A complementary power station equipment monitoring method, device, equipment and storage medium

Abstract

The present application discloses a method, device, equipment and storage medium for monitoring equipment of a complementary power station, and relates to the technical field of artificial intelligence systems in the production field. The method utilizes the characteristic of Joule's law that the resistance heating is positively correlated with the square of the current. When the output of the electrical equipment increases, the current increases accordingly, which will increase the heat generated by the electrical equipment. The entire process shows a positive correlation characteristic, so when the output of the electrical equipment increases, the equipment temperature will also rise accordingly. The present application also sets the detection interval based on the temperature. The higher the temperature, the shorter the detection interval, ensuring that the abnormal time of the electrical equipment becomes shorter and shorter as the temperature increases, effectively protecting the service life of the electrical equipment. The present application collects information based on visual thermal imaging. Compared with the signal integration and manual inspection of the existing technology, the monitoring method of the present application has good robustness and has good versatility in the scope of three-party power generation equipment such as wind power generation, photovoltaic power generation, and hydropower generation.

Description

Technical Field

[0001] The present application relates to the technical field of artificial intelligence systems in the production field, and in particular to a method, apparatus, equipment and storage medium for monitoring equipment in a complementary power station. Background Art

[0002] To meet the needs of building a new power system, the construction of an integrated clean energy base featuring solar, hydro, and storage has a positive practical impact. The base uses hydropower (including pumped storage) within the basin as a regulating power source, coupling the output of hydropower and solar power to complement each other and deliver them.

[0003] In the mainstream renewable energy power generation of hydropower and photovoltaic complementarity, the real-time power of photovoltaic power is greatly affected by climate, weather, and day and night, causing the photovoltaic power generation curve to fluctuate violently, with steep rises and falls, and has obvious randomness and volatility; hydropower generation, on the other hand, has strong stability. To ensure the balance and stability of the active power of the power grid, clean energy bases need hydropower units to adjust the active power output according to the load characteristics of the power grid and the real-time power of renewable energy, smooth out the fluctuations in renewable energy load, and ensure the stability of the power grid while meeting the load requirements of the power system and the stable operation of hydropower units.

[0004] At present, the equipment monitoring of complementary power stations usually adopts the method of signal monitoring system combined with manual inspection. The signals are collected separately and then aggregated to the centralized control end for monitoring. Usually, a warning signal is generated when the electrical equipment of the complementary power station is abnormal, and the warning signal is sent to the dispatching monitoring end. When the dispatching personnel monitor the warning signal at the dispatching monitoring end, they will notify the substation personnel to inspect and handle the substation equipment. The substation personnel will eliminate the abnormality of the substation equipment to eliminate the warning signal, and conduct manual inspections regularly during non-abnormal periods to ensure stable operation of the equipment.

[0005] Due to the high randomness of photovoltaic power generation, photovoltaic power generation equipment often experiences load fluctuations, forcing the hydropower station to implement real-time adjustments to smooth out these load fluctuations. Electrical equipment on both the photovoltaic and hydropower stations often operates under fluctuating or unstable loads, which can easily lead to false alarms in the signal monitoring system. For example, non-abnormal conditions may be incorrectly reported, while abnormal conditions may not be reported at all, making the signal monitoring system inaccurate. Summary of the Invention

[0006] The main purpose of the present application is to provide a device monitoring method, apparatus, equipment and storage medium for a complementary power station, so as to solve the problem of inaccurate signal monitoring systems of complementary power stations in the prior art.

[0007] In order to achieve the above objectives, this application provides the following technical solutions:

[0008] A method for monitoring equipment in a complementary power station, the complementary power station comprising a hydropower terminal and a photovoltaic terminal connected to each other, wherein all electrical equipment in the hydropower terminal and the photovoltaic terminal are monitored by at least one external camera terminal, the method comprising:

[0009] Step S1, respectively obtaining the minimum historical output and the maximum historical output of the hydropower terminal and the photovoltaic terminal;

[0010] Step S2, taking the minimum of the two minimum historical outputs as the minimum value of the output interval, and taking the maximum of the two maximum historical outputs as the maximum value of the output interval, to construct the output interval;

[0011] Step S3, dividing the output interval into at least two consecutive sub-intervals;

[0012] Step S4, defining a time period based on each sub-interval, wherein the length of all time periods increases from long to short as the value of the output interval increases from small to large;

[0013] Step S5, respectively obtaining the real-time output of the hydropower terminal and the photovoltaic terminal;

[0014] Step S6, classifying all real-time outputs into all sub-intervals through a classification algorithm, and obtaining a time period based on a real-time output;

[0015] Step S7, collecting thermal images of all electrical devices through the external shooting terminal at intervals of the time period;

[0016] Step S8, determining whether the temperature of each electrical device exceeds a preset temperature threshold based on all thermal images;

[0017] Step S9: Mark the electrical equipment that exceeds the preset temperature threshold as an abnormally over-temperature device.

[0018] As a further improvement of the present application, step S6 classifies all real-time outputs into all subintervals using a classification algorithm, and obtains a time period based on a real-time output, including:

[0019] Step S61: define a set to be classified A based on all real-time outputs = {a1, a2, ..., a j ,…,a m}, where a j is the jth real-time output in the set A to be classified, and m is the number of all real-time outputs;

[0020] Step S62: define a category set C = {c1, c2, ..., c i ,…,c n}, where ci is the i-th time period in the category set C;

[0021] Step S63, calculating the conditional probability of the signal set A to be classified in each time period according to formula (1):

[0022]

[0023] Among them, P(Ac i ) is the conditional probability of the signal set A to be classified in the i-th time period, P(c i ) is the marginal probability of the i-th time period, P(a j c i ) is the conditional probability of the jth real-time output in the i-th time period;

[0024] Step S64, classifying each real-time output into the time period with the highest conditional probability;

[0025] Step S65 , respectively obtain the time period of each real-time output, and obtain a time period based on one real-time output.

