Energy efficiency anomaly prediction agent construction method based on large model

Abstract

The invention relates to the field of energy efficiency anomaly analysis, in particular to an energy efficiency anomaly prediction agent construction method based on a large model, and the method comprises the steps: collecting and obtaining the operation data of a plurality of transformers in a transformer substation in a unit period, so as to extract the corresponding work operation characteristics of each transformer; in combination with the working operation characteristics and the operation state switching times, evaluating an operation energy consumption abnormity characterization value of the transformer substation so as to divide energy consumption degree categories of the transformer substation; and adaptively performing energy efficiency abnormity evaluation analysis on the transformer substation according to the energy consumption degree category. According to the invention, the accuracy and timeliness of substation energy efficiency abnormity identification are improved, and the energy loss is reduced. Meanwhile, potential faults can be recognized in advance, safety risks are reduced, and stable operation of the transformer substation is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of energy efficiency anomaly analysis, and in particular to a method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model. Background Technology

[0002] Currently, the operating environment of substations is becoming increasingly complex. On the one hand, the volatility of electricity load is constantly increasing, and the dynamic changes in various loads such as industrial production and residential life cause transformers to frequently operate in different states, such as no-load and load switching. On the other hand, substation equipment is gradually developing towards larger size and intelligence. Furthermore, as a core hub in the power system's transmission and distribution chain, the operational stability and energy efficiency of substations directly affect the reliability of power supply and energy utilization efficiency. Traditional manual monitoring and experience-based energy efficiency assessment methods are no longer sufficient to meet the real-time processing needs of massive amounts of operational data.

[0003] Driven by energy conservation and environmental protection policies, substation energy efficiency optimization has become an important development direction for the power industry. Accurately identifying energy efficiency anomalies and predicting potential faults in advance can not only reduce energy loss and operation and maintenance costs, but also avoid safety risks such as large-scale power outages caused by equipment failures.

[0004] Therefore, this invention provides a method for predicting energy efficiency anomalies in substations, based on a more comprehensive and systematic analytical framework, to achieve accurate prediction and efficient management of energy efficiency anomalies.

[0005] Chinese Patent Application Publication No. CN119561253A discloses a standardized intelligent control system for substations. This system achieves real-time response to equipment operating status by acquiring real-time data on current, voltage, power factor, temperature and humidity, operating status, insulation status, and ambient temperature and humidity of substation units. It employs a gradient boosting decision tree combined with a variational autoencoder and a self-attention mechanism to extract state vector features, accurately identify complex abnormal states, and detect potential faults early, significantly improving monitoring efficiency and early warning accuracy. By combining a long short-term memory network model with historical and real-time load data, it accurately predicts future power demand, overcoming the limitations of traditional static load management. Based on the predicted power demand, a reinforcement learning model dynamically adjusts the number of operating substation units, achieving optimal resource allocation while meeting demand, significantly improving the overall operating efficiency of the substation.

[0006] However, the following problems still exist in the existing technology. Many studies focus on real-time response to equipment operating status or single-dimensional energy consumption analysis. However, substation energy efficiency anomalies are affected by a combination of factors, including equipment characteristics, load fluctuations, and operating modes. This makes it difficult to comprehensively cover multiple sources of loss, resulting in low accuracy in identifying energy efficiency anomalies and delayed prediction of potential faults. Summary of the Invention

[0007] To address this, the present invention provides a method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model. This method overcomes the limitations of existing technologies, which often focus on real-time response of equipment operating status or single-dimensional energy consumption analysis. However, substation energy efficiency anomalies are influenced by multiple factors, making it difficult to comprehensively cover multiple sources of loss, which leads to low accuracy in identifying energy efficiency anomalies and delayed prediction of potential faults.

[0008] To achieve the above objectives, this invention provides a method for constructing an energy efficiency anomaly prediction agent based on a large model, comprising: The operation data of several transformers in the substation within a unit cycle are collected to extract the corresponding operating characteristics of each transformer. The operating characteristics include the average no-load time ratio and the average operating interval. Based on the operational characteristics and the number of operational state switching, the abnormal energy consumption characteristics of the substation are evaluated to classify the energy consumption level of the substation. Based on the energy consumption level category, an energy efficiency anomaly assessment and analysis is performed on the substation, including: For the transformers involved in the substation, the operating power time-domain variation curves are constructed. Based on the maximum load deviation value of each transformer in the energized state and the average temperature rise per unit time during state switching, it is determined whether the substation meets the energy efficiency loss benchmark, so as to determine whether the substation has an abnormal energy efficiency tendency. Based on the voltage fluctuation characteristics of the load served by the substation, the energy efficiency anomaly characterization parameters of the substation are calculated to determine the corresponding energy efficiency anomaly level and issue a corresponding early warning signal. Alternatively, based on the abnormal energy consumption characteristics, adjust the data collection and monitoring frequency for the substation; The voltage fluctuation characteristics include the maximum duration of a single fluctuation and the range of fluctuation amplitude.

