Transformer intelligent control method and system based on artificial intelligence

Abstract

The invention discloses a transformer intelligent control method and system based on artificial intelligence, relates to the technical field of transformer intelligent control, and solves the technical problem that abnormity of a transformer is difficult to find and process in time due to the fact that a transformer monitoring control analysis scheme is unreasonable in the prior art. Comprising the following steps: analyzing the working state of the transformer based on use data to obtain a first priority coefficient of the transformer; analyzing the supply and demand state of the transformer based on the supply and demand data to obtain a second priority coefficient of the transformer; generating a data processing priority corresponding to the transformer based on the first priority coefficient and the second priority coefficient; sequentially acquiring monitoring data of each transformer based on the data processing priority of each transformer, and analyzing a fault risk based on the monitoring data to obtain a fault analysis result corresponding to the transformer; generating a control strategy of the transformer based on the fault analysis result; and the abnormity of the transformer can be timely and accurately found.

Description

Technical Field

[0001] This application belongs to the field of transformer intelligent control technology, specifically a transformer intelligent control method and system based on artificial intelligence. Background Technology

[0002] With the large-scale development of renewable energy, distributed power sources are increasingly being integrated into power distribution networks. Transformers, as a new generation of power electronic transformer equipment, have been widely used in regional power distribution systems, microgrids, and multi-energy complementary energy stations. Transformers typically possess functions such as high-frequency conversion, precise voltage regulation, and current direction control. Beyond traditional step-down transmission, they can further support bidirectional power flow regulation, dynamic power response, harmonic suppression, and fault isolation, making them indispensable power equipment in environments with a high proportion of renewable energy integration.

[0003] Existing technologies typically collect and analyze transformer data through fixed acquisition cycles, and then control the transformers accordingly based on the data analysis results. However, large substations usually have a large number of transformer groups. Using the above method to analyze transformers will lead to data processing congestion, making it difficult to detect and handle transformer anomalies in a timely manner. Therefore, an intelligent transformer control method and system based on artificial intelligence is needed. Summary of the Invention

[0004] This application provides an artificial intelligence-based intelligent control method for transformers, which solves the technical problem that makes it difficult to detect and handle transformer anomalies in a timely manner due to unreasonable transformer monitoring, control and analysis schemes in the prior art.

[0005] To achieve the above objectives, this application adopts the following technical solution: Firstly, an artificial intelligence-based intelligent control method for transformers is provided, including: Obtain usage data for each transformer; the usage data includes current, voltage, and other data collected from the transformers during this evaluation period; analyze the operating status of the transformers based on the usage data to obtain the priority coefficient of the transformer. Obtain supply and demand data for each transformer, which includes relevant data from the power supply side connected to the transformer input terminal and relevant data from the power consumption side connected to the transformer output terminal; including power supply data and power consumption data; analyze the supply and demand status of the transformer based on the supply and demand data to obtain the priority coefficient of the transformer. The data processing priority corresponding to the transformer is generated based on priority coefficient one and priority coefficient two; Based on the data processing priority of each transformer, the monitoring data of each transformer is obtained sequentially, and the fault risk is analyzed based on the monitoring data to obtain the fault analysis result corresponding to the transformer. The control strategy for the transformer is generated based on the fault analysis results.

[0006] Based on the above technical solution, in the transformer intelligent control method and system based on artificial intelligence provided in this application, the following steps are taken: First, the usage data of each transformer is acquired. Then, the working status of the transformer is analyzed based on the usage data to obtain a priority coefficient one for the transformer. Next, the supply and demand data of each transformer is acquired, and the supply and demand status of the transformer is analyzed based on the supply and demand data to obtain a priority coefficient two for the transformer. Finally, a data processing priority is generated for the transformer based on priority coefficient one and priority coefficient two. Then, monitoring data of each transformer is acquired sequentially according to the data processing priority of each transformer, and fault risk is analyzed based on the monitoring data to obtain the fault analysis result for the transformer. A control strategy for the transformer is generated based on the fault analysis result. By setting different processing priorities for different transformers, targeted decentralized monitoring, analysis, and control of each transformer in a transformer group are achieved. This enables timely and accurate detection and handling of transformer anomalies.

[0007] In conjunction with the first aspect above, in one possible implementation, obtaining the priority coefficient of the transformer based on data analysis includes: Extract the current and voltage values ​​collected at various times from the data; use the current and voltage values ​​(which are measured values ​​on the secondary side of the transformer); fit several current values ​​into a current variation curve according to the chronological order of their corresponding collection times; fit several voltage values ​​into a voltage variation curve according to the chronological order of their corresponding collection times. Substituting the voltage change curve and the current change curve into the set priority coefficient evaluation function one, the corresponding priority coefficient one is obtained; one expression of the priority coefficient evaluation function one includes: ; Wherein, YCO is the priority coefficient corresponding to the transformer, U(t) is the function symbol corresponding to the voltage change curve, I(t) is the function symbol corresponding to the current change curve, t0 is the set neighborhood step size, the smaller the neighborhood step size is, the more accurate the result will be; the amount of data to be processed will also increase accordingly. In this embodiment, the neighborhood step size is set to one percent of the acquisition period; t∈[t0, T-t0]; T is the time length of this acquisition period; t is the independent variable corresponding to U(t) and I(t), i.e. time; t' is the independent variable corresponding to U(t') and I(t'), i.e. time.

