, it is necessary to consider whether the change in the light intensity in the light intensity sequence at which moment will affect the total power generation. If it will not affect, this type of sequence can be determined as a normal category, which improves the accuracy of subsequent clustering, and at the same time improves the accuracy of the calculated standard normal sequence, thereby improving the accuracy of the detection result.

[0026] In one embodiment, a normal light intensity sequence is obtained, including: after segmenting each light intensity sequence by using ordered sample clustering, by comparing the light intensity variance and the variance threshold in each data segment, the abnormal fluctuation segments and the stable segments in each light intensity sequence are determined. Those with a light intensity variance greater than the variance threshold are taken as abnormal fluctuation segments, and vice versa as stable segments. The empirical value of the variance threshold is 60. In other embodiments, the empirical value of the variance threshold is 75 or 80, etc. Then it is determined whether the abnormal fluctuation segment will affect the total power generation. If it has an impact, the abnormal fluctuation segment is taken as an effective waveband. If it has no impact, the abnormal fluctuation segment is taken as an invalid waveband (that is, this abnormal fluctuation segment belongs to normal fluctuation and will not affect the total power generation). Then the light intensity corresponding to the invalid waveband in the light intensity sequence is corrected to the mean value of the light intensity of any adjacent stable segment, and the corrected light intensity sequence is used as the normal light intensity sequence, so as to reduce the influence of the invalid waveband on the clustering result during the subsequent clustering process and make the clustering result more accurate.

[0027] In one embodiment, determining whether the abnormal fluctuation segment is an effective waveband or an invalid waveband includes: collecting a plurality of data pairs, where a data pair is composed of any two light intensity sequences, and the light intensities at other times in the data pair are the same except for the abnormal fluctuation segment. Also, the duration of the dynamic segment in the abnormal fluctuation segment, the fluctuation amplitude of the abnormal fluctuation segment (where the fluctuation amplitude of the abnormal fluctuation segment is the difference between the mean value of the light intensity of the abnormal fluctuation segment and the mean value of the light intensity of the previous stable segment), and the label of the abnormal fluctuation segment (the label is 0 or 1. When the absolute value of the difference between the total power generations of the two light intensity sequences in the data pair is less than the difference threshold, the label is 0, and vice versa the label is 1. The empirical value of the difference threshold is 0.2; 0 indicates that it will not cause a change in the total power generation (invalid waveband), and 1 indicates that it will cause a change in the total power generation (effective waveband)). Then, using the fluctuation amplitude and duration of the abnormal fluctuation segment in the data pair, and the label to train a classification model. Stop when the cross-entropy loss function is less than the loss threshold or the number of training times is greater than the number of iterations to obtain the final classification model. Input the fluctuation amplitude and duration of the abnormal fluctuation segment into the final classification model, and the result is 1 or 0. 1 indicates an effective waveband, and 0 indicates an invalid waveband.

[0028] S104: Obtain the standard normal light intensity sequence of the current moment light intensity sequence.

[0029] In one embodiment, after screening all the normal light intensity sequences at any light intensity, cluster all the normal light intensity sequences to obtain a plurality of clustering clusters. One clustering cluster corresponds to one type of sequence, so as to obtain multiple types of normal light intensity sequences. The clustering algorithm is the K-means clustering algorithm. Then, the sequence composed of the mean values of all the normal light intensity sequences in each clustering cluster is used as the standard normal light intensity sequence of various types at this light intensity.

[0030] S105: Determine the standard operating parameters at the current moment's light intensity.

[0031] In one embodiment, calculate the similarity between the current moment's light intensity sequence and standard light intensity sequences of various types, determine the type with the highest similarity as the type to which the current moment's light intensity sequence belongs, and then take the average of the operating parameters corresponding to all normal light intensity sequences of the corresponding type at this light intensity as the standard operating parameters at this light intensity. The standard operating parameters include standard power generation, standard voltage, and standard current.

[0032] In one embodiment, a calculation formula for calculating the similarity between the current moment's light intensity sequence and any type of normal initial sequence is provided. The calculation formula is: ; where represents the similarity between the current moment's light intensity sequence and the th type of standard light intensity sequence, represents the light intensity value at the th moment in the current light intensity sequence, represents the th type of standard light intensity sequence at the th moment of the light intensity value, represents the number of moments in the light intensity sequence, represents the standard normalization function.