[0026] As a further improvement of the present application, in step S7, thermal images of all electrical devices are collected by the external camera at intervals of the time period. Then, in step S8, based on all thermal images, it is determined whether the temperature of each electrical device exceeds a preset temperature threshold. The above steps include:

[0027] Step S10, packing all pixel color values of a thermal image into a color value data set;

[0028] Step S20, normalizing all color value data sets to obtain normalized data sets;

[0029] Step S30, dividing the normalized data set into a training set and a validation set according to a preset ratio;

[0030] Step S40, defining a neural network model in which the input layer, the hidden layer, and the output layer are sequentially connected;

[0031] Step S50: outputting the training set to the input layer, performing several trainings through the neural network model, and obtaining a root mean square error between the validation set and the current training result based on each training;

[0032] Step S60, obtaining a minimum value of the root mean square error among all training results, and obtaining a training result corresponding to the minimum value as a prediction model;

[0033] Step S70, outputting the verification set to the prediction model to obtain a predicted color value data set;

[0034] Step S80: Restore the predicted color value data set to a predicted thermal image and substitute it into step S8.

[0035] As a further improvement of the present application, step S20 normalizes all color value data sets to obtain normalized data sets, including:

[0036] According to formula (2), all color values of the current color value dataset are normalized:

[0037]

[0038] Among them, θ R ' GB is the normalized color value obtained after normalization, θ RGB is the color value of the current color value dataset, is the average value of all color values in the current color value dataset, and δ is the standard deviation of all color values in the current color value dataset.

[0039] As a further improvement of the present application, the neural network model is represented by formula (3):

[0040]

[0041] Wherein, y is the neural network model; x q is the qth input node of the input layer, each input node corresponds to a normalized data set in the training set, is the weight from the p-th input node of the input layer to the q-th input node of the hidden layer; is the bias of the qth input node connected to the hidden layer; is the bias of the output layer; tansig(x) is the transfer function, and The numbers in the brackets of the symbols are the number of layers. The superscript (1) is the first layer, that is, the input layer, and the superscript (1, 2) is the first layer to the second layer, that is, the input layer to the hidden layer.

[0042] As a further improvement of the present application, the root mean square error is represented by formula (4):

[0043]

[0044] Wherein, RMSE is the root mean square error, q is the number of the input nodes or the normalized data set, and x p is the color value of the pth normalized data set, This is the training result after the color value training of the p-th normalized dataset is completed.

[0045] As a further improvement of the present application, step S9 marks the electrical equipment exceeding the preset temperature threshold as an abnormally over-temperature device, and then includes:

[0046] Step S100, obtaining the location information of the abnormally overheated device and the color value of the thermal imaging corresponding to the abnormally overheated device;

[0047] Step S200: sending the location information and the color value of the thermal imaging corresponding to the abnormally overheated device to an external receiving end.

[0048] In order to achieve the above objectives, this application also provides the following technical solutions:

[0049] An equipment monitoring device for a complementary power station, the equipment monitoring device being applied to the above-mentioned equipment monitoring method, the equipment monitoring device comprising:

[0050] A historical output maximum value acquisition module is used to respectively obtain the minimum historical output and the maximum historical output of the hydropower terminal and the photovoltaic terminal;

[0051] an output interval construction module, configured to construct the output interval by taking the minimum of the two minimum historical outputs as the minimum value of the output interval and the maximum of the two maximum historical outputs as the maximum value of the output interval;

[0052] An output sub-interval division module, configured to evenly divide the output interval into at least two consecutive sub-intervals;

[0053] A time period definition module is used to define a time period based on each sub-interval, and the length of all time periods increases from small to large as the value of the output interval increases;

[0054] A real-time output acquisition module, used to respectively acquire the real-time outputs of the hydropower terminal and the photovoltaic terminal;

[0055] A real-time output classification module is used to classify all real-time outputs into all sub-intervals through a classification algorithm, and obtain a time period based on a real-time output;

[0056] A thermal imaging acquisition module, configured to acquire thermal images of all electrical devices via the external shooting terminal at intervals of the time period;

[0057] A thermal imaging judgment module, configured to judge whether the temperature of each electrical device exceeds a preset temperature threshold based on all thermal images;

[0058] The abnormally over-temperature device marking module is used to mark electrical equipment that exceeds the preset temperature threshold as an abnormally over-temperature device.

[0059] In order to achieve the above objectives, this application also provides the following technical solutions:

[0060] An electronic device includes a processor and a memory coupled to the processor, wherein the memory stores program instructions that can be executed by the processor; when the processor executes the program instructions stored in the memory, the device monitoring method of the complementary power station as described above is implemented.

[0061] To achieve the above objectives, this application also provides the following technical solutions:

[0062] A storage medium stores program instructions, which, when executed by a processor, implement the above-mentioned device monitoring method for a complementary power station.