[0008] In conjunction with the first aspect mentioned above, in one possible implementation, the priority coefficient two for the transformer is obtained by analyzing supply and demand data, including: Extract power supply data and power consumption data from the supply and demand data; extract the unit power supply quantity at each collection time from the power supply data, and fit several of the unit power supply quantities into a power supply change curve according to the chronological order of their corresponding collection times; Extract the unit electricity consumption at each collection time from the electricity consumption data; fit the several unit electricity consumptions into an electricity consumption change curve according to the chronological order of their corresponding collection times; Substituting the power supply change curve and the power consumption change curve into the set priority coefficient evaluation function two, the corresponding priority coefficient two is obtained; one expression of the priority evaluation function two includes: ; Wherein, YCT is the priority coefficient 2 corresponding to the transformer, YQ(t) is the function symbol corresponding to the power consumption change curve, GQ(t) is the function symbol corresponding to the power supply change curve, t0 is the set neighborhood step size, the smaller the neighborhood step size is set, the more accurate the result, the corresponding amount of data to be processed will also increase, in this embodiment the neighborhood step size is set to one percent of the acquisition period; t∈[t0, T-t0]; T is the time length of this acquisition period; t is the independent variable corresponding to YQ(t) and GQ(t), i.e. time; t' is the independent variable corresponding to YQ(t') and GQ(t'), i.e. time.

[0009] In conjunction with the first aspect above, one possible implementation method for setting the acquisition time includes: Obtain the current and voltage values ​​acquired at least three previous acquisition times; and the interval between two acquisition times; substitute the current values, voltage values, and interval times into a set acquisition time adjustment function to obtain the interval time for the next acquisition time; one expression of the acquisition time adjustment function is as follows: ; Where Ti+1 is the adjusted interval time; Ti is the interval time corresponding to the i-th acquisition time; Ui is the voltage value acquired at the i-th acquisition time, and Ii is the current value acquired at the i-th acquisition time; Ti-1 is the interval time corresponding to the (i-1)-th acquisition time; Ui-1 is the voltage value acquired at the (i-1)-th acquisition time, and Ii-1 is the current value acquired at the (i-1)-th acquisition time; Ui-2 is the voltage value acquired at the (i-2)-th acquisition time, and Ii-2 is the current value acquired at the (i-2)-th acquisition time; K1 is the set maximum adjustable coefficient, and K1∈[0,1]; δ1 is the weighting coefficient of voltage, and δ2 is the weighting coefficient of current; the time after the interval time is recorded as the acquisition time.

[0010] In conjunction with the first aspect above, in one possible implementation, the data processing priority is generated based on priority coefficient one and priority coefficient two, including: Obtain priority coefficient 1 and priority coefficient 2 for each transformer, and perform a weighted summation of priority coefficient 1 and priority coefficient 2 to obtain the comprehensive priority coefficient of the transformer; obtain the comprehensive priority coefficient of each transformer in sequence; group the transformers into several processing groups according to the comprehensive priority coefficient in descending order, with a set number of transformers in each processing group; the set number is the number of transformers that the system can analyze simultaneously at one time; obtain the largest comprehensive priority coefficient in each processing group; sort the processing groups according to the comprehensive priority coefficient in descending order, and number them, using the number as the data processing priority of each transformer in the corresponding processing group.

[0011] In conjunction with the first aspect above, in one possible implementation, the fault analysis result is obtained by analyzing the fault risk based on monitoring data, including: Extract several monitoring values ​​collected from each monitoring item in the monitoring data; generate time-oriented analysis data groups and several monitoring item analysis data groups based on the several monitoring values ​​of each monitoring item; The fault risk assessment model is input into the time-based analysis data group and several monitoring project analysis data groups to obtain the fault analysis results corresponding to the transformer. The fault risk assessment model is obtained through artificial intelligence model training. The fault analysis results include several fault types and fault risk scores corresponding to the fault types.

[0012] In conjunction with the first aspect above, in one possible implementation, one training method for the fault risk assessment model includes: Acquire several historical monitoring data points and corresponding fault risk scores for each fault type. These fault risk scores are the result of expert analysis of the monitoring data, assessing the risk of the transformer exhibiting a specific fault type. A higher fault risk score indicates a higher probability of the transformer exhibiting the corresponding fault type. Generate time-based analysis data sets and several monitoring item analysis data sets based on several monitoring values ​​for each monitoring item. Integrate the time-based analysis data sets, the several monitoring item analysis data sets, and the fault risk scores corresponding to several fault types into training data and validation data. The AI ​​model is trained using training data and tested using validation data. The input consists of a time analysis data set and several monitoring item analysis data sets, and the output consists of several fault types and their corresponding fault risk scores. The fault types and their corresponding fault risk scores are integrated into a fault analysis result and then output. Finally, a fault risk assessment model is obtained, with the input consisting of a time analysis data set and several monitoring item analysis data sets, and the output consisting of fault analysis results. The AI ​​model includes one of the following: recurrent neural networks, deep neural networks, etc.

