Abnormality detection method, device and system for photovoltaic power station control system

Abstract

The present invention provides an abnormality detection method, device and system for a photovoltaic power station control system, the method comprising: obtaining weather perception data of each sensor and its adjacent sensors in a detection time period; judging whether the weather perception data of the sensor is abnormal according to the weather perception data of the sensor and its adjacent sensors in the detection time period; if the weather perception data of the sensor is abnormal in a plurality of consecutive detection time periods, judging that the sensor is operating abnormally, and performing relevant fault warning. The present invention can detect the operation abnormality of the sensor online, improve the accuracy of the abnormality detection of the tracker, and improve the comprehensiveness of the abnormality detection of the whole station.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation, and in particular to an abnormality detection method, device and system for a photovoltaic power station control system. Background Art

[0002] Usually, a photovoltaic power station includes several photovoltaic sub-arrays, each of which is equipped with several control boxes, one communication box and one to several meteorological sensors (such as anemometers or snow depth, flood sensors, etc.). One to several trackers share one control box, and each control box communicates with the communication box of the photovoltaic sub-array via wired or wireless communication.

[0003] The communication box of the photovoltaic sub-array collects information such as wind speed and direction through anemometers, and collects snow depth and flood level status through snow depth and flood sensors. It makes judgments based on the above information collected and notifies the control box when it senses disaster weather. The control box adjusts the operating posture of the tracker based on the received information.

[0004] The current mainstream control mode is that the communication box of each sub-array collects data independently and manages the operating posture of the tracker of this sub-array independently based on the collected data.

[0005] Since the number of sensors deployed in a subarray is limited, it is difficult to identify sensor operation anomalies online based on the control mode of a single subarray. Sensors such as anemometers, flood sensors, and snow depth sensors are electronic mechanical devices. Due to sand in the bearings, bearing wear, and other conditions, sensors may fail. Currently, sensor failures can only be discovered by performing offline calibration of sensors every year. For example, anemometers need to be returned to the manufacturer for wind tunnel testing or repairs, which has a long actual operation cycle and is difficult to implement. Customers usually do not have the means to calibrate anemometers and other sensors on site at the photovoltaic power station, and cannot promptly determine whether the sensors are faulty, require maintenance, or need to be repaired.

[0006] In addition, the abnormal sensor failed to sense the disastrous weather in time, resulting in the inability to adjust the tracker of its sub-array to the disaster protection posture in time, which may cause greater losses due to the delayed response.

[0007] For example, a certain anemometer may have a fault and the wind speed reading may be too low, but the power station operation and maintenance personnel cannot promptly determine the data error of the anemometer, resulting in delayed response or failure to issue early warnings during wind disasters. Another example is when a flood sensor detects a flood, due to the limited detection range of the sensor itself, when the flood level in this area varies greatly at various points due to factors such as surface undulations, some trackers in this area may have already touched water. Summary of the invention

[0008] One of the purposes of the present invention is to overcome at least some of the deficiencies in the prior art and to provide an abnormality detection method, device and system for a photovoltaic power station control system.

[0009] The technical solution provided by the present invention is as follows:

[0010] A photovoltaic power station control system anomaly detection method is applied to a photovoltaic power station including a plurality of photovoltaic sub-arrays, wherein the photovoltaic power station control system includes a sensor for sensing the weather in an area where each photovoltaic sub-array is located; the anomaly detection method includes:

[0011] Obtain weather perception data of each sensor and its neighboring sensors during a detection period;

[0012] Determining whether the weather perception data of the sensor is abnormal based on the weather perception data of the sensor and its neighboring sensors during the detection period;

[0013] If the weather sensing data of the sensor is abnormal within multiple consecutive detection time periods, it is determined that the sensor is operating abnormally and a relevant fault warning is issued.

[0014] In some embodiments, judging whether the weather perception data of the sensor is abnormal according to the weather perception data of the sensor and its neighboring sensors in a detection time period includes:

[0015] Calculate the mean and standard deviation of the weather perception data according to the weather perception data of the sensor and its neighboring sensors during the detection period;

[0016] Based on the mean and standard deviation, a reasonable interval is constructed according to the normal distribution rule;

[0017] If the weather sensing data of the sensor falls outside the reasonable range, the weather sensing data of the sensor is abnormal.

[0018] In some embodiments, it also includes:

[0019] Dividing the photovoltaic power station into several areas in advance;

[0020] If the number of sensors detecting disastrous weather in an area reaches a preset ratio relative to the total number of sensors in the area within a preset time range, a disastrous weather warning will be issued in the area.

