Data management method and system for thermal power scene
By predicting historical power generation of generator sets and analyzing the features of thermal imaging images, an anomaly vector is constructed, which solves the problem of insufficient intelligent analysis in the data management of thermal power plants, enables timely detection and accurate identification of faults, and improves regulatory capabilities.
Patent Information
- Application Number
- CN202510961953.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-11
- Publication Date
- 2025-11-11
AI Technical Summary
Existing technologies lack intelligent analysis of collected data in thermal power plant data management, resulting in insufficient supervision and an inability to detect and determine fault types in a timely manner.
By predicting the historical power generation of generator sets, combining thermal imaging image features and noise features, an anomaly vector is constructed, and a pre-set fault analysis model is used to determine the fault type, including image fusion and outlier calculation.
It has improved the monitoring of faults in thermal power plants, enabling timely detection of suspected faults and accurate determination of fault types, and enhancing the intelligence level of data analysis.
Smart Images
Figure HDA0005496576230000011 
Figure HDA0005496576230000021
Abstract
Description
Technical Field
[0001] This invention relates to the field of power plant management technology, and in particular to a data management method and system for thermal power plants. Background Technology
[0003] First, data management is key to achieving refined operation of thermal power plants. By collecting, processing, and analyzing massive amounts of data during the production process in real time, power plants can accurately grasp key indicators such as equipment operating status, energy consumption, and environmental emissions. This data provides a scientific basis for optimizing production processes and improving energy conversion efficiency, ensuring the safe and stable operation of the power plant.
[0004] However, existing data management technologies often only store power plant operation data for easy viewing later, but lack further intelligent analysis of the collected data to improve the level of power plant supervision.
[0005] In view of this, the present invention is proposed. Summary of the Invention
[0006] The purpose of this invention is to provide a data management method and system for thermal power plants. This solution can further analyze the collected data to identify suspected faults. When a suspected fault occurs, the type of fault can be determined by further analyzing the data. By performing further intelligent analysis on the collected data, the level of supervision of power plants can be improved.
[0007] This invention provides a data management method for thermal power plant scenarios, the method comprising the following steps:
[0008] The power generation of the generator set is predicted based on the historical power generation of the generator set, and the power generation of the predicted time period is determined based on the predicted power generation.
[0009] If a fault analysis is to be performed on the generator set, the thermal imaging video of the generator set is retrieved, and multiple thermal images are extracted from the thermal imaging video.
[0010] Sub-images of the same location in multiple thermal images are fused to obtain a fused sub-image for each location in the corresponding thermal image, and the outlier value of each fused sub-image is calculated.
[0011] An anomaly vector is constructed based on the outliers of each fused sub-image. The anomaly vector is then input into a preset fault analysis model, and the fault type is output based on the fault analysis model.
[0012] The above-mentioned scheme first predicts the power generation of the generator set based on its historical power output. By comparing the predicted power output with the actual power output, it determines whether the generator set is suspected of having a fault. Since generator set faults are often accompanied by differences in heat distribution and noise, when a suspected fault occurs, an anomaly vector is constructed by calling the image features and noise features of the thermal imaging video. By analyzing the anomaly vector, the fault type is determined. This scheme can further analyze the collected data to identify suspected faults. When a suspected fault occurs, the type of fault can be determined by further analyzing the data. By performing further intelligent analysis on the collected data, the level of power plant supervision can be improved.
[0013] In some embodiments of the present invention, in the step of determining whether to perform fault analysis on the generator set based on the predicted power generation, the predicted power generation is compared with the actual power generation of the generator set during the predicted time period to determine whether to perform fault analysis on the generator set.
[0014] In some embodiments of the present invention, in the step of comparing the predicted power generation with the actual power generation of the generator set during the predicted time period and determining whether to perform fault analysis on the generator set, if the calculated predicted power generation is less than the actual power generation of the generator set during the predicted time period, and the difference between the predicted power generation and the actual power generation of the generator set during the predicted time period is less than a preset power threshold, then it is determined that fault analysis should be performed on the generator set; otherwise, fault analysis should not be performed on the generator set.
[0015] Using the above scheme, this scheme compares the predicted power generation with the actual power generation of the generator set during the predicted time period. If the actual power generation is less than the predicted power generation, it indicates that the generator set may be generating abnormally. Furthermore, the difference between the two is compared with a preset power threshold. Since the difference is negative, the smaller the negative value, the greater the difference between the prediction and the actual value, which indicates that the probability of the generator set being abnormal is greater. Through two-level analysis, the necessity of fault analysis is ensured.