[0063] The present application obtains the minimum historical output and maximum historical output of the hydropower end and the photovoltaic end respectively; takes the minimum value of the two minimum historical outputs as the minimum value of the output interval and the maximum value of the two maximum historical outputs as the maximum value of the output interval to construct an output interval; divides the output interval into at least two continuous sub-intervals on average; defines a time period based on each sub-interval, and the length of all time periods increases from long to short as the value of the output interval increases from small to large; obtains the real-time output of the hydropower end and the photovoltaic end respectively; classifies all real-time outputs into all sub-intervals through a classification algorithm, and obtains a time period based on one real-time output; collects thermal images of all electrical equipment through an external shooting end during the interval time period; determines whether the temperature of each electrical equipment exceeds a preset temperature threshold based on all thermal images; and marks electrical equipment that exceeds the preset temperature threshold as abnormally overheated equipment. This application utilizes the characteristic of Joule's law that the resistance heating is positively correlated with the square of the current. When the output of the electrical equipment increases, the current increases accordingly, which will increase the heat generated by the electrical equipment. The entire process shows a positive correlation characteristic, so when the output of the electrical equipment increases, the equipment temperature will also rise accordingly. This application also sets the detection interval based on the temperature. The higher the temperature, the shorter the detection interval, ensuring that the abnormal time of the electrical equipment becomes shorter and shorter as the temperature increases, effectively protecting the service life of the electrical equipment. This application collects information based on visual thermal imaging. Compared with the signal integration and manual inspection of the existing technology, the monitoring method of this application has good robustness and good versatility within the scope of wind, solar and hydropower generation equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 A schematic diagram of the process steps of an embodiment of the device monitoring method of the complementary power station of the present application;

[0065] Figure 2 This is a functional module diagram of an embodiment of the equipment monitoring device of the complementary power station of the present application;

[0066] Figure 3 This is a schematic structural diagram of an embodiment of the electronic device of the present application;

[0067] Figure 4 This is a structural diagram of an embodiment of the storage medium of the present application. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0069] The terms "first", "second" and "third" in this application are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, a feature defined as "first", "second" and "third" may explicitly or implicitly include at least one of such features. In the description of this application, "multiple" means at least two, for example, two, three, etc., unless otherwise clearly and specifically defined. All directional indications in the embodiments of this application (such as up, down, left, right, front, back...) are only used to explain the relative positional relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally also include steps or units that are not listed, or may optionally also include other steps or units inherent to these processes, methods, products or devices.

[0070] References to "embodiments" herein mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0071] like Figure 1As shown, this embodiment provides an embodiment of a method for monitoring equipment in a complementary power station. In this embodiment, the complementary power station includes a hydropower terminal and a photovoltaic terminal that are connected to the grid. All electrical equipment at the hydropower terminal and the photovoltaic terminal are monitored through at least one external shooting terminal.

[0072] Preferably, this embodiment can also be applied to a wind, solar and water three-terminal complementary integrated power station.

[0073] Specifically, the device monitoring method includes the following steps:

[0074] Step S1, respectively obtaining the minimum historical output and the maximum historical output of the hydropower terminal and the photovoltaic terminal.

[0075] Preferably, output can be understood as power generation. The power generation of a photovoltaic power station is calculated based on its installed capacity and lighting conditions. Generally speaking, the annual power generation of a photovoltaic power station can be calculated by multiplying the installed capacity by the equivalent full-time hours. For example, a one-megawatt photovoltaic power station refers to an installed capacity of 1 megawatt, or 1,000 kilowatts. The operation of a photovoltaic power station also involves reactive voltage control, including the configuration of reactive power sources and the satisfaction of basic voltage control requirements to ensure the effective operation of the power station and the stability of the power system. Specifically in terms of output range, taking a photovoltaic power station in a certain area as an example, the maximum photovoltaic output in spring (February-May) exceeds that in summer (June-September). For example, the maximum photovoltaic output from February to May 2022 was 34,740 kW, 31,590 kW, 30,990 kW, and 35,270 kW, respectively, while the maximum output from June to September of the same year was 33,630 kW, 32,970 kW, 26,190 kW, and 25,020 kW, respectively. This shows that although there is plenty of sunshine in the summer, due to other factors (such as equipment maintenance, shadow blocking, etc.), the photovoltaic output capacity in spring exceeds that in summer. In addition, the output power curves of domestic distributed photovoltaic power stations show that the photovoltaic output varies in different seasons and months. For example, the typical daily output power curve of an 800kWp rooftop distributed photovoltaic power station in a certain area of Guangdong shows that the proportion of power generation in each month is different, which reflects that the photovoltaic output is affected by seasonal changes.

[0076] Preferably, the output of a hydropower station is relatively stable. The output of a non-regulated and daily-regulated hydropower station is usually calculated based on the average daily output of the designed dry days. Since the site selection of a hydropower station is based on a stable water flow, the output range of the hydropower station is usually little affected by the change in daily flow. The output of a seasonally regulated hydropower station and an annually regulated hydropower station involves the average output after reservoir regulation. The output of a seasonally regulated hydropower station refers to the average output of the water supply period in the designed dry year after reservoir regulation, while the output of an annually regulated hydropower station is based on runoff regulation calculations and hydropower calculations based on a long series of hydrological data. The average output of each water supply period is calculated, and a curve of the average output guarantee rate of the water supply period is drawn. The average output value on the curve corresponding to the design guarantee rate (Pd) of the hydropower station is the guaranteed output value of the annually regulated hydropower station. The output calculation of a multi-year regulated hydropower station is similar to that of a one-year regulated hydropower station, but the flow and head in the formula refer to the average flow and average head of the multi-year series of designed dry years. This type of hydropower station has a wider output range because it can better utilize water resources, especially during the dry season.

[0077] For example, a hydropower station on a certain river is equipped with four mixed-flow turbine generator sets with a single unit capacity of 50,000 kilowatts, with a total installed capacity of 200,000 kilowatts. That is, when operating at full load, the output of the power station can reach 200,000 kilowatts. When there is no influence of head and scheduling policy, the output of the hydropower station generally will not fluctuate significantly.