[0021] In some embodiments, it also includes:

[0022] Grouping the trackers in the photovoltaic power station, grouping the trackers with the same design and installed in close geographical locations into one group;

[0023] Obtain the operating data of each tracker during the detection period;

[0024] According to the operation data of each tracker and other trackers in its tracker group during the detection period, it is determined whether the operation of the tracker is abnormal; if the tracker is abnormal, a relevant fault warning is issued.

[0025] In some embodiments, judging whether the operation of each tracker is abnormal according to the operation data of each tracker and other trackers of the tracker group to which it belongs during the detection time period includes:

[0026] Analyze the operation data of the tracker in the detection time period to determine whether the tracker has abnormal fluctuations and the time point of occurrence of the abnormal fluctuations when the abnormal fluctuations occur;

[0027] Divide the detection time period into a number of time sub-segments, traverse each of the time sub-segments, and compare and analyze the operation data of each tracker in the tracker group to which the tracker belongs in the same time sub-segment to determine whether there is a heterogeneous tracker in the tracker group in each time sub-segment and the number of the heterogeneous tracker if there is a heterogeneous tracker; the heterogeneous tracker is a tracker that operates significantly differently from most trackers in the tracker group;

[0028] If the tracker generates data with abnormal fluctuations within the detection time period, and the tracker group to which the tracker belongs has a heterogeneous tracker in the time sub-segment to which the abnormal fluctuation occurs, and the heterogeneous tracker includes the tracker, then it is determined that the tracker is operating abnormally.

[0029] In some embodiments, obtaining operation data of a tracker during a detection period includes:

[0030] Acquiring current data and bracket angle data of a tracker during a detection period;

[0031] According to the current data and the bracket angle data of the tracker at any two adjacent time points within the detection time period, the current change data required for the unit angle change of the tracker at the corresponding time point is obtained;

[0032] The current change data required by the tracker for the unit angle change at each time point in the detection time period constitutes the operation data of the tracker in the detection time period.

[0033] In some embodiments, the SmoothedZ-score algorithm is used to analyze the operating data of each tracker in the detection time period to determine whether the tracker has abnormal fluctuations and the time point when the abnormal fluctuations occur.

[0034] In some embodiments, comparative analysis is performed on the operation data of each tracker in the tracker group to which the tracker belongs in the same time sub-segment to determine whether there is a heterogeneous tracker in the tracker group in each time sub-segment and the number of the heterogeneous tracker when there is a heterogeneous tracker, including:

[0035] Acquire the operation data of each tracker in the tracker group in the time subsegment;

[0036] Calculate the mean and standard deviation of the operating data according to the operating data of each tracker in the tracker group in the time sub-segment;

[0037] Based on the mean and standard deviation, a reasonable interval is constructed according to the normal distribution rule;

[0038] The operation data of each tracker in the tracker group in the time subsegment is traversed. If the operation data of the tracker in the time subsegment falls outside the reasonable interval, then there is a heterogeneous tracker in the tracker group in the time subsegment, and the tracker is the heterogeneous tracker.

[0039] The present invention also provides an abnormality detection device for a photovoltaic power station control system, comprising:

[0040] Memory for storing computer programs;

[0041] A processor is used to implement the steps of any of the above-mentioned methods for detecting anomalies in a photovoltaic power station control system when running the computer program.

[0042] The present invention also provides a photovoltaic power station control system, comprising a control center and a plurality of control subsystems, each control subsystem being used to control a photovoltaic subarray;

[0043] The control subsystem includes a communication box, a sensor and several control boxes; the control subsystem communicates with the control center through the communication box;

[0044] The communication box is electrically connected to the sensor and is used to collect weather information of the area where the corresponding photovoltaic sub-array is located through the sensor;

[0045] The communication box is also connected to the control box by wire or wirelessly, and is used to collect current data and bracket angle data of the corresponding tracker in the photovoltaic sub-array through the control box;

[0046] The control center is used to execute any of the above-mentioned methods for detecting anomalies in a photovoltaic power station control system.

[0047] The abnormality detection method, device and system of a photovoltaic power station control system provided by the present invention can at least bring the following beneficial effects:

[0048] 1. Compared with the traditional single-subarray independent anomaly detection mode, the present invention can identify the operating anomalies of the sensor online by combining sensors of other subarrays and comparing the weather perception data of the sensor to be detected with those of its neighboring sensors.

[0049] 2. The present invention improves the accuracy of abnormality detection of the tracker by monitoring whether the running data of the tracker itself has sudden changes and comparing it with the running data of other similar trackers in the tracker group to which it belongs.

[0050] 3. The present invention improves the comprehensiveness of abnormality detection of the entire station by discovering that a certain proportion of sensors have issued warnings about disastrous weather and notifying the sub-arrays that have not detected extreme weather to enter protection mode. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The preferred implementation scheme will be described below in a clear and understandable manner with reference to the accompanying drawings to further illustrate the above characteristics, technical features, advantages and implementation methods of an abnormality detection method for a photovoltaic power station control system.