[0016] In some embodiments of the present invention, in the step of image fusion of sub-images at the same location in multiple thermal imaging images, the thermal imaging images are segmented to divide each thermal imaging image into multiple sub-images.
[0017] In some embodiments of the present invention, in the step of segmenting the thermal image into multiple sub-images, the thermal image is uniformly divided into sub-images of a predetermined size, and each segmented sub-image is numbered, with sub-images at the same position in multiple thermal images having the same number.
[0018] In some embodiments of the present invention, the steps of fusing sub-images at the same location in multiple thermal imaging images to obtain a fused sub-image for each location in the corresponding thermal imaging image, and calculating the outlier value of each fused sub-image are as follows:
[0019] The fused sub-image is determined based on the sub-image number. For sub-images with the same number, the average pixel value of pixels at the same position is calculated to obtain the fused sub-image.
[0020] Each fused sub-image is input into a pre-defined graph neural network model to obtain the outlier value of each fused sub-image.
[0021] Using the above scheme, the thermal images of the same size corresponding to multiple time points are first segmented and labeled. Sub-images of the same location at multiple time points are then fused and analyzed to obtain the anomalies at that location over a period of time. This scheme can perform fragmented analysis on each location and improve the accuracy of the overall analysis through detailed local analysis.
[0022] In some embodiments of the present invention, in the step of constructing an anomaly degree vector based on the outliers of each fused sub-image, the weight value corresponding to the fused sub-image is determined based on the number corresponding to each fused sub-image, the weight value and the outlier corresponding to the fused sub-image are weighted and calculated to obtain a weighted outlier, and the anomaly degree vector is constructed based on the weighted outlier of each fused sub-image.
[0023] Using the above scheme, for the generator set, the normal operating temperature is different at each location. Therefore, this scheme sets different weight values for the corresponding fused sub-image of each location. Based on the weight values, the initially obtained outliers are further calculated to obtain the outlier vector of each location, ensuring the accuracy of the outlier calculation at each location.
[0024] In some embodiments of the present invention, in the step of constructing an anomaly degree vector based on the weighted outlier values of each fused sub-image, noise data corresponding to the time period of the thermal imaging video is called, noise dimension values are calculated based on the noise data, and the noise dimension values and the weighted outlier values corresponding to each fused sub-image are used as the values of each dimension in the anomaly degree vector to construct the anomaly degree vector.
[0025] In some embodiments of the present invention, in the step of calling noise data corresponding to the time period of the thermal imaging video recording and calculating the noise dimension value based on the noise data, the sound frequency at each time point in the noise data is statistically analyzed, the average value of the sound frequency is calculated, and normalized; the normalized sound frequency is used as the noise dimension value.
[0026] By adopting the above scheme, in the final anomaly vector construction step, both thermal imaging information and noise information are used as information in the anomaly vector. By fusing information from multiple dimensions, the accuracy of the final fault analysis is guaranteed.
[0027] Another aspect of the present invention relates to an electronic device having a computer program stored thereon, which, when executed by a processor, implements the steps of the aforementioned data management method for thermal power scenarios.
[0028] Another aspect of the present invention relates to a data management system for thermal power plant scenarios. The system includes a computer device, which includes a processor and a memory. The memory stores computer instructions, and the processor executes the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps of the method.
[0029] In summary, the present invention has the following beneficial effects:
[0030] 1. This solution first predicts the power generation of the generator set based on its historical power output. By comparing the predicted power output with the actual power output, it determines whether the generator set is suspected of having a fault. Since generator set faults are often accompanied by differences in heat distribution and noise, when a suspected fault occurs, an anomaly vector is constructed by calling the image features and noise features of the thermal imaging video. The fault type is determined by analyzing the anomaly vector. This solution can further analyze the collected data to identify suspected faults. When a suspected fault occurs, the type of fault can be determined by further analyzing the data. By performing further intelligent analysis on the collected data, the level of power plant supervision can be improved.
[0031] 2. This scheme compares the predicted power generation with the actual power generation of the generator set during the predicted time period. If the actual power generation is less than the predicted power generation, it indicates that the generator set may be generating abnormally. Furthermore, the difference between the two is compared with a preset power threshold. Since the difference is negative, the smaller the negative value, the greater the difference between the prediction and the actual value, which indicates that the probability of the generator set being abnormal is greater. Through two-level analysis, the necessity of fault analysis is ensured.