[0033] S106: Calculate the degree of fault of the current moment's photovoltaic energy storage device for fault detection.

[0034] In one embodiment, based on the difference value between the actual operating parameters at the current moment's light intensity and the standard operating parameters at the same light intensity, it is used as the degree of fault of the photovoltaic energy storage device at the current moment, and the photovoltaic energy storage device is detected for faults according to the size of the degree of fault. If the degree of fault of the photovoltaic energy storage device at the current moment is greater than the fault threshold, it is determined that the photovoltaic energy storage device has a fault, and an alarm is given to warn the staff. If the degree of fault of the photovoltaic energy storage device at the current moment is less than or equal to the fault threshold, it is determined that the photovoltaic energy storage device has no fault, and the photovoltaic energy storage device can continue to work. Among them, the empirical value of the fault threshold is 0.5. In other embodiments, the empirical value of the fault threshold can be 0.6 or 0.55, which can be adjusted according to the specific implementation situation.

[0035] It should be noted that when calculating the degree of fault of the photovoltaic energy storage device at the current moment, the degree of fault at the current moment can be calculated according to at least one of the operating parameters, or according to two or three of the operating parameters.

[0036] In one embodiment, calculating the fault degree at the current moment according to at least one of the operating parameters includes: selecting any one of the operating parameters as the target parameter, calculating the difference between the actual target parameter at the current moment under the illumination intensity and the standard target parameter under the same illumination intensity, and taking the normalized difference as the fault degree at the current moment.

[0037] Exemplarily, calculating the fault degree at the current moment according to the voltage among the operating parameters, calculating the difference between the actual voltage at the current moment under the illumination intensity and the standard voltage under the same illumination intensity, and taking the normalized difference as the fault degree at the current moment.

[0038] In another embodiment, calculating the fault degree at the current moment according to any two of the operating parameters includes: selecting any two of the operating parameters as the first target parameter and the second target parameter, respectively calculating the first difference between the actual first target parameter at the current moment under the illumination intensity and the standard first target parameter under the same illumination intensity, and the second difference between the actual second target parameter and the standard second target parameter, and taking the mean value of the normalized first difference and the second difference as the fault degree at the current moment.

[0039] Exemplarily, selecting the current and voltage among the operating parameters as the first target parameter and the second target parameter, respectively calculating the first difference between the actual voltage at the current moment under the illumination intensity and the standard voltage under the same illumination intensity, and the second difference between the actual current and the standard current, and taking the mean value of the normalized first difference and the second difference as the fault degree at the current moment.

[0040] So far, by collecting operating parameters such as illumination intensity and power generation, constructing an illumination intensity sequence with the illumination intensity as the link, obtaining the standard illumination intensity sequence through screening and clustering, and calculating the similarity to determine the type of the current illumination intensity sequence, the illumination intensity characteristic law under the normal operation mode can be accurately grasped. On this basis, with the standard operating parameters as a reference, comparing the actual operating parameters with them, and quantifying the fault degree with the difference value, the accurate detection of the faults of the photovoltaic energy storage device is realized.

[0041] The present invention also provides a fault detection system for a photovoltaic energy storage device. The system includes a processor and a memory, and the memory stores computer program instructions. When the computer program instructions are executed by the processor, a fault detection method for a photovoltaic energy storage device according to the first aspect of the present invention is implemented.

[0042] The system further includes a communication bus, a communication interface and other components well known to those skilled in the art. Their settings and functions are known in the art, so they will not be elaborated here.

[0043] In the present invention, the foregoing memory may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory (RRAM), a dynamic random access memory (DRAM), a static random access memory (SRAM), an enhanced dynamic random access memory (EDRAM), a high-bandwidth memory (HBM), a hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions stored or otherwise held by such a computer-readable medium.

[0044] The above-described embodiments merely represent several implementation manners of the present invention. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention.