[0052] Figure 1 It is a flow chart of an embodiment of an abnormality detection method of a photovoltaic power station control system of the present invention;

[0053] Figure 2 It is a structural schematic diagram of an embodiment of an abnormality detection device of a photovoltaic power station control system of the present invention;

[0054] Figure 3 It is a structural schematic diagram of an embodiment of a photovoltaic power station control system of the present invention;

[0055] Figure 4 It is a structural schematic diagram of another embodiment of a photovoltaic power station control system of the present invention. DETAILED DESCRIPTION

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the specific implementation methods of the present invention will be described below with reference to the accompanying drawings. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings and other implementation methods can be obtained based on these drawings without creative work.

[0057] In order to simplify the drawings, only the parts related to the present invention are schematically shown in each figure, and they do not represent the actual structure of the product. In addition, in order to simplify the drawings and facilitate understanding, in some figures, only one of the parts with the same structure or function is schematically drawn or marked. In this article, "one" not only means "only one", but also means "more than one".

[0058] In one embodiment of the present invention, Figure 1 As shown, a method for detecting anomalies in a photovoltaic power station control system is applied to a photovoltaic power station including a plurality of photovoltaic sub-arrays, wherein the control system of the photovoltaic power station includes a sensor for sensing the weather in the area where each photovoltaic sub-array is located. The method for detecting anomalies includes:

[0059] Step S100 obtains weather sensing data of each sensor and its neighboring sensors during a detection period;

[0060] Step S110 determines whether the weather sensing data of a sensor is abnormal based on the weather sensing data of the sensor and its neighboring sensors in the detection time period;

[0061] In step S120, if the weather sensing data of the sensor is abnormal in multiple consecutive detection time periods, it is determined that the sensor is operating abnormally and a related fault warning is issued.

[0062] Specifically, the sensor is a meteorological sensor used to measure the weather in the area, including anemometers, snow depth sensors, flood sensors, etc. Anemometers are used to measure wind speed and wind direction, snow depth sensors are used to measure snow depth, and flood sensors are used to measure flood levels. Photovoltaic subarrays are generally equipped with anemometers, and snow depth sensors and flood sensors can be selected according to actual conditions.

[0063] The neighboring sensors of the sensor to be detected are sensors within a preset distance range from the sensor to be detected under the photovoltaic power station. If the photovoltaic power station is small, all other sensors in the entire station can be used as neighboring sensors of the sensor to be detected. The neighboring sensors and the sensor to be detected may be in the same photovoltaic subarray or across subarrays.

[0064] Since the number of sensors deployed in a subarray is limited (for example, only one anemometer is deployed in a subarray), it is difficult to identify sensor operation anomalies online based on the traditional single subarray independent anomaly detection mode, and sensor failures can only be discovered by routine offline calibration of the sensor. This embodiment combines sensors from other subarrays to determine the neighboring sensors of the sensor to be detected within the entire power station, and compares the weather perception data of the sensor to be detected and its neighboring sensors, so as to identify the sensor with abnormal operation online and issue an early warning.

[0065] In one embodiment, it further includes:

[0066] Step S200 pre-divides the photovoltaic power station into several areas;

[0067] In step S210, if the number of sensors detecting that disastrous weather has occurred in a region reaches a preset ratio relative to the total number of sensors in the region within a preset time range, a disastrous weather warning is issued in the region.

[0068] Specifically, disaster weather includes strong winds, heavy snow, floods, etc. When disastrous weather comes, some sensors may fail to detect extreme weather due to factors such as their placement and external interference. When a certain proportion of sensors detect extreme weather in a short period of time, the sub-arrays that have not detected extreme weather are notified to start the strong wind, heavy snow or flood mode, to avoid the failure to start the trackers under the sub-array to enter the corresponding protection mode due to the abnormality of some sensors.

[0069] The entire power station can be used as one area, or divided into multiple areas according to actual conditions. If the number of sensors that detect disastrous weather in the same area reaches a preset ratio, a disastrous weather warning can be activated in the area, allowing the trackers of all photovoltaic sub-arrays in the area to enter the corresponding protection mode.

[0070] In one embodiment, it further includes:

[0071] Step S300 groups the trackers in the photovoltaic power station, and groups the trackers with the same design and installed in close geographical locations into one group.

[0072] Different trackers may control different PV modules and use different drive motors, which will result in incomparability of their operating data. For this reason, trackers with the same design and installed in a preset geographical location are generally grouped together. The basic principle of grouping is to ensure that the operating rules of each tracker in the group are the same and the operating data are similar, which can be used for engineering comparison to identify whether the operating data of certain trackers is abnormal. Trackers in the same group can belong to the same PV sub-array or across PV sub-arrays.