[0032] 3. This solution first segments and labels thermal images of the same size corresponding to multiple time points, then merges sub-images of the same location at multiple time points and analyzes the anomalies at that location over a period of time. This solution can perform fragmented analysis on each location, and improve the accuracy of the overall analysis through detailed local analysis.
[0033] 4. For generator sets, the normal operating temperature is different at each location. Therefore, this scheme sets different weight values for the corresponding fused sub-images at each location. Based on these weight values, the initially obtained outliers are further calculated to obtain the outlier vector at each location, ensuring the accuracy of the outlier calculation at each location. Attached Figure Description
[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0035] Figure 1 This is a schematic diagram of one embodiment of the data management method for thermal power plant scenarios according to the present invention;
[0036] Figure 2 This is a schematic diagram of another embodiment of the data management method of the present invention for thermal power scenarios. Detailed Implementation
[0037] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.
[0038] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The singular forms “a,” “the,” and “the” used in this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.
[0039] like Figure 1 As shown, the present invention provides a data management method for thermal power plant scenarios, the method comprising the following steps:
[0040] Step S100: Based on the historical power generation of the generator set, predict the power generation of the predicted time period, and determine whether to perform fault analysis on the generator set based on the predicted power generation.
[0041] In some embodiments of the present invention, a pre-set prediction model is used for prediction;
[0042] Specifically, the historical power generation includes power generation at multiple historical time points. An input vector is constructed from the power generation at multiple historical time points, and the input model is input into a pre-set prediction model. The prediction model outputs the predicted power generation.
[0043] The prediction model is either a Long Short-Term Memory (LSTM) network model or a Convolutional Neural Network (CNN) model.
[0044] In some embodiments of this invention, LSTM controls the retention and forgetting of information at different times by introducing three important gating mechanisms—the forget gate, the input gate, and the output gate. These gating mechanisms allow the network to dynamically decide which information should be retained in memory and which should be forgotten, thereby effectively solving the long-term dependency problem.
[0045] Forgotten Gate:
[0046] The forget gate determines which information is discarded from the cell state.
[0047] It generates a value between 0 and 1 using a sigmoid function, representing the degree to which each state value is preserved.
[0048] Values close to 0 indicate "forgotten", and values close to 1 indicate "retained".
[0049] Input Gate:
[0050] The input gate consists of two parts: a sigmoid layer and a tanh layer.
[0051] The sigmoid layer determines which values will be updated in the cell state.
[0052] The tanh layer generates a new candidate value vector.
[0053] The two outputs are multiplied to obtain the updated candidate value, which is then added to the cell state after the forget gate to obtain the new cell state.
[0054] Output gate:
[0055] The output gate determines the value of the next hidden state.
[0056] It first uses a sigmoid layer to determine which cell states will be output.
[0057] Then, candidate values for the output state are generated through a tanh layer.
[0058] Finally, the output of the sigmoid layer is multiplied by the output of the tanh layer to form the final output.
[0059] Step S200: If a fault analysis is to be performed on the generator set, the thermal imaging video of the generator set is retrieved, and multiple thermal images are extracted from the thermal imaging video.
[0060] In the specific implementation process, continuous shooting is carried out through thermal imaging technology. The thermal images obtained by continuous shooting are thermal imaging videos. The thermal images are also known as infrared thermal images (Far East Maps). They are graphic displays of the surface temperature distribution of an object, which are obtained by converting invisible light waves emitted by an object into visible light signals after being received by a detector and processed by an electronic computer.
[0061] Thermal imaging technology works on the fact that all objects generate heat. Infrared radiation is a type of electromagnetic wave with a wavelength longer than visible light, making it invisible to the human eye. However, when infrared radiation strikes an object, the object absorbs and reflects the radiation, creating hot or cold spots on an infrared detector. By converting these hot or cold spots into visible light signals and processing them with a computer, a thermal image of the object's surface temperature distribution can be obtained.
[0062] Step S300: The sub-images at the same location in multiple thermal imaging images are fused to obtain the fused sub-image at each location in the corresponding thermal imaging image, and the outlier value of each fused sub-image is calculated.