[0078] Step S2: The minimum value of the two minimum historical outputs is used as the minimum value of the output interval, and the maximum value of the two maximum historical outputs is used as the maximum value of the output interval to construct the output interval.

[0079] Step S3: Divide the output interval into at least two consecutive sub-intervals on average.

[0080] In step S4, a time period is defined based on each sub-interval, and the lengths of all time periods are shortened as the value of the output interval increases from small to large.

[0081] For example, the output intervals are [0.1MW, 0.5MW), [0.5MW, 1.0MW), [1.0MW, 1.5MW), [1.5MW, 2.0MW), [2.0MW, 2.5MW], and the time periods can be defined as 2.5h, 2.0h, 1.5h, 1.0h, and 0.5h respectively.

[0082] It should be noted that the above output range examples do not represent the actual situation, and the output range should be set according to the actual statistical data of the power station.

[0083] Step S5: obtaining the real-time output of the hydropower terminal and the photovoltaic terminal respectively.

[0084] Step S6: Classify all real-time outputs into all sub-intervals through a classification algorithm, and obtain a time period based on one real-time output.

[0085] Step S7: collecting thermal images of all electrical devices through an external camera at intervals.

[0086] Thermal imaging is based on the principle that all objects emit infrared radiation, the intensity of which is proportional to the object's temperature. Thermal imaging technology detects this infrared radiation, converts it into electrical signals, and ultimately generates a thermal image on a display, showing the object's temperature distribution. The thermal imaging process involves several steps: First, the thermal imager receives the infrared radiation energy distribution pattern of the target through an infrared detector and an optical imaging objective lens. This energy is then reflected onto the infrared detector's photosensitive element, resulting in an infrared thermal image. Finally, this image corresponds to the heat distribution field on the object's surface, with different colors representing different temperatures. Temperatures represented by different colors in thermal imaging: In thermal images, different colors represent different temperatures. Typically, red indicates high temperatures, blue indicates low temperatures, and black indicates the lowest temperature areas.

[0087] Step S8: determining whether the temperature of each electrical device exceeds a preset temperature threshold based on all thermal images.

[0088] Preferably, the commonly used thermal imaging color red represents high temperature, and blue represents low temperature, and the temperature can be judged by the color value.

[0089] Preferably, the normal operating temperature range of the electrical equipment of the complementary power station is generally -40°C to 85°C, and the recommended operating temperature range is -20°C to 50°C. When the temperature exceeds 85°C, it can be considered that the temperature of the electrical equipment is abnormal.

[0090] Step S9: Mark the electrical equipment that exceeds the preset temperature threshold as an abnormally over-temperature device.

[0091] Furthermore, step S6 classifies all real-time outputs into all sub-intervals using a classification algorithm, and obtains a time period based on a real-time output, which specifically includes the following steps:

[0092] Step S61: define a set to be classified A based on all real-time outputs = {a1, a2, ..., a j ,…,a m}, where a j is the jth real-time output in the set A to be classified, and m is the number of all real-time outputs.

[0093] Step S62: define a category set C = {c1, c2, ..., c i ,…,c n}, where c i is the i-th time period in the category set C.

[0094] Step S63, calculate the conditional probability of the signal set A to be classified in each time period according to formula (1):

[0095]

[0096] Among them, P(Ac i ) is the conditional probability of the signal set A to be classified in the i-th time period, P(c i ) is the marginal probability of the i-th time period, P(a j c i ) is the conditional probability of the jth real-time output in the i-th time period.

[0097] Step S64 , classifying each real-time output into the time period with the highest conditional probability.

[0098] Step S65 , respectively obtain the time period of each real-time output, and obtain a time period based on one real-time output.

[0099] Furthermore, in step S7, thermal images of all electrical devices are collected through an external camera at intervals. Then, in step S8, it is determined based on all thermal images whether the temperature of each electrical device exceeds a preset temperature threshold. The above steps include:

[0100] Step S10: Pack all pixel color values of a thermal image into a color value data set.

[0101] Preferably, this embodiment uses RGB color values.

[0102] Step S20 , normalizing all color value data sets to obtain normalized data sets.

[0103] Preferably, this embodiment prefers the normalization method of zero-mean normalization (Z-score normalization), which gives the mean (mean) and standard deviation (standard deviation) of the original data to standardize the data, and the processed data conforms to the standard normal distribution, that is, the mean is 0 and the standard deviation is 1. For the normalization method, this embodiment can also use batch normalization. Compared with simple normalization, in previous neural network training, only the input layer data is normalized, but no normalization is performed in the middle layer. Although the data set of the input node is normalized, the data distribution of the input data after matrix multiplication is likely to change greatly, and as the number of hidden layer network layers continues to deepen, the change in data distribution will become greater and greater. Therefore, the normalization performed by batch normalization in the middle layer of the neural network makes the training effect better.

[0104] Step S30: Divide the normalized data set into a training set and a validation set according to a preset ratio.

[0105] Preferably, the preset ratio is 8:2, so as to divide the normalized data set into a training set and a sample set in a ratio of 8:2.

[0106] Preferably, in actual application, a certain proportion of validation sets is also required to verify the accuracy of the model, that is, verification is required after training is completed, and the sample set prediction is started only after successful verification. Usually, the image data is divided into training set, validation set, and sample set in a ratio of 70%:15%:15%, that is, 70% of the data is the training set, 15% of the data is the validation set, and 15% of the data is the sample set.

[0107] Step S40, defining a neural network model in which the input layer, the hidden layer, and the output layer are sequentially connected.

[0108] In step S50, the training set is output to the input layer, and the neural network model is used for several trainings. The root mean square error between the validation set and the current training result is obtained based on each training.