[0013] In conjunction with the first aspect above, in one possible implementation, the generation of time-oriented analysis data groups and analysis data groups for several monitoring items based on several monitoring values ​​of various monitoring items includes: Extract several monitoring values ​​collected from each monitoring item in the monitoring data; fit the monitoring values ​​of each monitoring item into a monitoring value change curve corresponding to the monitoring item according to the chronological order of their corresponding collection time; The various monitoring items and their corresponding monitoring value change curves are integrated into a time-oriented analysis data set; The monitoring values ​​at the same moment are extracted from the curves of the changes in various monitoring values, and the monitoring values ​​corresponding to the same moment for different monitoring items are integrated into a monitoring item analysis data group; several monitoring item analysis data groups are obtained in sequence.

[0014] In conjunction with the first aspect above, in one possible implementation, the control strategy is generated based on the fault analysis results, including: Extract the fault risk score for each fault type from the fault analysis results; The acquisition cycle of the transformer is generated based on the risk score of each fault type; When the fault risk score of a fault type is greater than the set fault risk threshold, the fault type is set as a risky fault type, and a solution for the fault type is obtained. The aforementioned solution and acquisition cycle are integrated into a control strategy.

[0015] In conjunction with the first aspect above, in one possible implementation, the collection period is generated based on the risk score for each fault type, including: Obtain the fault risk score for each fault type; When there is a fault risk score that is greater than the set fault risk threshold, the average of all fault risk scores that are greater than the fault risk threshold is calculated as the adjustment benchmark value. When all fault risk scores are less than the set fault risk threshold, the average of each fault risk score is calculated as the adjustment benchmark value. Substituting the adjusted reference value into the set period adjustment function yields the adjusted acquisition period; one expression of the period adjustment function is as follows: Wherein, CT is the adjusted acquisition period, YT is the acquisition period set before adjustment, which is a fixed value; K2 is the set adjustment coefficient, and K2∈[0,1], K2 is set to 0.5 in this embodiment; TJ is the adjustment benchmark value, and FY is the fault risk threshold.

[0016] Secondly, this application provides an artificial intelligence-based transformer intelligent control device, comprising: a processor and a storage medium; the storage medium includes instructions, and the processor is used to execute the instructions to implement the methods described in the first aspect and any possible implementation thereof. This artificial intelligence-based transformer intelligent control device can be an electronic device or a chip within an electronic device.

[0017] Thirdly, this application provides an artificial intelligence-based intelligent transformer control system, comprising: a data acquisition module, an acquisition and control module, a data processing module, and a database; wherein, The data acquisition module includes a monitoring data acquisition unit and a supply and demand data acquisition unit; The monitoring data acquisition unit acquires the usage data and monitoring data of each transformer through data acquisition devices installed on the transformers; the data acquisition devices include at least devices and sensors for acquiring voltage and current. The supply and demand data acquisition unit is used to acquire the supply and demand data of each transformer; The acquisition control module is used to generate the acquisition time for controlling the data acquisition module to acquire data. The data processing module includes a preliminary partitioning unit and a fault analysis unit. The preliminary division unit is used to analyze the operating status of the transformer based on usage data to obtain the transformer's priority coefficient one; to analyze the supply and demand status of the transformer based on supply and demand data to obtain the transformer's priority coefficient two; and to generate the data processing priority corresponding to the transformer based on priority coefficient one and priority coefficient two. The fault analysis unit is used to sequentially acquire monitoring data of each transformer based on the data processing priority of each transformer, analyze the fault risk based on the monitoring data to obtain the fault analysis result corresponding to the transformer, and generate the control strategy of the transformer based on the fault analysis result; the control strategy includes a solution and a data acquisition cycle.

[0018] Fourthly, this application provides a computer-readable storage medium storing instructions that, when executed on an AI-based transformer intelligent control device, cause the AI-based transformer intelligent control device to perform the methods described in the first aspect and any possible implementation thereof.

[0019] Fifthly, this application provides a computer program product containing instructions that, when run on an AI-based transformer intelligent control device, causes the AI-based transformer intelligent control device to perform the methods described in the first aspect and any possible implementation thereof.

[0020] This application provides an artificial intelligence-based intelligent control method and system for transformers. The system can acquire usage data from each transformer; analyze the operating status of the transformers based on the usage data to obtain a priority coefficient one; acquire supply and demand data from each transformer, analyze the supply and demand status of the transformers based on the supply and demand data to obtain a priority coefficient two; generate a data processing priority for each transformer based on priority coefficient one and priority coefficient two; acquire monitoring data from each transformer sequentially according to the data processing priority of each transformer, analyze fault risks based on the monitoring data to obtain fault analysis results for each transformer; generate a control strategy for the transformer based on the fault analysis results; and set processing priorities for different transformers to achieve targeted, decentralized monitoring, analysis, and control of each transformer in a transformer group. This enables timely and accurate detection and handling of transformer anomalies.