[0073] Step S310 obtains the operating data of each tracker during the detection period.

[0074] The operation data of the tracker may be the driving current data of the motor and / or the bracket angle data, or may be data derived from the driving current data of the motor and the bracket angle data, or other data that can reflect the operation status of the tracker.

[0075] The control box collects the tracker's current data, bracket angle data, or other data that can reflect the tracker's operating status. The PV subarray's communication box regularly collects this data from the subordinate control box, or the control box actively reports this data to the communication box. These data or their derivative data are used as the tracker's operating data. The derivative data can be generated by the control box or the communication box.

[0076] Step S320 determines whether the operation of each tracker is abnormal based on the operation data of each tracker and other trackers in its tracker group during the detection period; if the tracker is abnormal, a related fault warning is issued.

[0077] Specifically, based on the operation data of a single tracker over a period of time, it is possible to detect abnormal operation of the tracker. However, this abnormality is not necessarily caused by a tracker failure. For example, in a sudden strong wind, the operation data of the tracker will fluctuate abnormally compared to normal times. In order to improve the accuracy of tracker fault detection, the operation data of the tracker is compared with the operation data of other trackers in the tracker group to which it belongs; if the operation difference with most trackers is large, it is judged that the tracker may be faulty.

[0078] Compared with the traditional single-subarray independent anomaly detection mode, this embodiment combines sensors of other subarrays and compares the weather perception data of the sensor to be detected with those of its neighboring sensors, so as to identify the operation anomaly of the sensor online; by comparing the operation data of the tracker with that of the tracker group, the accuracy of the tracker's anomaly detection can be improved; by discovering that a certain proportion of sensors issue warnings for disastrous weather, the photovoltaic subarrays that have not detected extreme weather are notified to enter the protection mode, thereby improving the comprehensiveness of the anomaly detection of the entire station.

[0079] In one embodiment, step S110 includes:

[0080] Step S111 calculates the mean and standard deviation of the weather perception data according to the weather perception data of the sensor and its neighboring sensors during the detection period;

[0081] Step S112 constructs a reasonable interval according to the mean and standard deviation and the normal distribution rule;

[0082] In step S113, if the weather sensing data of the sensor falls outside a reasonable range, the weather sensing data of the sensor is abnormal.

[0083] For example, the reasonable interval is set to [u-3σ, u+3σ], where u is the mean and σ is the standard deviation. Data outside the reasonable interval are outliers or outliers.

[0084] In one embodiment, obtaining the operation data of a tracker in the detection time period in step S310 includes:

[0085] Step S311 obtains current data and bracket angle data of a tracker during a detection period;

[0086] Step S312 obtains the current change data required for the unit angle change of the tracker at the corresponding time point according to the current data and the bracket angle data of any two adjacent time points of the tracker within the detection time period;

[0087] Step S313: The current change data required by the tracker for the unit angle change at each time point in the detection time period constitutes the operation data of the tracker in the detection time period.

[0088] Specifically, based on the current data I of the tracker and the bracket angle data α, the formula The current change data β required for the unit angle change is obtained. β can significantly represent the change in the amplitude of the current when the system resistance changes, which is helpful to find abnormal points in abnormal data detection. Therefore, using β as operating data is conducive to improving the accuracy of abnormal data detection.

[0089] In one embodiment, step S313 includes:

[0090] In step S3131, the current change data required for the unit angle change of the tracker at each time point in the detection time period are amplified according to preset rules, and the amplified data constitute the operation data of the tracker in the detection time period.

[0091] In order to amplify the difference more obviously, the current change data β required for the unit angle change is further amplified, for example, according to γ ​​= a*β δ + b, calculate γ, and use γ as the operating data. Among them, a, b, and δ are preset parameters.

[0092] In one embodiment, step S320 includes:

[0093] Step S321 analyzes the running data of the tracker in the detection time period to determine whether the tracker has abnormal fluctuations and the time point of the abnormal fluctuations when the abnormal fluctuations occur.

[0094] Under normal circumstances, the running data of the tracker fluctuates within a certain range. However, when the tracker fails or is disturbed by external factors (for example, a sudden strong wind), the running data will fluctuate abnormally. Abnormal fluctuation means that the data fluctuation exceeds the normal fluctuation range. The normal fluctuation range can be obtained based on the historical running data under normal circumstances. Perform anomaly detection on the running data of the tracker over a period of time. If there are abnormal fluctuation points in the running data, it means that the tracker has abnormal fluctuations.