[0109] Step S60: Obtain the minimum value of the root mean square error among all training results, and obtain the training result corresponding to the minimum value as the prediction model.

[0110] Step S70: Output the validation set to the prediction model to obtain a predicted color value data set.

[0111] Step S80 , restoring the predicted color value dataset to the predicted thermal image and substituting it into step S8 .

[0112] Furthermore, in step S20, all color value data sets are normalized to obtain normalized data sets, including:

[0113] According to formula (2), all color values of the current color value dataset are normalized:

[0114]

[0115] Among them, θ R ' GB is the normalized color value obtained after normalization, θ RGB is the color value of the current color value dataset, is the average value of all color values in the current color value dataset, and δ is the standard deviation of all color values in the current color value dataset.

[0116] Furthermore, the neural network model is represented by formula (3):

[0117]

[0118] Among them, y is the neural network model; x q is the qth input node of the input layer, each input node corresponds to a normalized data set in the training set, is the weight from the p-th input node of the input layer to the q-th input node of the hidden layer; is the bias of the qth input node connected to the hidden layer; is the bias of the output layer; tansig(x) is the transfer function, and The numbers in the brackets of the symbols in formula (3) are the number of layers. The superscript (1) is the first layer, that is, the input layer, and the superscript (1, 2) is the first layer to the second layer, that is, the input layer to the hidden layer.

[0119] Preferably, the training model training a neural network usually requires providing a large amount of data, namely a data set; the data set is generally divided into three categories, namely the above-mentioned training set (training set), validation set (valid data set) and test set (test set).

[0120] Among them, one epoch is the process of training once with all the samples in the training set. The so-called training once refers to one forward pass and one back pass. When the number of samples in an epoch (i.e., training set) is too large, training once may consume too much time, and it is not necessary to use all the data in the training set for each training. In this case, the entire training set needs to be divided into multiple small blocks, that is, divided into multiple batches for training. An epoch consists of one or more batches. A batch is a part of the training set. Each training process only uses a part of the data, i.e., a batch. The process of training a batch is an iteration.

[0121] Preferably, the neural network training specifically includes a perceptron, which is composed of two layers of neurons. The input layer receives external input signals and transmits them to the output layer. The output layer is MP neurons, and the step function is y=f(∑w·x-β).

[0122] Preferably, given a training data set, the weight w and the training bias β can be obtained through learning, and β can be understood as a fixed value with an input of -1,0.

[0123] It should be noted that the symbolic meaning of the above formula is not interchangeable with other formulas in the embodiment. This step function is only used to illustrate the principle and does not participate in the calculation of other formulas.

[0124] Preferably, in this embodiment, the number of neural network training times can be set to 1000 times.

[0125] Preferably, the learning rate from the 1st to the 500th epoch can be set to 0.01, the learning rate from the 501st to the 750th epoch can be set to 0.001, and the learning rate from the 751st to the 1000th epoch can be set to 0.0001.

[0126] It can be understood that the neural network training of this embodiment mainly includes the following ideas:

[0127] ① Initialize the weights and bias items in the network.

[0128] Initializing the parameter values (output unit weights, bias terms and hidden unit weights, bias terms are all model parameters) is to activate forward propagation, obtain the output value of each layer element, and then obtain the value of the loss function.

[0129] ②Activate forward propagation to obtain the output value of each layer and the expected value of the loss function of each layer.

[0130] ③ According to the loss function, calculate the error term of the output unit and the error term of the hidden unit.

[0131] Compute various errors, calculate the gradient of the parameters with respect to the loss function, or calculate partial derivatives using the chain rule of calculus. For partial derivatives of vectors or matrices within a composite function, always multiply the composite function's inner functions with the left derivative. For partial derivatives of scalars within a composite function, either multiply the composite function's inner functions with the right derivative can be used.

[0132] ④Update the weights and bias items in the neural network.

[0133] ⑤ Repeat ② to ④ until the loss function is less than the preset bias or the number of iterations is exhausted, and the parameters output at this time are the current optimal parameters.

[0134] Furthermore, the root mean square error is represented by formula (4):

[0135]

[0136] Among them, RMSE is the root mean square error, q is the number of input nodes or normalized data sets, and x p is the color value of the pth normalized data set, This is the training result after the color value training of the p-th normalized dataset is completed.

[0137] Furthermore, in step S9, the electrical equipment exceeding the preset temperature threshold is marked as abnormally over-temperature equipment, and then the following steps are included:

[0138] Step S100: Acquire location information of an abnormally overheated device and a color value of a thermal image corresponding to the abnormally overheated device.

[0139] Step S200: sending the location information and the color value of the thermal imaging corresponding to the abnormally overheated device to an external receiving end.

[0140] This embodiment obtains the minimum historical output and maximum historical output of the hydropower end and the photovoltaic end respectively; takes the minimum value of the two minimum historical outputs as the minimum value of the output interval, and takes the maximum value of the two maximum historical outputs as the maximum value of the output interval to construct an output interval; divides the output interval into at least two continuous sub-intervals on average; defines a time period based on each sub-interval, and the length of all time periods increases from long to short as the value of the output interval increases from small to large; obtains the real-time output of the hydropower end and the photovoltaic end respectively; classifies all real-time outputs into all sub-intervals through a classification algorithm, and obtains a time period based on one real-time output; collects thermal images of all electrical equipment through an external shooting end during the interval time period; determines whether the temperature of each electrical equipment exceeds a preset temperature threshold based on all thermal images; and marks electrical equipment that exceeds the preset temperature threshold as an abnormally overheated device. This embodiment utilizes the characteristic of Joule's law that the resistance heating is positively correlated with the square of the current. When the output of the electrical equipment increases, the current increases accordingly, which will increase the heat generated by the electrical equipment. The entire process shows a positive correlation characteristic, so when the output of the electrical equipment increases, the equipment temperature will also rise accordingly. This embodiment also sets the detection interval based on the temperature. The higher the temperature, the shorter the detection interval, ensuring that the abnormal time of the electrical equipment becomes shorter and shorter as the temperature increases, effectively protecting the service life of the electrical equipment. This embodiment collects information based on visual thermal imaging. Compared with the signal integration and manual inspection of the existing technology, the monitoring method of this embodiment has good robustness and good versatility within the scope of wind, solar and hydropower generation equipment.