[0021] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of this application 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram illustrating the steps of the transformer intelligent control method in this application; Figure 2 This is a schematic diagram of the module connections of the transformer intelligent control system in this application. Detailed Implementation

[0024] The technical solutions of this application will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0025] Please see Figure 1 The first aspect of this application provides an artificial intelligence-based intelligent control method for transformers, comprising: Obtain usage data for each transformer; usage data includes current, voltage, and other data collected from the transformers during this evaluation period; analyze the operating status of the transformers based on the usage data to obtain a priority coefficient 1 for the transformers; priority coefficient 1 is used to indicate the impact of the transformer's operating status on its monitoring priority. The supply and demand data of each transformer are obtained. The supply and demand data includes relevant data from the power supply side connected to the transformer input terminal and relevant data from the power consumption side connected to the transformer output terminal; including power supply data and power consumption data. The supply and demand status of the transformer is analyzed based on the supply and demand data to obtain the priority coefficient 2 of the transformer. The priority coefficient 2 is used to indicate the impact of the transformer's supply and demand data on its monitoring priority. The data processing priority for the transformer is generated based on priority coefficient one and priority coefficient two; the data processing priority is the order in which the data corresponding to the transformer is processed. Based on the data processing priority of each transformer, the monitoring data of each transformer is obtained sequentially. The monitoring data is the monitoring data corresponding to the transformer, including several monitoring items and their corresponding monitoring values. Based on the monitoring data, the fault risk is analyzed to obtain the fault analysis result corresponding to the transformer. The control strategy for the transformer is generated based on the fault analysis results; the control strategy includes a solution and a data acquisition cycle.

[0026] Based on the above technical solution, in the transformer intelligent control method and system based on artificial intelligence provided in this application, the following steps are taken: First, the usage data of each transformer is acquired. Then, the working status of the transformer is analyzed based on the usage data to obtain a priority coefficient one for the transformer. Next, the supply and demand data of each transformer is acquired, and the supply and demand status of the transformer is analyzed based on the supply and demand data to obtain a priority coefficient two for the transformer. Finally, a data processing priority is generated for the transformer based on priority coefficient one and priority coefficient two. Then, monitoring data of each transformer is acquired sequentially according to the data processing priority of each transformer, and fault risk is analyzed based on the monitoring data to obtain the fault analysis result for the transformer. A control strategy for the transformer is generated based on the fault analysis result. By setting different processing priorities for different transformers, targeted decentralized monitoring, analysis, and control of each transformer in a transformer group are achieved. This enables timely and accurate detection and handling of transformer anomalies.

[0027] In one possible implementation, the priority coefficient of the transformer is obtained by analyzing the usage data, including: extracting the current and voltage values ​​collected at various times in the usage data; using the current and voltage values ​​as measured values ​​on the secondary side of the transformer; fitting several current values ​​into a current change curve according to the chronological order of their corresponding collection times; and fitting several voltage values ​​into a voltage change curve according to the chronological order of their corresponding collection times. Substituting the voltage and current variation curves into the set priority coefficient evaluation function one yields the corresponding priority coefficient one. Priority coefficient one evaluates the probability of transformer anomalies based on the fluctuations in current and voltage at the transformer output terminal, and sets a coefficient for priority handling. The stronger the current or voltage fluctuations, the greater the likelihood of transformer anomalies, requiring timely further data analysis; therefore, the priority coefficient one for this transformer is set relatively high. One expression of the priority coefficient evaluation function one includes: ; Wherein, YCO is the priority coefficient corresponding to the transformer, U(t) is the function symbol corresponding to the voltage change curve, I(t) is the function symbol corresponding to the current change curve, t0 is the set neighborhood step size, the smaller the neighborhood step size is, the more accurate the result will be; the amount of data to be processed will also increase accordingly. In this embodiment, the neighborhood step size is set to one percent of the acquisition period; t∈[t0, T-t0]; T is the time length of this acquisition period; t is the independent variable corresponding to U(t) and I(t), i.e. time; t' is the independent variable corresponding to U(t') and I(t'), i.e. time.

[0028] In this embodiment, the priority coefficient 1 is calculated using the above formula. The current change curve and voltage change curve collected within this acquisition period are analyzed. When the difference between each point on the current change curve or voltage change curve and the average value in its neighborhood is large, it indicates that the current or voltage fluctuation at the acquisition time is strong. At this time, the transformer may be abnormal, and further data analysis is required in time. Therefore, the priority coefficient 1 of the transformer is set to be large.