[0095] Step S322 divides the detection time period into several time sub-segments, traverses each time sub-segment, and compares and analyzes the operation data of each tracker in the tracker group to which the tracker belongs in the same time sub-segment to determine whether there is a heterogeneous tracker in the tracker group of each time sub-segment and the number of the heterogeneous tracker when there is a heterogeneous tracker; the heterogeneous tracker is a tracker that operates significantly differently from most trackers in the tracker group.

[0096] Specifically, although the operation paces of the trackers belonging to the same tracker group are similar, they are not completely consistent. Therefore, instead of directly comparing and analyzing the operation data of each tracker at the same time point, the operation data within a certain time tolerance (such as 3 minutes) is used as the analysis data. This time tolerance is set based on experience.

[0097] Divide the detection time period into several time sub-segments according to the time tolerance. Analyze the operation data of each tracker in each time sub-segment and find outliers. For example, if the detection time period is 30 minutes and the time tolerance is 3 minutes, there are 10 time sub-segments. Compare the tracker data within the group for each of the 10 time sub-segments.

[0098] If there is an outlier, it means there is a heterogeneous tracker. The number of the tracker to which the outlier belongs is the number of the heterogeneous tracker.

[0099] In step S323, if the tracker generates data with abnormal fluctuations within the detection time period, and the tracker group to which the tracker belongs has a heterogeneous tracker in the time sub-segment to which the abnormal fluctuation occurs, and the heterogeneous tracker includes the tracker, then it is determined that the tracker is operating abnormally.

[0100] Specifically, there are two types of abnormal data detection, the first type is the context time series abnormality detection of a single tracker, which is completed by step S321, and the second type is the data abnormality detection within the tracker group, which is completed by step S322.

[0101] Step S321 is to complete data anomaly detection of a single tracker with the tracker as a unit. The detection result of the tracker includes whether there is an abnormal fluctuation. If there is an abnormal fluctuation, it also includes the time point when the abnormal fluctuation occurs.

[0102] Step S322 is to complete the detection of data anomalies within each time sub-segment in the tracker group. The detection result of the tracker group in each time sub-segment includes whether there is a heterogeneous tracker. The heterogeneous tracker is not limited to 1, and may be 2 or more. If there is a heterogeneous tracker, the number of the heterogeneous tracker is also included.

[0103] A tracker that detects an abnormality in only one type of abnormal data detection does not necessarily malfunction. For example, in windy weather, the tracker detects an abnormality in the first type of abnormal data detection, but it is not abnormal in the second type of abnormal data detection, that is, the tracker is normal. After the strong wind disappears, the operating data returns to normal.

[0104] If a tracker is detected as abnormal in both abnormal data detections, the tracker is an abnormal tracker. The abnormal tracker is determined by combining the detection results of the tracker and the detection results of the tracker group in each time sub-segment, which improves the judgment accuracy of the abnormal tracker.

[0105] In one embodiment, step S321 includes:

[0106] Step S3211 uses the SmoothedZ-score algorithm to analyze the running data of the tracker in the detection time period to determine whether the tracker has abnormal fluctuations and the time point when the abnormal fluctuations occur.

[0107] The SmoothedZ-score algorithm uses the most recent historical data to make predictions. If the actual value exceeds the predicted data range, the actual value is an outlier.

[0108] In one embodiment, step S322 compares and analyzes the operation data of each tracker in a tracker group in the same time sub-segment to determine whether there are heterogeneous trackers in the tracker group of each time sub-segment and the numbers of the heterogeneous trackers, including:

[0109] Step S3221 obtains the operation data of each tracker in the tracker group in the time sub-segment;

[0110] Step S3222 calculates the mean and standard deviation of the operating data according to the operating data of each tracker in the tracker group in the time sub-segment;

[0111] Step S3223 constructs a reasonable interval according to the mean and standard deviation and the normal distribution rule;

[0112] Step S3224 traverses the operation data of each tracker in the tracker group in the time sub-segment. If the operation data of the tracker in the time sub-segment falls outside the reasonable range, there is a heterogeneous tracker in the tracker group in the time sub-segment, and the tracker is a heterogeneous tracker.

[0113] In one embodiment, step S322 compares and analyzes the operation data of each tracker in a tracker group in the same time sub-segment to determine whether there are heterogeneous trackers in the tracker group of each time sub-segment and the numbers of the heterogeneous trackers, including:

[0114] Obtain the operation data of each tracker in the tracker group in the time sub-segment;

[0115] Sort the operating data of each tracker in the tracker group in the time sub-segment to obtain the first quartile and the third quartile;

[0116] Calculate the difference between the third quartile and the first quartile to get the interquartile range IQR;

[0117] According to the first quartile and the interquartile range IQR, the lower limit of the reasonable interval is obtained;

[0118] According to the third quartile and the interquartile range IQR, the upper limit of the reasonable interval is obtained;

[0119] The running data of each tracker in the tracker group in the time sub-segment is traversed. If the running data of the tracker in the time sub-segment falls outside the reasonable range, there is a heterogeneous tracker in the tracker group in the time sub-segment, and the tracker is a heterogeneous tracker.