[0063] Step S400: Construct an anomaly degree vector based on the outliers of each fused sub-image, input the anomaly degree vector into a preset fault analysis model, and output the fault type based on the fault analysis model.
[0064] In the specific implementation process, the fault analysis model is a pre-trained convolutional neural network model. The convolutional neural network model includes a classification layer, which outputs the label corresponding to the fault type to determine the fault type.
[0065] In the specific implementation process, an anomaly degree vector is constructed based on the outliers of each fused sub-image, and the anomaly degree vector is input into a preset fault analysis model. The fault type is output based on the fault analysis model.
[0066] The above-mentioned scheme first predicts the power generation of the generator set based on its historical power output. By comparing the predicted power output with the actual power output, it determines whether the generator set is suspected of having a fault. Since generator set faults are often accompanied by differences in heat distribution and noise, when a suspected fault occurs, an anomaly vector is constructed by calling the image features and noise features of the thermal imaging video. By analyzing the anomaly vector, the fault type is determined. This scheme can further analyze the collected data to identify suspected faults. When a suspected fault occurs, the type of fault can be determined by further analyzing the data. By performing further intelligent analysis on the collected data, the level of power plant supervision can be improved.
[0067] In some embodiments of the present invention, in the step of determining whether to perform fault analysis on the generator set based on the predicted power generation, the predicted power generation is compared with the actual power generation of the generator set during the predicted time period to determine whether to perform fault analysis on the generator set.
[0068] In some embodiments of the present invention, in the step of comparing the predicted power generation with the actual power generation of the generator set during the predicted time period and determining whether to perform fault analysis on the generator set, if the calculated predicted power generation is less than the actual power generation of the generator set during the predicted time period, and the difference between the predicted power generation and the actual power generation of the generator set during the predicted time period is less than a preset power threshold, then it is determined that fault analysis should be performed on the generator set; otherwise, fault analysis should not be performed on the generator set.
[0069] In the specific implementation process, the historical power generation of the generator set is used for prediction. The predicted power generation is compared with the actual power generation to determine whether the generator set is suspected of having a fault. When a suspected fault occurs, an anomaly vector is constructed by calling the image features and noise features of the thermal imaging video. The fault type is determined by analyzing the anomaly vector.
[0070] Using the above scheme, this scheme compares the predicted power generation with the actual power generation of the generator set during the predicted time period. If the actual power generation is less than the predicted power generation, it indicates that the generator set may be generating abnormally. Furthermore, the difference between the two is compared with a preset power threshold. Since the difference is negative, the smaller the negative value, the greater the difference between the prediction and the actual value, which indicates that the probability of the generator set being abnormal is greater. Through two-level analysis, the necessity of fault analysis is ensured.
[0071] In some embodiments of the present invention, in the step of image fusion of sub-images at the same location in multiple thermal imaging images, the thermal imaging images are segmented to divide each thermal imaging image into multiple sub-images.
[0072] In some embodiments of the present invention, in the step of segmenting the thermal image into multiple sub-images, the thermal image is uniformly divided into sub-images of a predetermined size, and each segmented sub-image is numbered, with sub-images at the same position in multiple thermal images having the same number.
[0073] In specific implementation, the sub-image is an image of size 3*3 pixels, 4*4 pixels, or 5*5 pixels.
[0074] In the specific implementation process, when numbering the segmented sub-images, numerical order is used. Specifically, the segmented sub-images are numbered sequentially row by row.
[0075] like Figure 2 As shown, in some embodiments of the present invention, the step of fusing sub-images at the same location in multiple thermal imaging images to obtain a fused sub-image for each location in the corresponding thermal imaging image, and calculating the outlier value of each fused sub-image, includes:
[0076] Step S310: Determine the fused sub-image based on the sub-image number; calculate the average pixel value of pixels at the same position for sub-images with the same number to obtain the fused sub-image.
[0077] In practice, sub-images with the same number are located in the same position in different thermal imaging images.
[0078] Step S320: Input each fused sub-image into a preset graph neural network model to obtain the outlier value of each fused sub-image.
[0079] Using the above scheme, the thermal images of the same size corresponding to multiple time points are first segmented and labeled. Sub-images of the same location at multiple time points are then fused and analyzed to obtain the anomalies at that location over a period of time. This scheme can perform fragmented analysis on each location and improve the accuracy of the overall analysis through detailed local analysis.