[0029] In one possible implementation, the priority coefficient of the transformer is obtained by analyzing the supply and demand data, including: extracting the power supply data and power consumption data from the supply and demand data; extracting the unit power supply at each collection time from the power supply data; and fitting the several unit power supply quantities into a power supply change curve according to the order of their corresponding collection times. Extract the unit electricity consumption at each collection time from the electricity consumption data; fit the several unit electricity consumptions into an electricity consumption change curve according to the chronological order of their corresponding collection times; Substitute the power supply change curve and the power consumption change curve into the set priority coefficient evaluation function two to obtain the corresponding priority coefficient two; Priority coefficient two is used to evaluate the impact of supply and demand fluctuations at both ends of the transformer on the transformer's operational volatility, and sets a coefficient for priority processing. When the supply and demand fluctuations at both ends of the transformer are relatively large, the transformer's operating state needs to be switched frequently, thus increasing the possibility of anomalies and requiring timely further data analysis. Therefore, the priority coefficient one for this transformer is set relatively high. One expression form of the priority coefficient evaluation function two includes: ; Wherein, YCT is the priority coefficient 2 corresponding to the transformer, YQ(t) is the function symbol corresponding to the power consumption change curve, GQ(t) is the function symbol corresponding to the power supply change curve, t0 is the set neighborhood step size, the smaller the neighborhood step size is set, the more accurate the result, the corresponding amount of data to be processed will also increase, in this embodiment the neighborhood step size is set to one percent of the acquisition period; t∈[t0, T-t0]; T is the time length of this acquisition period; t is the independent variable corresponding to YQ(t) and GQ(t), i.e. time; t' is the independent variable corresponding to YQ(t') and GQ(t'), i.e. time.

[0030] In this embodiment, the priority coefficient 2 is calculated using the above formula. The power supply change curve and power consumption change curve collected within this collection period are analyzed. When the difference between each point on the power supply change curve or power consumption change curve and the average value in its neighborhood is large, it indicates that the power supply demand or power consumption demand corresponding to the collection time is highly volatile. At this time, the working state of the transformer needs to be switched frequently, and the possibility of transformer abnormality is greater. Further data analysis needs to be carried out in a timely manner. Therefore, the priority coefficient 2 of the transformer is set to be large.

[0031] It is understandable that the priority coefficient evaluation function one and priority coefficient evaluation function two have the same form in this embodiment in order to ensure that the range of priority coefficient one and priority coefficient two is within a unified scale, so as to facilitate subsequent joint calculation.

[0032] In one possible implementation, setting the acquisition time includes: acquiring current and voltage values ​​acquired at least three previous acquisition times; and the interval between two acquisition times; in this acquisition cycle, the initial interval between the three acquisition times is a set acquisition interval, and the interval is adjusted at least after the third acquisition time; substituting the current value, voltage value, and interval into a set acquisition time adjustment function to obtain the interval for the next acquisition time; one expression of the acquisition time adjustment function is: ; Where Ti+1 is the adjusted interval time; Ti is the interval time corresponding to the i-th acquisition time; Ui is the voltage value acquired at the i-th acquisition time, and Ii is the current value acquired at the i-th acquisition time; Ti-1 is the interval time corresponding to the (i-1)-th acquisition time; Ui-1 is the voltage value acquired at the (i-1)-th acquisition time, and Ii-1 is the current value acquired at the (i-1)-th acquisition time; Ui-2 is the voltage value acquired at the (i-2)-th acquisition time, and Ii-2 is the current value acquired at the (i-2)-th acquisition time; K1 is the set maximum adjustable coefficient, and K1∈[0,1], K1 is set to 0.2 in this embodiment; δ1 is the weighting coefficient of voltage, and δ2 is the weighting coefficient of current; the initial values ​​of δ1 and δ2 in this embodiment are both 0.5; in another embodiment, δ1 and δ2 are calculated by the following formula: ; ; That is, when the current changes are more complex, δ2 is set to a larger value; when the voltage changes are more complex, δ1 is set to a larger value; and the time after the interval is recorded as the acquisition time.

[0033] This embodiment adaptively adjusts the timing of data acquisition using the aforementioned method. When the differences in the degree of change of current and voltage in recent acquisitions are relatively large, it indicates that the changes in current and voltage are more complex, requiring more data acquisition to ensure that the characteristics of subsequent analysis are as close to the real situation as possible, thereby guaranteeing the accuracy of the analysis results. Therefore, this embodiment increases the amount of data by reducing the interval between acquisition times. Conversely, when the differences in the degree of change of current and voltage in recent acquisitions are relatively small, it indicates that the changes in current and voltage are relatively smooth and simple. Acquiring too much data will not significantly improve the final analysis accuracy and will also waste computational resources. Therefore, this embodiment reduces the amount of data by appropriately increasing the interval between acquisition times, thereby reducing the waste of computational resources.

[0034] In one possible implementation, data processing priorities are generated based on priority coefficient one and priority coefficient two, including: obtaining priority coefficient one and priority coefficient two for each transformer; performing a weighted summation of priority coefficient one and priority coefficient two to obtain a comprehensive priority coefficient for the transformer; sequentially obtaining the comprehensive priority coefficient for each transformer; grouping each transformer into several processing groups according to the comprehensive priority coefficient in descending order, with each processing group containing a set number of transformers; the set number being the number of transformers that the system can analyze simultaneously at one time; obtaining the largest comprehensive priority coefficient in each processing group; sorting each processing group according to the comprehensive priority coefficient in descending order and numbering them, using the number as the data processing priority for each transformer in the corresponding processing group.