[0120] Specifically, the quartile data calculation method is used. The interquartile range IQR = Q3-Q1, the lower limit of the reasonable interval (for example = Q1-1.5*IQR), and the upper limit of the reasonable interval (for example = Q3+1.5*IQR) are calculated, where 1.5 is a preset coefficient that can be adjusted as needed. Outliers are defined as data that is lower than the lower limit of the reasonable interval or higher than the upper limit of the reasonable interval.

[0121] In one embodiment of the present invention, Figure 2 As shown, an abnormality detection device 200 for a photovoltaic power station control system includes a memory 210 and a processor 220.

[0122] The memory 210 is used to store a computer program 230 .

[0123] When the processor 220 runs the computer program 230 , the steps of the abnormality detection method for the photovoltaic power station control system described in the above embodiment are implemented.

[0124] In one embodiment of the present invention, Figure 3 As shown, a photovoltaic power station control system includes a control center 10 and a plurality of control subsystems 20, each of which is used to control a photovoltaic subarray.

[0125] Each control subsystem 20 includes a communication box 21 , a sensor 22 and a plurality of control boxes 23 ; the control subsystem communicates with the control center 10 via the communication box 20 .

[0126] The communication box 21 is electrically connected to the sensor 21 and is used to collect weather information of the area where the corresponding photovoltaic sub-array is located through the sensor; it is also connected to the control box 23 by wire or wirelessly and is used to collect current data and bracket angle data of the corresponding tracker in the photovoltaic sub-array through the control box.

[0127] The control center 10 is used to execute the abnormality detection method of a photovoltaic power station control system described in the above embodiment.

[0128] Specifically, a photovoltaic power station includes several photovoltaic sub-arrays. Each photovoltaic sub-array is equipped with a communication box, sensors and several control boxes. One control box can control 1 to N trackers, and the tracker controls the photovoltaic components to rotate with the sun to obtain more radiation. Each control box communicates with the communication box of the photovoltaic sub-array via wired or wireless communication.

[0129] Sensors are used to measure weather information in the area where the photovoltaic sub-array is located, including anemometers, snow depth sensors (optional), and flood sensors (optional). Anemometers are used to measure wind speed and wind direction, snow depth sensors are used to measure snow depth, and flood sensors are used to measure flood levels. For each type of sensor, one or more sensors can be set in the photovoltaic sub-array, and they can be set in different locations. Based on the weather perception data collected by these sensors, disaster weather such as strong winds, heavy snow, and floods can be sensed.

[0130] The communication box can regularly inspect and collect the current and angle data of the tracker collected by each control box and the weather data collected by the sensor. The communication box can report these data to the control center, and the control center executes the abnormality detection method of the photovoltaic power station control system described in the above embodiment.

[0131] The present invention also provides a specific application scenario embodiment, in which the above-mentioned abnormality detection method of the photovoltaic power station control system is applied to a photovoltaic power station.

[0132] The control system structure of a photovoltaic power station is as follows: Figure 4 As shown:

[0133] The photovoltaic power station includes multiple photovoltaic sub-arrays and SCADA servers. Each photovoltaic sub-array includes a communication box, a meteorological sensor and several control boxes. Each control box controls a tracker, and the tracker controls the rotation of the photovoltaic module. The communication box of the photovoltaic sub-array is connected to each control box and meteorological sensor in the photovoltaic sub-array through wireless or wired. The SCADA server is connected to each communication box of the photovoltaic sub-array, and the SCADA server acts as a control center. The meteorological sensors include anemometers, snow depth sensors and flood sensors ( Figure 4 not shown).

[0134] The communication box regularly inspects and collects the current and angle data of the tracker collected by each control box, as well as the wind speed and direction data collected by the anemometer, and the data collected by meteorological sensors such as snow depth and flood, and transmits the collected data to the SCADA server through optical fiber. The server software detects the operation anomalies of the tracker and anemometer, and counts the number of anomalies detected by the meteorological sensor. When disaster weather occurs, it triggers the entire station to enter protection mode.

[0135] There is a default polling time interval between the communication box and the control box. SCADA can issue commands to the communication box to adjust the polling interval as needed or conduct centralized polling for certain control boxes within a certain period of time. The control box runs according to its own program and is set with maximum current and limit angle alarms.