[0080] Specifically, when a generator set is operating, it generates a large amount of heat internally. This heat mainly comes from the energy released during fuel combustion (for internal combustion engines such as diesel generator sets) or electromagnetic induction (for most electric generators). The uneven distribution of heat has a significant impact on the performance and lifespan of the generator set.
[0081] Firstly, for internal combustion engine generator sets, such as diesel generator sets, the high-temperature, high-pressure gas generated by the combustion of fuel in the cylinder is the main power source driving the piston and thus the generator rotor. During this process, components such as the cylinder liner, cylinder head, piston, and valves absorb a large amount of heat. The cylinder liner and cylinder head, in particular, are in direct contact with the high-temperature, high-pressure gas, resulting in extremely high temperatures. If not cooled in time, these components may experience thermal expansion that disrupts their normal clearances, or they may seize due to lubricant failure, or even lead to reduced mechanical strength and damage.
[0082] Secondly, for electric generators, such as turbine generators, the heat distribution is mainly concentrated in components such as the turbine, stator, and rotor. When the turbine rotates at high speed, a large amount of heat is generated due to the friction and compression of the gas medium. In particular, the turbine blades are not only subjected to the impact of high-temperature, high-pressure gas but also to the centrifugal force from high-speed rotation, resulting in extremely uneven temperature distribution on the blades. Hot spots often appear in specific locations on the blades, where the temperature is much higher than other parts, making them prone to overheating and hot-spot corrosion. The stator and rotor also generate heat during electromagnetic induction, especially when current passes through the stator coils, producing a large amount of Joule heat.
[0083] The cooling system plays a crucial role in effectively controlling heat distribution during generator operation. For internal combustion engine generator sets, air-cooled or water-cooled systems are typically used to reduce the temperature of components such as cylinder liners and cylinder heads. For electric generators, a well-designed cooling system, such as a circulating cooling air system or a water cooling system, ensures that the temperature of components such as the turbine, stator, and rotor remains within a safe range. These cooling systems not only effectively reduce the generator set's temperature but also improve its operating efficiency and reliability, and extend its service life.
[0084] In some embodiments of the present invention, in the step of constructing an anomaly degree vector based on the outliers of each fused sub-image, the weight value corresponding to the fused sub-image is determined based on the number corresponding to each fused sub-image, the weight value and the outlier corresponding to the fused sub-image are weighted and calculated to obtain a weighted outlier, and the anomaly degree vector is constructed based on the weighted outlier of each fused sub-image.
[0085] In the specific implementation process, the weight value of the fused sub-image at each location is preset. The staff sets the weight value of the fused sub-image according to the normal working temperature of the fused sub-image at different locations, and makes the weight value correspond to the number.
[0086] Using the above scheme, for the generator set, the normal operating temperature is different at each location. Therefore, this scheme sets different weight values for the corresponding fused sub-image of each location. Based on the weight values, the initially obtained outliers are further calculated to obtain the outlier vector of each location, ensuring the accuracy of the outlier calculation at each location.
[0087] In some embodiments of the present invention, in the step of constructing an anomaly degree vector based on the weighted outlier values of each fused sub-image, noise data corresponding to the time period of the thermal imaging video is called, noise dimension values are calculated based on the noise data, and the noise dimension values and the weighted outlier values corresponding to each fused sub-image are used as the values of each dimension in the anomaly degree vector to construct the anomaly degree vector.
[0088] In the specific implementation process, this solution uses recording equipment to record the audio data of the generator set during operation as noise data.
[0089] In some embodiments of the present invention, in the step of calling noise data corresponding to the time period of the thermal imaging video recording and calculating the noise dimension value based on the noise data, the sound frequency at each time point in the noise data is statistically analyzed, the average value of the sound frequency is calculated, and normalized; the normalized sound frequency is used as the noise dimension value.
[0090] In the specific implementation process, the average value of the sound frequency is normalized to between 0 and 1.
[0091] By adopting the above scheme, in the final anomaly vector construction step, both thermal imaging information and noise information are used as information in the anomaly vector. By fusing information from multiple dimensions, the accuracy of the final fault analysis is guaranteed.
[0092] Another aspect of the present invention relates to an electronic device having a computer program stored thereon, which, when executed by a processor, implements the steps of the aforementioned data management method for thermal power scenarios. The electronic device may be a tangible storage medium, such as random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, register, floppy disk, hard disk, removable storage disk, CD-ROM, or any other form of storage medium known in the art.