[0035] In one possible implementation, fault analysis results are obtained by analyzing fault risks based on monitoring data, including: extracting several monitoring values ​​collected from each monitoring item in the monitoring data; generating a time-oriented analysis data group and several monitoring item analysis data groups based on the several monitoring values ​​of each monitoring item; inputting the time-oriented analysis data group and several monitoring item analysis data groups into a fault risk assessment model to obtain the fault analysis results corresponding to the transformer, wherein the fault risk assessment model is trained through an artificial intelligence model; the fault analysis results include several fault types and fault risk scores corresponding to the fault types.

[0036] In one possible implementation, a training method for the fault risk assessment model includes: acquiring several historical monitoring data sets and fault risk scores for each fault type corresponding to the monitoring data; the fault risk scores are the results of expert analysis of the detection data, assessing the risk of the transformer corresponding to the detection data exhibiting the corresponding fault type, with a higher fault risk score indicating a higher probability of the transformer exhibiting the corresponding fault type; generating time-based analysis data sets and several monitoring item analysis data sets based on several monitoring values ​​of each monitoring item; and integrating the time-based analysis data sets, several monitoring item analysis data sets, and fault risk scores corresponding to several fault types into training data and validation data; The AI ​​model is trained using training data and tested using validation data. The input consists of a time analysis data set and several monitoring item analysis data sets, and the output consists of several fault types and their corresponding fault risk scores. The fault types and their corresponding fault risk scores are integrated into a fault analysis result and then output. Finally, a fault risk assessment model is obtained, with the input consisting of a time analysis data set and several monitoring item analysis data sets, and the output consisting of fault analysis results. The AI ​​model includes one of the following: recurrent neural networks, deep neural networks, etc.

[0037] In one possible implementation, a time-oriented analysis data group and a number of monitoring project analysis data groups are generated based on several monitoring values ​​of each monitoring project, including: extracting several monitoring values ​​collected from each monitoring project in the monitoring data; and fitting the monitoring values ​​of each monitoring project into a monitoring value change curve corresponding to the monitoring project according to the chronological order of their corresponding collection times. The various monitoring items and their corresponding monitoring value change curves are integrated into a time-oriented analysis data set; The monitoring values ​​at the same moment are extracted from the curves of various monitoring values. The monitoring values ​​corresponding to the same moment for different monitoring items are integrated into a monitoring item analysis data group. Several monitoring item analysis data groups are obtained in sequence. It can be understood that the monitoring item analysis data group contains the monitoring values ​​of different monitoring items at the same moment, which facilitates the correlation analysis between monitoring items. The time corresponding to different item analysis data groups is different.

[0038] In one possible implementation, generating a control strategy based on fault analysis results includes: extracting fault risk scores for each fault type from the fault analysis results; generating a collection cycle for the transformer based on the risk scores for each fault type; when the fault risk score for a fault type is greater than a set fault risk threshold; setting the fault type as a risk fault type; obtaining a solution for the fault type; and integrating the solution and the collection cycle into a control strategy.

[0039] In one possible implementation, the collection period is generated based on the risk score of each fault type, including: obtaining the fault risk score of each fault type; When there is a fault risk score that is greater than the set fault risk threshold, the average of all fault risk scores that are greater than the fault risk threshold is calculated as the adjustment benchmark value. When all fault risk scores are less than the set fault risk threshold, the average of each fault risk score is calculated as the adjustment benchmark value. Substituting the adjusted reference value into the set period adjustment function yields the adjusted acquisition period; one expression of the period adjustment function is as follows: Wherein, CT is the adjusted acquisition period, YT is the acquisition period set before adjustment, which is a fixed value; K2 is the set adjustment coefficient, and K2∈[0,1], K2 is set to 0.5 in this embodiment; TJ is the adjustment benchmark value, and FY is the fault risk threshold.

[0040] This embodiment uses a fault risk score feedback mechanism to adjust the next monitoring and control cycle for the transformer. The higher the fault risk score, the higher the probability of a problem occurring in the transformer. Therefore, it is necessary to conduct the next data acquisition and monitoring in a timely manner to ensure that the problem is detected promptly, the relevant parameters of the transformer are controlled, and the fault is resolved. Thus, this embodiment achieves the above function by reducing the acquisition cycle.

[0041] Secondly, this application provides an artificial intelligence-based transformer intelligent control device, comprising: a processor and a storage medium; the storage medium includes instructions, and the processor is used to execute the instructions to implement the methods described in the first aspect and any possible implementation thereof. This artificial intelligence-based transformer intelligent control device can be an electronic device or a chip within an electronic device.