[0141] like Figure 2 As shown, this embodiment provides an embodiment of an equipment monitoring device for a complementary power station. In this embodiment, the equipment monitoring device is applied to the equipment monitoring method in the above embodiment.

[0142] Specifically, the equipment monitoring device includes a historical output maximum value acquisition module 1, an output interval construction module 2, an output sub-interval division module 3, a time period definition module 4, a real-time output acquisition module 5, a real-time output classification module 6, a thermal imaging acquisition module 7, a thermal imaging judgment module 8, and an abnormal over-temperature equipment marking module 9, which are electrically connected in sequence.

[0143] Among them, the historical output maximum value acquisition module 1 is used to obtain the minimum historical output and maximum historical output of the hydropower end and the photovoltaic end respectively; the output interval construction module 2 is used to take the minimum value of the two minimum historical outputs as the minimum value of the output interval and the maximum value of the two maximum historical outputs as the maximum value of the output interval to construct the output interval; the output sub-interval division module 3 is used to divide the output interval into at least two consecutive sub-intervals on average; the time period definition module 4 is used to define a time period based on each sub-interval, and the length of all time periods increases from small to large as the value of the output interval increases. to short; the real-time output acquisition module 5 is used to obtain the real-time output of the hydropower end and the photovoltaic end respectively; the real-time output classification module 6 is used to classify all real-time outputs into all sub-intervals through a classification algorithm, and obtain a time period based on a real-time output; the thermal imaging acquisition module 7 is used to collect thermal images of all electrical equipment through an external shooting end at intervals; the thermal imaging judgment module 8 is used to judge whether the temperature of each electrical equipment exceeds the preset temperature threshold based on all thermal images; the abnormal over-temperature equipment marking module 9 is used to mark electrical equipment that exceeds the preset temperature threshold as abnormal over-temperature equipment.

[0144] Furthermore, the real-time output classification module 6 specifically includes a first real-time output classification submodule, a second real-time output classification submodule, a third real-time output classification submodule, a fourth real-time output classification submodule, and a fifth real-time output classification submodule, which are electrically connected in sequence; the first real-time output classification submodule is electrically connected to the real-time output acquisition module 5, and the fifth real-time output classification submodule is electrically connected to the thermal imaging acquisition module 7.

[0145] The first real-time output classification submodule is used to define a set to be classified A = {a1, a2, ..., a j ,…,a m}, where a j is the jth real-time output in the set A to be classified, and m is the number of all real-time outputs.

[0146] The second real-time output classification submodule is used to define a category set C = {c1, c2, ..., c i ,…,c n}, where c i is the i-th time period in the category set C.

[0147] The third real-time output classification submodule is used to calculate the conditional probability of the signal set A to be classified in each time period according to formula (1):

[0148]

[0149] Among them, P(Ac i) is the conditional probability of the signal set A to be classified in the i-th time period, P(c i ) is the marginal probability of the i-th time period, P(a j c i ) is the conditional probability of the jth real-time output in the i-th time period.

[0150] The fourth real-time output classification submodule is used to classify each real-time output into a time period with the highest conditional probability.

[0151] The fifth real-time output classification submodule is used to respectively obtain the time period of each real-time output, and obtain a time period based on one real-time output.

[0152] Furthermore, the thermal imaging acquisition module 7 and the thermal imaging judgment module 8 also include a pixel color value packaging module, a color value data set normalization processing module, a normalized data set partitioning module, a neural network model definition module, a neural network model training module, a prediction model acquisition module, a prediction color value data set acquisition module, and a prediction thermal imaging restoration and substitution module, which are electrically connected in sequence; the pixel color value packaging module is electrically connected to the thermal imaging acquisition module 7, and the prediction thermal imaging restoration and substitution module is electrically connected to the thermal imaging judgment module 8.

[0153] Among them, the pixel color value packaging module is used to package all pixel color values of a thermal image into a color value data set; the color value data set normalization processing module is used to normalize all color value data sets to obtain a normalized data set; the normalized data set division module is used to divide the normalized data set into a training set and a verification set according to a preset ratio; the neural network model definition module is used to define a neural network model in which the input layer, hidden layer, and output layer are signal-connected in sequence; the neural network model training module is used to output the training set to the input layer, perform several trainings through the neural network model, and obtain the root mean square error between the verification set and the current training result based on each training; the prediction model acquisition module is used to obtain the minimum value of the root mean square error in all training results, and obtain the training result corresponding to the minimum value as the prediction model; the predicted color value data set acquisition module is used to output the verification set to the prediction model to obtain a predicted color value data set; the predicted thermal imaging restoration and substitution module is used to restore the predicted color value data set to the predicted thermal imaging and substitute it into the thermal imaging judgment module 8.

[0154] Furthermore, the color value dataset normalization processing module is specifically used to perform standard normalization on all color values of the current color value dataset according to formula (2):

[0155]

[0156] Among them, θ R 'GB is the normalized color value obtained after normalization, θ RGB is the color value of the current color value dataset, is the average value of all color values in the current color value dataset, and δ is the standard deviation of all color values in the current color value dataset.