[0093] This invention also provides a data management system for thermal power plant scenarios. The system includes a computer device, which includes a processor and a memory. The memory stores computer instructions, and the processor executes the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps of the method.
[0094] Those skilled in the art will understand that the exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention. When implemented in hardware, it can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the desired tasks. The programs or code segments can be stored in a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave.
[0095] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.
[0096] In this invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or in place of features of other embodiments.
[0097] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations of the embodiments of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A data management method for thermal power plant scenarios, characterized in that, The steps of the method include: The power generation of the generator set is predicted based on the historical power generation of the generator set, and the power generation of the predicted time period is determined based on the predicted power generation. If a fault analysis is to be performed on the generator set, the thermal imaging video of the generator set is retrieved, and multiple thermal images are extracted from the thermal imaging video. Sub-images of the same location in multiple thermal images are fused to obtain a fused sub-image for each location in the corresponding thermal image, and the outlier value of each fused sub-image is calculated. An anomaly vector is constructed based on the outliers of each fused sub-image. The anomaly vector is then input into a preset fault analysis model, and the fault type is output based on the fault analysis model.
2. The data management method for thermal power plant scenarios according to claim 1, characterized in that: In the step of determining whether to perform fault analysis on the generator set based on the predicted power generation, the predicted power generation is compared with the actual power generation of the generator set during the predicted time period to determine whether to perform fault analysis on the generator set.
3. The data management method for thermal power plant scenarios according to claim 2, characterized in that: In the step of comparing the predicted power generation with the actual power generation of the generator set during the predicted time period to determine whether to perform fault analysis on the generator set, if the calculated predicted power generation is less than the actual power generation of the generator set during the predicted time period, and the difference between the predicted power generation and the actual power generation of the generator set during the predicted time period is less than a preset power threshold, then it is determined that fault analysis should be performed on the generator set; otherwise, fault analysis should not be performed on the generator set.
4. The data management method for thermal power plant scenarios according to claim 1, characterized in that: In the step of image fusion of sub-images at the same location in multiple thermal imaging images, the thermal imaging images are segmented to divide each thermal imaging image into multiple sub-images.
5. The data management method for thermal power plant scenarios according to claim 4, characterized in that: In the step of segmenting the thermal image into multiple sub-images, the thermal image is uniformly divided into sub-images of a predetermined size, and each sub-image is numbered. Sub-images at the same position in multiple thermal images are numbered the same.
6. The data management method for thermal power plant scenarios according to claim 5, characterized in that: In the step of fusing sub-images of the same location from multiple thermal images to obtain a fused sub-image for each location in the corresponding thermal image, and calculating the outlier value of each fused sub-image: The fused sub-image is determined based on the sub-image number. For sub-images with the same number, the average pixel value of pixels at the same position is calculated to obtain the fused sub-image. Each fused sub-image is input into a pre-defined graph neural network model to obtain the outlier value of each fused sub-image.
7. The data management method for thermal power plant scenarios according to claim 6, characterized in that: In the step of constructing an anomaly degree vector based on the outliers of each fused sub-image, the weight value corresponding to each fused sub-image is determined based on the number corresponding to each fused sub-image, the weight value and the outlier are weighted and calculated to obtain the weighted outlier value, and the anomaly degree vector is constructed based on the weighted outlier value of each fused sub-image.
8. The data management method for thermal power plant scenarios according to claim 1, characterized in that: In the step of constructing an anomaly degree vector based on the weighted outlier values of each fused sub-image, noise data corresponding to the time period of the thermal imaging video is called, noise dimension value is calculated based on the noise data, and the noise dimension value and the weighted outlier values corresponding to each fused sub-image are used as the values of each dimension in the anomaly degree vector to construct the anomaly degree vector.
9. The data management method for thermal power plant scenarios according to claim 8, characterized in that: In the step of calling the noise data corresponding to the time period of the thermal imaging video and calculating the noise dimension value based on the noise data, the sound frequency at each time point in the noise data is counted, the average sound frequency is calculated, and normalization is performed. Normalized sound frequencies are used as noise dimension values.
10. A data management system for thermal power plant scenarios, characterized in that: The system includes a computer device, which includes a processor and a memory. The memory stores computer instructions, and the processor executes the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps of the data management method for thermal power scenarios according to any one of claims 1-9.