[0042] Please see Figure 2 Thirdly, this application provides an artificial intelligence-based transformer intelligent control system, comprising: a data acquisition module, an acquisition and control module, a data processing module, and a database; wherein, The data acquisition module includes a monitoring data acquisition unit and a supply and demand data acquisition unit; The monitoring data acquisition unit acquires the usage data and monitoring data of each transformer through data acquisition devices installed on the transformers; the data acquisition devices include at least devices and sensors for acquiring voltage and current. The supply and demand data acquisition unit is used to acquire the supply and demand data of each transformer; The data acquisition control module is used to generate the acquisition time for controlling the data acquisition module to acquire data; The data processing module includes a preliminary partitioning unit and a fault analysis unit; The preliminary division unit is used to analyze the operating status of the transformer based on usage data to obtain the transformer's priority coefficient one; to analyze the supply and demand status of the transformer based on supply and demand data to obtain the transformer's priority coefficient two; and to generate the data processing priority corresponding to the transformer based on priority coefficient one and priority coefficient two. The fault analysis unit is used to sequentially acquire monitoring data of each transformer based on the data processing priority of each transformer, analyze the fault risk based on the monitoring data to obtain the fault analysis result corresponding to the transformer, and generate the control strategy of the transformer based on the fault analysis result; the control strategy includes a solution and a data acquisition cycle; The database is used to store all the data that this system needs and processes.

[0043] Fourthly, this application provides a computer-readable storage medium storing instructions that, when executed on an AI-based transformer intelligent control device, cause the AI-based transformer intelligent control device to perform the methods described in the first aspect and any possible implementation thereof.

[0044] Fifthly, this application provides a computer program product containing instructions that, when run on an AI-based transformer intelligent control device, causes the AI-based transformer intelligent control device to perform the methods described in the first aspect and any possible implementation thereof.

[0045] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0046] How this application works: By acquiring usage data of each transformer; analyzing the operating status of the transformers based on the usage data to obtain a priority coefficient one for each transformer; acquiring supply and demand data of each transformer, analyzing the supply and demand status of the transformers based on the supply and demand data to obtain a priority coefficient two for each transformer; generating a data processing priority for each transformer based on priority coefficient one and priority coefficient two; acquiring monitoring data of each transformer sequentially according to the data processing priority of each transformer, analyzing fault risks based on the monitoring data to obtain fault analysis results for each transformer; generating a control strategy for the transformer based on the fault analysis results; setting processing priorities for different transformers to achieve targeted decentralized monitoring, analysis, and control of each transformer in the transformer group; thereby enabling timely and accurate detection and handling of transformer anomalies.

[0047] The above embodiments are only used to illustrate the technical methods of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of this application without departing from the spirit and scope of the technical methods of this application.

Claims

1. A transformer intelligent control method based on artificial intelligence, characterized in that, include: Obtain usage data for each transformer; Based on the analysis of the operating status of the transformer using the data, the priority coefficient of the transformer is obtained. Obtain the supply and demand data of each transformer, analyze the supply and demand status of the transformer based on the supply and demand data, and obtain the priority coefficient of the transformer. The data processing priority corresponding to the transformer is generated based on priority coefficient one and priority coefficient two; Based on the data processing priority of each transformer, the monitoring data of each transformer is obtained sequentially, and the fault risk is analyzed based on the monitoring data to obtain the fault analysis result corresponding to the transformer. The control strategy for the transformer is generated based on the fault analysis results.

2. The intelligent transformer control method based on artificial intelligence according to claim 1, characterized in that, The priority coefficient one for the transformer, obtained through analysis of the usage data, includes: Extract the current and voltage values ​​collected at various times from the data; combine the current and voltage values; fit several current values ​​into a current change curve according to the chronological order of their corresponding collection times; fit several voltage values ​​into a voltage change curve according to the chronological order of their corresponding collection times. Substitute the voltage change curve and the current change curve into the set priority coefficient evaluation function one to obtain the corresponding priority coefficient one.

3. The intelligent transformer control method based on artificial intelligence according to claim 1, characterized in that, Based on the analysis of supply and demand data, the priority coefficient two for the transformer is obtained, including: Extract power supply data and power consumption data from the supply and demand data; extract the unit power supply quantity at each collection time from the power supply data, and fit several of the unit power supply quantities into a power supply change curve according to the chronological order of their corresponding collection times; Extract the unit electricity consumption at each collection time from the electricity consumption data; fit the several unit electricity consumptions into an electricity consumption change curve according to the chronological order of their corresponding collection times; Substitute the power supply change curve and the power consumption change curve into the set priority coefficient evaluation function two to obtain the corresponding priority coefficient two.