[0157] Furthermore, the neural network model definition module is equipped with a neural network model represented by formula (3):

[0158]

[0159] Among them, y is the neural network model; x q is the qth input node of the input layer, each input node corresponds to a normalized data set in the training set, is the weight from the p-th input node of the input layer to the q-th input node of the hidden layer; is the bias of the qth input node connected to the hidden layer; is the bias of the output layer; tansig(x) is the transfer function, and The numbers in the brackets of the symbols are the number of layers. The superscript (1) is the first layer, that is, the input layer, and the superscript (1, 2) is the first layer to the second layer, that is, the input layer to the hidden layer.

[0160] Furthermore, the prediction model acquisition module is equipped with the root mean square error represented by formula (4):

[0161]

[0162] Among them, RMSE is the root mean square error, q is the number of input nodes or normalized data sets, and x p is the color value of the pth normalized data set, This is the training result after the color value training of the p-th normalized dataset is completed.

[0163] Furthermore, the equipment monitoring device also includes an abnormal over-temperature equipment information acquisition module and a normal over-temperature equipment information sending module electrically connected in sequence; the abnormal over-temperature equipment information acquisition module is electrically connected to the abnormal over-temperature equipment marking module 9.

[0164] Among them, the abnormal overheating device information acquisition module is used to obtain the location information of the abnormal overheating device and the color value of the thermal imaging corresponding to the abnormal overheating device; the abnormal overheating device information sending module is used to send the location information and the color value of the thermal imaging corresponding to the abnormal overheating device to the external receiving end.

[0165] It should be noted that this embodiment is a functional module embodiment based on the above method embodiment. The preferred, expanded, limited and exemplary parts of this embodiment can be referred to the above embodiment, and will not be described in detail in this embodiment.

[0166] This embodiment obtains the minimum historical output and maximum historical output of the hydropower end and the photovoltaic end respectively; takes the minimum value of the two minimum historical outputs as the minimum value of the output interval, and takes the maximum value of the two maximum historical outputs as the maximum value of the output interval to construct an output interval; divides the output interval into at least two continuous sub-intervals on average; defines a time period based on each sub-interval, and the length of all time periods increases from long to short as the value of the output interval increases from small to large; obtains the real-time output of the hydropower end and the photovoltaic end respectively; classifies all real-time outputs into all sub-intervals through a classification algorithm, and obtains a time period based on one real-time output; collects thermal images of all electrical equipment through an external shooting end during the interval time period; determines whether the temperature of each electrical equipment exceeds a preset temperature threshold based on all thermal images; and marks electrical equipment that exceeds the preset temperature threshold as an abnormally overheated device. This embodiment utilizes the characteristic of Joule's law that the resistance heating is positively correlated with the square of the current. When the output of the electrical equipment increases, the current increases accordingly, which will increase the heat generated by the electrical equipment. The entire process shows a positive correlation characteristic, so when the output of the electrical equipment increases, the equipment temperature will also rise accordingly. This embodiment also sets the detection interval based on the temperature. The higher the temperature, the shorter the detection interval, ensuring that the abnormal time of the electrical equipment becomes shorter and shorter as the temperature increases, effectively protecting the service life of the electrical equipment. This embodiment collects information based on visual thermal imaging. Compared with the signal integration and manual inspection of the existing technology, the monitoring method of this embodiment has good robustness and good versatility within the scope of wind, solar and hydropower generation equipment.

[0167] Figure 3 An embodiment of the electronic device of the present application is shown, see Figure 3 The electronic device 10 includes a processor 101 and a memory 102 coupled to the processor 101.

[0168] The memory 102 stores program instructions for implementing the device monitoring method of the complementary power plant according to any one of the above embodiments.

[0169] The processor 101 is configured to execute program instructions stored in the memory 102 to perform equipment monitoring of the complementary power station.

[0170] The processor 101 may also be referred to as a CPU (Central Processing Unit). The processor 101 may be an integrated circuit chip having signal processing capabilities. The processor 101 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. The general-purpose processor may be a microprocessor or any conventional processor.

[0171] Further, Figure 4 This is a schematic diagram of the structure of the storage medium of an embodiment of the present application, see Figure 4 The storage medium 11 of the embodiment of the present application stores program instructions 111 that can implement all the above methods, wherein the program instructions 111 can be stored in the above storage medium in the form of a software product, including a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0172] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0173] In addition, the functional units in the various embodiments of the present application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units. The above is only an implementation method of the present application and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the description and drawings of this application, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.

[0174] The above detailed description of the specific embodiments of the invention is intended only as an example, and the present application is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions of the invention are also within the scope of the present application. Therefore, equivalent changes, modifications, and improvements made without departing from the spirit and scope of the present application should be included within the scope of the present application.