[0136] SCADA server software features include:

[0137] 1. Abnormal operation tracking of the tracker:

[0138] 1) Create groups based on the design of the tracker. Depending on the project size or engineering judgment, the entire subarray or multiple subarrays may be grouped together. Trackers in each group that have the same design or can be assigned to the same group based on engineering judgment can be grouped together. The size of the group is not limited by the size of the subarray and can cross subarrays. The definition of a group is only related to the design and load characteristics of the trackers in the group.

[0139] Each sub-array has a specific number, and the control boxes in the sub-array have corresponding numbers for system positioning. The tracker can be determined based on the number of the control box.

[0140] 2) The communication box collects the current and inclination data of the tracker and transmits these data to the SCADA server, which stores the data.

[0141] 3) The SCADA server performs anomaly detection on the collected data.

[0142] SCADA software introduces an algorithm for judging abnormal data. Abnormal data includes two categories: one is the context time series anomaly caused by a sudden failure of the same tracker; the other is that the operating data of one or several trackers are abnormal compared with other trackers in the group.

[0143] For the first anomaly: Based on the current value I and angle value α of a tracker, the derivative value is introduced This value is used as the operating data of the tracker, which can significantly represent the change in the amplitude of the current when the system resistance changes.

[0144] For the second anomaly: the operating data of each tracker in the tracker group are compared. Since the operating pace of each tracker in the group is similar but not completely consistent, the comparison frequency is to take a certain time tolerance after each data collection, such as the angle value and current value within 3 minutes (adjustable) for analysis data, and find outliers in the currently collected data.

[0145] The mean square error method can be used, that is, assuming that these data are normally distributed to calculate the statistical characteristics of the data at that moment. If a data point exceeds 3 times the standard deviation, then these points are likely to be outliers or outliers, giving the tracker a "comparison anomaly" mark.

[0146] The quartile data calculation method can also be used. The interquartile range = the difference between the third quartile and the first quartile (IQR = Q3-Q1). In this case, outliers are defined as observations that are lower than the lower limit of the reasonable interval (such as Q1-1.5xIQR) or higher than the upper limit of the reasonable interval (such as Q3+1.5xIQR). When the data of a tracker is found to be an outlier, the tracker is given an abnormal mark.

[0147] 2. Abnormal operation tracking of anemometer:

[0148] The anemometers in the photovoltaic power station are numbered for the positioning of the entire station. For example, the anemometer number can be composed of the subarray number and the anemometer number in the subarray.

[0149] The wind speed and direction data of each anemometer are transmitted to the communication box and then to the SCADA server in chronological order.

[0150] Due to the randomness of wind speed, it is not possible to perform anomaly analysis based on the time series data of a single anemometer, but to compare the data of a specific anemometer with its neighboring anemometers. The neighboring anemometers of each anemometer can be pre-set.

[0151] Considering that time acquisition cannot be completely synchronized, the average wind speed / wind direction data of a certain period of time is taken for comparison. Based on the average wind speed / wind direction data of an anemometer and its neighboring anemometers in the detection period, the mean square error method is used to determine whether the data of the anemometer is abnormal. If the data of the anemometer is abnormal in multiple consecutive detection time periods, it is determined that the anemometer is operating abnormally and relevant fault warnings are issued.

[0152] 3. Early warning and monitoring of disastrous weather:

[0153] The meteorological sensors such as anemometers, snow depth sensors, and flood sensors in the power station are numbered. The sensors and their number that have detected the disaster weather are determined based on the data measured by the meteorological sensors. If the number of sensors that have detected the disaster weather reaches a preset ratio relative to the total number of sensors of the same category in the entire station, a disaster weather warning will be issued and a protection mechanism will be activated in the entire station. The SCADA server sends the corresponding protection command to the communication box to activate the protection action of all sub-arrays, including the sub-arrays that have not detected the corresponding disaster weather.

[0154] This embodiment performs abnormality detection on the trackers and sensors in each photovoltaic subarray at the level of the entire power station, and by discovering that a certain proportion of sensors issue warnings for disastrous weather, the entire station is activated to enter a protection mode for occasional disastrous weather, thereby improving the accuracy and comprehensiveness of the abnormality detection of the entire station.