4. A transformer intelligent control method based on artificial intelligence according to claim 2 or 3, characterized in that, One method for setting the acquisition time includes: Obtain the current and voltage values ​​acquired at least three previous acquisition times; and the interval between two acquisition times; substitute the current values, voltage values, and interval times into a set acquisition time adjustment function to obtain the interval time for the next acquisition time; one expression of the acquisition time adjustment function is as follows: ； Where Ti+1 is the adjusted interval time; Ti is the interval time corresponding to the i-th acquisition time; Ui is the voltage value acquired at the i-th acquisition time, and Ii is the current value acquired at the i-th acquisition time; Ti-1 is the interval time corresponding to the (i-1)-th acquisition time; Ui-1 is the voltage value acquired at the (i-1)-th acquisition time, and Ii-1 is the current value acquired at the (i-1)-th acquisition time; Ui-2 is the voltage value acquired at the (i-2)-th acquisition time, and Ii-2 is the current value acquired at the (i-2)-th acquisition time; K1 is the set maximum adjustable coefficient, and K1∈[0,1]; δ1 is the weighting coefficient of voltage, and δ2 is the weighting coefficient of current; the time after the interval time is recorded as the acquisition time.

5. The intelligent transformer control method based on artificial intelligence according to claim 1, characterized in that, The data processing priority is generated based on priority coefficient one and priority coefficient two, including: Obtain priority coefficient 1 and priority coefficient 2 for each transformer, and perform a weighted summation of priority coefficient 1 and priority coefficient 2 to obtain the comprehensive priority coefficient of the transformer; obtain the comprehensive priority coefficient of each transformer in sequence; group the transformers into several processing groups according to the comprehensive priority coefficient in descending order, with a set number of transformers in each processing group; obtain the largest comprehensive priority coefficient in each processing group; sort the processing groups according to the comprehensive priority coefficient in descending order, and number them, using the number as the data processing priority of each transformer in the corresponding processing group.

6. The intelligent transformer control method based on artificial intelligence according to claim 1, characterized in that, The fault analysis results are obtained by analyzing the fault risks based on monitoring data, including: Extract several monitoring values ​​collected from each monitoring item in the monitoring data; generate time-oriented analysis data groups and several monitoring item analysis data groups based on the several monitoring values ​​of each monitoring item; The fault risk assessment model is obtained by inputting time into the analysis data group and several monitoring project analysis data groups to obtain the fault analysis results corresponding to the transformer. The fault risk assessment model is obtained by training an artificial intelligence model.

7. The intelligent transformer control method based on artificial intelligence according to claim 6, characterized in that, The data set for time-oriented analysis based on several monitoring values ​​of various monitoring items and the data set for analysis of several monitoring items include: Extract several monitoring values ​​collected from each monitoring item in the monitoring data; fit the monitoring values ​​of each monitoring item into a monitoring value change curve corresponding to the monitoring item according to the chronological order of their corresponding collection time; The various monitoring items and their corresponding monitoring value change curves are integrated into a time-oriented analysis data set; The monitoring values ​​at the same moment are extracted from the curves of the changes in various monitoring values, and the monitoring values ​​corresponding to the same moment for different monitoring items are integrated into a monitoring item analysis data group; several monitoring item analysis data groups are obtained in sequence.

8. The intelligent transformer control method based on artificial intelligence according to claim 1, characterized in that, The control strategy is generated based on the fault analysis results, including: Extract the fault risk score for each fault type from the fault analysis results; The acquisition cycle of the transformer is generated based on the risk score of each fault type; When the fault risk score of a fault type is greater than the set fault risk threshold, the fault type is set as a risky fault type, and a solution for the fault type is obtained. The aforementioned solution and acquisition cycle are integrated into a control strategy.

9. The intelligent transformer control method based on artificial intelligence according to claim 8, characterized in that, The collection cycle is generated based on the risk score for each fault type, including: Obtain the fault risk score for each fault type; When there is a fault risk score that is greater than the set fault risk threshold, the average of all fault risk scores that are greater than the fault risk threshold is calculated as the adjustment benchmark value. When all fault risk scores are less than the set fault risk threshold, the average of each fault risk score is calculated as the adjustment benchmark value. Substitute the adjusted baseline value into the set period adjustment function to obtain the adjusted acquisition period.

10. An artificial intelligence-based transformer intelligent control system, based on the application of an artificial intelligence-based transformer intelligent control method according to any one of claims 1 to 9; characterized in that, include: The system comprises a data acquisition module, an acquisition control module, a data processing module, and a database; among which, The data acquisition module includes a monitoring data acquisition unit and a supply and demand data acquisition unit; The monitoring data acquisition unit acquires the usage data and monitoring data of each transformer through data acquisition devices installed on the transformers; The supply and demand data acquisition unit is used to acquire the supply and demand data of each transformer; The acquisition control module is used to generate the acquisition time for controlling the data acquisition module to acquire data. The data processing module includes a preliminary partitioning unit and a fault analysis unit. The preliminary division unit is used to analyze the operating status of the transformer based on usage data to obtain the transformer's priority coefficient one; to analyze the supply and demand status of the transformer based on supply and demand data to obtain the transformer's priority coefficient two; and to generate the data processing priority corresponding to the transformer based on priority coefficient one and priority coefficient two. The fault analysis unit is used to sequentially acquire monitoring data of each transformer based on the data processing priority of each transformer, analyze the fault risk based on the monitoring data to obtain the fault analysis result corresponding to the transformer, and generate the control strategy of the transformer based on the fault analysis result; the control strategy includes a solution and a data acquisition cycle.

Images