Claims

1. A fault detection method for a photovoltaic energy storage device, characterized in that, including: collecting the operating parameters of the light energy storage device, where the operating parameters include light intensity and power generation; obtaining the light intensity at the current moment, selecting the moment with the same light intensity as it from the historical operating parameters, and constructing the light intensity at this moment and the light intensities of the previous N moments into a light intensity sequence to obtain all light intensity sequences under the same light intensity as the current moment; correcting all light intensity sequences to obtain multiple normal light intensity sequences, where correcting the light intensity sequence includes: after segmenting each light intensity sequence by using ordered sample clustering, comparing the light intensity variance and the variance threshold in each data segment to obtain the abnormal fluctuation segments and stable segments in each light intensity sequence, and determining whether the abnormal fluctuation segment is an effective band or an ineffective band, where the effective band or ineffective band reflects whether it will affect the total power generation of the light intensity sequence, and correcting the light intensity corresponding to the ineffective band in any light intensity sequence to the average value of the light intensities of any adjacent stable segment; clustering all normal light intensity sequences to obtain multiple clustering clusters, obtaining the average value of the corresponding operating parameters of each normal light intensity sequence in the clustering cluster to which the current moment light intensity sequence belongs as the standard operating parameter under this light intensity, and calculating the fault degree at the current moment based on the standard operating parameter and the actual operating parameter under the current moment light intensity to detect the photovoltaic energy storage device.

2. The fault detection method for a photovoltaic energy storage device according to claim 1, wherein detecting the photovoltaic energy storage device, including: if the fault degree of the photovoltaic energy storage device at the current moment is greater than the fault threshold, determining that the photovoltaic energy storage device has a fault, and then giving an alarm to warn the staff.

3. The fault detection method for a photovoltaic energy storage device according to claim 1, characterized in that, calculating the fault degree at the current moment, including: calculating the fault degree at the current moment according to at least one of the operating parameters.

4. The fault detection method for a photovoltaic energy storage device according to claim 3, characterized in that calculating the fault degree at the current moment according to at least one of the operating parameters, including: selecting any one of the operating parameters as the target parameter, calculating the difference between the actual target parameter at the current moment light intensity and the standard target parameter under the same light intensity, and taking the normalized difference as the fault degree at the current moment.

5. The fault detection method for a photovoltaic energy storage device according to claim 3, wherein, calculating the fault degree at the current moment according to at least one of the operating parameters, including: selecting any two of the operating parameters as the first target parameter and the second target parameter, respectively calculating the first difference between the actual first target parameter at the current moment light intensity and the standard first target parameter under the same light intensity, and the second difference between the actual second target parameter and the standard second target parameter, and taking the average value of the normalized first difference and the second difference as the fault degree at the current moment.

6. The fault detection method for a photovoltaic energy storage device according to claim 1, wherein determining whether the abnormal fluctuation segment is an effective band or an ineffective band, including: determining whether this abnormal fluctuation segment will affect the total power generation. If it has an impact, taking this abnormal fluctuation segment as an effective band. If it has no impact, taking this abnormal fluctuation segment as an ineffective band.

7. The fault detection method for a photovoltaic energy storage device according to claim 6, characterized in that, obtaining the clustering cluster to which the current moment light intensity sequence belongs, including: calculating the similarity between the current moment light intensity sequence and the standard light intensity sequence of any clustering cluster, where the standard light intensity sequence of each clustering cluster is the average value of all normal light intensity sequences in this clustering, and determining the one with the highest similarity as the clustering cluster to which the current moment light intensity sequence belongs.

8. The fault detection method for a photovoltaic energy storage device according to claim 7, characterized in that Calculate the similarity between the current light intensity sequence and the standard light intensity sequence of any clustering cluster. The calculation formula is as follows: ； Wherein, represents the similarity between the current light intensity sequence and the th type of standard light intensity sequence, represents the light intensity value at the th moment in the current light intensity sequence, represents the th type of standard light intensity sequence at the th moment of the light intensity value, represents the number of moments in the light intensity sequence, represents the standard normalization function.

9. The fault detection method for a photovoltaic energy storage device according to claim 1, wherein, Cluster all normal light intensity sequences, including: Use the K-means clustering algorithm to cluster all normal light intensity sequences.

10. A fault detection system for a photovoltaic energy storage device, comprising a processor and a memory, characterized in that, The memory stores a computer program, and the processor executes the computer program to implement the fault detection method for a photovoltaic energy storage device according to any one of claims 1-9.

Images