[0155] It should be noted that the above embodiments can be freely combined as needed. The above are only preferred embodiments of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A method for detecting anomalies in a photovoltaic power station control system, applied to a photovoltaic power station including a plurality of photovoltaic sub-arrays, wherein the photovoltaic power station control system includes a sensor for sensing the weather in an area where each photovoltaic sub-array is located; the method for detecting anomalies is characterized in that: include: Obtain weather perception data of each sensor and its neighboring sensors during a detection period; Determining whether the weather perception data of the sensor is abnormal based on the weather perception data of the sensor and its neighboring sensors during the detection period; If the weather sensing data of the sensor is abnormal in multiple consecutive detection time periods, it is determined that the sensor is operating abnormally and a relevant fault warning is issued; Grouping the trackers in the photovoltaic power station, grouping the trackers with the same design and installed in close geographical locations into one group; Obtain the operating data of each tracker during the detection period; Determine whether the operation of each tracker is abnormal according to the operation data of other trackers in the tracker group to which each tracker belongs during the detection period; If the tracker operates abnormally, a relevant fault warning is issued; The step of judging whether the operation of each tracker is abnormal according to the operation data of each tracker and other trackers in the tracker group to which the tracker belongs during the detection period includes: Analyze the operation data of the tracker in the detection time period to determine whether the tracker has abnormal fluctuations and the time point of occurrence of the abnormal fluctuations when the abnormal fluctuations occur; Divide the detection time period into a number of time sub-segments, traverse each time sub-segment, compare and analyze the operation data of each tracker in the tracker group to which the tracker belongs in the same time sub-segment, and determine whether there is a heterogeneous tracker in the tracker group in each time sub-segment and the number of the heterogeneous tracker if there is a heterogeneous tracker; the heterogeneous tracker is a tracker that operates significantly differently from most trackers in the tracker group; If the tracker generates data with abnormal fluctuations within the detection time period, and the tracker group to which the tracker belongs has a heterogeneous tracker in the time sub-segment to which the abnormal fluctuation occurs, and the heterogeneous tracker includes the tracker, then it is determined that the tracker is operating abnormally.

2. The abnormality detection method of a photovoltaic power station control system according to claim 1, characterized in that: Judging whether the weather perception data of the sensor is abnormal according to the weather perception data of the sensor and its neighboring sensors in the detection time period includes: Calculate the mean and standard deviation of the weather perception data according to the weather perception data of the sensor and its neighboring sensors during the detection period; Based on the mean and standard deviation, a reasonable interval is constructed according to the normal distribution rule; If the weather sensing data of the sensor falls outside the reasonable range, the weather sensing data of the sensor is abnormal.

3. The abnormality detection method of a photovoltaic power station control system according to claim 1, characterized in that: Also includes: Dividing the photovoltaic power station into several areas in advance; If the number of sensors detecting disastrous weather in an area reaches a preset ratio relative to the total number of sensors in the area within a preset time range, a disastrous weather warning will be issued in the area.

4. The abnormality detection method of a photovoltaic power station control system according to claim 1, characterized in that: Obtaining the running data of a tracker during the detection period includes: Acquiring current data and bracket angle data of a tracker during a detection period; According to the current data and the bracket angle data of the tracker at any two adjacent time points within the detection time period, the current change data required for the unit angle change of the tracker at the corresponding time point is obtained; The current change data required by the tracker for the unit angle change at each time point in the detection time period constitutes the operation data of the tracker in the detection time period.

5. The abnormality detection method of a photovoltaic power station control system according to claim 1, characterized in that: The Smoothed Z-score algorithm is used to analyze the operation data of the tracker in the detection time period to determine whether the tracker has abnormal fluctuations and the time point of occurrence of the abnormal fluctuations when the abnormal fluctuations occur.

6. The abnormality detection method of a photovoltaic power station control system according to claim 1, characterized in that: Comparing and analyzing the operation data of each tracker in the tracker group to which the tracker belongs in the same time sub-segment, determining whether there is a heterogeneous tracker in the tracker group in each time sub-segment and the number of the heterogeneous tracker when there is a heterogeneous tracker, including: Acquire the operation data of each tracker in the tracker group in the time subsegment; Calculate the mean and standard deviation of the operating data according to the operating data of each tracker in the tracker group in the time sub-segment; Based on the mean and standard deviation, a reasonable interval is constructed according to the normal distribution rule; The operation data of each tracker in the tracker group in the time subsegment is traversed. If the operation data of the tracker in the time subsegment falls outside the reasonable interval, then there is a heterogeneous tracker in the tracker group in the time subsegment, and the tracker is the heterogeneous tracker.

7. An abnormality detection device for a photovoltaic power station control system, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the abnormality detection method for a photovoltaic power station control system as described in any one of claims 1 to 6 when running the computer program.

8. A photovoltaic power station control system, characterized in that: It includes a control center and several control subsystems, each of which is used to control a photovoltaic subarray; The control subsystem includes a communication box, a sensor and several control boxes; the control subsystem communicates with the control center through the communication box; The communication box is electrically connected to the sensor and is used to collect weather information of the area where the corresponding photovoltaic sub-array is located through the sensor; The communication box is also connected to the control box by wire or wirelessly, and is used to collect current data and bracket angle data of the corresponding tracker in the photovoltaic sub-array through the control box; The control center is used to execute the abnormality detection method for a photovoltaic power station control system according to any one of claims 1-6.

Images