Claims

1. A method for monitoring equipment in a complementary power station, wherein the complementary power station comprises a hydropower terminal and a photovoltaic terminal connected to the grid, wherein all electrical equipment in the hydropower terminal and the photovoltaic terminal are monitored by at least one external camera terminal, characterized in that: The equipment monitoring method of the complementary power station includes: Step S1, respectively obtaining the minimum historical output and the maximum historical output of the hydropower terminal and the photovoltaic terminal; Step S2, taking the minimum of the two minimum historical outputs as the minimum value of the output interval, and taking the maximum of the two maximum historical outputs as the maximum value of the output interval, to construct the output interval; Step S3, dividing the output interval into at least two consecutive sub-intervals; Step S4, defining a time period based on each sub-interval, wherein the length of all time periods increases from long to short as the value of the output interval increases from small to large; Step S5, respectively obtaining the real-time output of the hydropower terminal and the photovoltaic terminal; Step S6, classifying all real-time outputs into all sub-intervals through a classification algorithm, and obtaining a time period based on a real-time output; Step S7, collecting thermal images of all electrical devices through the external shooting terminal at intervals of the time period; Step S8, determining whether the temperature of each electrical device exceeds a preset temperature threshold based on all thermal images; Step S9, marking the electrical equipment that exceeds the preset temperature threshold as an abnormally over-temperature device; Step S6, classifying all real-time outputs into all sub-intervals through a classification algorithm, and obtaining a time period based on a real-time output, including: Step S61: define a set to be classified based on all real-time outputs ,in The set to be classified The Real-time output, The number of all real-time outputs; Step S62: Define a category set based on all time periods ,in, For the category set The time period; Step S63: Calculate the signal set to be classified according to formula (1): The conditional probability in each time period is: （1）； in, For the The signal set to be classified in a time period The conditional probability of For the The marginal probability of a time period, For the Under the time period The conditional probability of real-time output; Step S64, classifying each real-time output into the time period with the highest conditional probability; Step S65 , respectively obtain the time period of each real-time output, and obtain a time period based on one real-time output.

2. The device monitoring method according to claim 1, characterized in that: Step S7, collecting thermal images of all electrical devices through the external camera at intervals of the time period, and then, step S8, determining whether the temperature of each electrical device exceeds a preset temperature threshold based on all thermal images, before which, the steps include: Step S10, packing all pixel color values of a thermal image into a color value data set; Step S20, normalizing all color value data sets to obtain normalized data sets; Step S30, dividing the normalized data set into a training set and a validation set according to a preset ratio; Step S40, defining a neural network model in which the input layer, the hidden layer, and the output layer are sequentially connected; Step S50: outputting the training set to the input layer, performing several trainings through the neural network model, and obtaining a root mean square error between the validation set and the current training result based on each training; Step S60, obtaining a minimum value of the root mean square error among all training results, and obtaining a training result corresponding to the minimum value as a prediction model; Step S70, outputting the verification set to the prediction model to obtain a predicted color value data set; Step S80: Restore the predicted color value data set to a predicted thermal image and substitute it into step S8.

3. The device monitoring method according to claim 2, characterized in that: Step S20, normalizing all color value data sets to obtain normalized data sets, including: According to formula (2), all color values of the current color value dataset are normalized: （2）； in, is the normalized color value obtained after normalization. is the color value of the current color value dataset, is the average value of all color values in the current color value dataset, The standard deviation of all color values in the current color value dataset.

4. The device monitoring method according to claim 2, characterized in that: The neural network model is represented by formula (3): （3）； in, is the neural network model; is the first input nodes, each input node corresponds to a normalized data set in the training set, is the first The input nodes to the hidden layer The weights of the input nodes; is connected to the hidden layer The bias of the input node; is the bias of the output layer; is the transfer function, and ; The numbers in the brackets of the symbols in the formula (3) are the number of layers. The superscript (1) is the first layer, that is, the input layer, and the superscript (1, 2) is the first layer to the second layer, that is, the input layer to the hidden layer.

5. The device monitoring method according to claim 2, characterized in that: The root mean square error is represented by formula (4): （4）； in, is the root mean square error, is the number of input nodes or the normalized data set, For the The color values of the normalized dataset, For the The training results after the color value training of the normalized dataset is completed.

6. The device monitoring method according to claim 1, characterized in that: Step S9, marking the electrical equipment exceeding the preset temperature threshold as abnormally over-temperature equipment, and then including: Step S100, obtaining the location information of the abnormally overheated device and the color value of the thermal imaging corresponding to the abnormally overheated device; Step S200: sending the location information and the color value of the thermal imaging corresponding to the abnormally overheated device to an external receiving end.

7. An equipment monitoring device for a complementary power station, the equipment monitoring device being applied to the equipment monitoring method according to any one of claims 1 to 6, characterized in that: The equipment monitoring device comprises: A historical output maximum value acquisition module is used to respectively obtain the minimum historical output and the maximum historical output of the hydropower terminal and the photovoltaic terminal; an output interval construction module, configured to construct the output interval by taking the minimum of the two minimum historical outputs as the minimum value of the output interval and the maximum of the two maximum historical outputs as the maximum value of the output interval; An output sub-interval division module, configured to evenly divide the output interval into at least two consecutive sub-intervals; A time period definition module is used to define a time period based on each sub-interval, and the length of all time periods increases from small to large as the value of the output interval increases; A real-time output acquisition module, used to respectively acquire the real-time outputs of the hydropower terminal and the photovoltaic terminal; A real-time output classification module is used to classify all real-time outputs into all sub-intervals through a classification algorithm, and obtain a time period based on a real-time output; A thermal imaging acquisition module, configured to acquire thermal images of all electrical devices via the external shooting terminal at intervals of the time period; A thermal imaging judgment module, configured to judge whether the temperature of each electrical device exceeds a preset temperature threshold based on all thermal images; The abnormally over-temperature device marking module is used to mark electrical equipment that exceeds the preset temperature threshold as an abnormally over-temperature device.

8. An electronic device, characterized in that: It comprises a processor and a memory coupled to the processor, wherein the memory stores program instructions that can be executed by the processor; when the processor executes the program instructions stored in the memory, the device monitoring method according to any one of claims 1 to 6 is implemented.

9. A storage medium, characterized in that: The storage medium stores program instructions, which, when executed by a processor, can implement the device monitoring method according to any one of claims 1 to 6.

Images