[0009] Furthermore, the process of evaluating the abnormal operating energy consumption characteristics of the substation includes: The sum of the ratio of the average idle time percentage to the average idle time percentage threshold and the ratio of the average working interval threshold to the average working interval is taken as the first operating energy consumption characteristic. The ratio of the number of times the operating state is switched to the threshold number of times the operating state is switched is used as the second operating energy consumption characteristic; The weighted summation of the first operating energy consumption feature and the second operating energy consumption feature is determined as the abnormal operating energy consumption characterization value.

[0010] Furthermore, the energy consumption levels of the substations are categorized, including: If the abnormal energy consumption characteristic value of a substation is greater than or equal to the abnormal energy consumption characteristic threshold, the energy consumption level of the substation will be classified as a high energy consumption level category. If the abnormal energy consumption characteristic value of a substation is less than the abnormal energy consumption characteristic threshold, then the energy consumption level of the substation will be classified as a low energy consumption level category.

[0011] Furthermore, an energy efficiency anomaly assessment and analysis are conducted on the substation, including: If a substation is classified as a high-energy-consumption category, then the operating power time-domain variation curves of several transformers involved in the substation are constructed. Based on the maximum load deviation value of each transformer in the energized state and the average temperature rise per unit time during state switching, it is determined whether the substation meets the energy efficiency loss benchmark, so as to determine whether the substation has an abnormal energy efficiency tendency. Based on the voltage fluctuation characteristics of the load served by the substation, the energy efficiency anomaly characterization parameters of the substation are calculated to determine the corresponding energy efficiency anomaly level and issue a corresponding early warning signal. If a substation is classified as a low-energy-consumption category, the frequency of data collection and monitoring for that substation will be adjusted based on the abnormal energy consumption characteristics.

[0012] Further, determining whether the substation meets the energy efficiency loss benchmark includes: If the maximum load deviation is less than the maximum load deviation threshold, and the average temperature rise per unit time during state switching is less than the average temperature rise threshold, then the substation is determined to meet the energy efficiency loss benchmark.

[0013] Furthermore, determining whether the substation exhibits an abnormal energy efficiency trend includes: If a substation does not meet the energy efficiency loss benchmark, it is determined that the substation has an abnormal energy efficiency tendency.

[0014] Furthermore, the process of calculating the energy efficiency anomaly characterization parameters of the substation includes: The ratio of the maximum duration of a single fluctuation to the threshold of the maximum duration of a single fluctuation is used as the first energy efficiency anomaly feature; The ratio of the fluctuation range to the fluctuation range threshold is used as the second energy efficiency anomaly feature. The sum of the first energy efficiency anomaly feature and the second energy efficiency anomaly feature is used as the energy efficiency anomaly degree characterization parameter.

[0015] Furthermore, the process of determining the corresponding energy efficiency anomaly level includes: The correspondence between the energy efficiency anomaly level and the predetermined range of energy efficiency anomaly degree characterization parameters is pre-defined; Determine the energy efficiency anomaly characterization parameter range to which the corresponding energy efficiency anomaly characterization parameter of the substation belongs; The energy efficiency anomaly level corresponding to the range of energy efficiency anomaly characterization parameters is taken as the energy efficiency anomaly level of the substation. Among them, the energy efficiency anomaly level corresponds one-to-one with the range of parameters representing the degree of energy efficiency anomaly.

[0016] Furthermore, the frequency of data acquisition and monitoring for the substation is adjusted, including: The frequency of data acquisition and monitoring for substations is increased, and the increase in the frequency of data acquisition and monitoring is positively correlated with the abnormal energy consumption characterization value.

[0017] Furthermore, the process of constructing the time-domain variation curve of operating power includes: Construct a rectangular coordinate system with time as the horizontal axis and operating power as the vertical axis; The coordinates of the operating power at each moment are marked in the rectangular coordinate system; Connect the coordinate points with a smooth curve to obtain the time-domain variation curve of the operating power.

[0018] Compared with existing technologies, this invention collects operational data from several transformers within a substation over a unit cycle to extract the corresponding operational characteristics of each transformer. Combining these operational characteristics with the number of operational state switching cycles, it assesses the anomaly values ​​of the substation's energy consumption to classify the substation's energy consumption levels. Based on these energy consumption levels, it adaptively performs energy efficiency anomaly assessment and analysis on the substation. This invention improves the accuracy and timeliness of substation energy efficiency anomaly identification, reducing energy loss. Simultaneously, it can identify potential faults in advance, reducing safety risks and ensuring the stable operation of the substation.

[0019] In particular, this invention integrates three core loss sources: equipment idle time loss, operational stability loss, and additional loss due to state switching. The average no-load time percentage serves as a core indicator for assessing implicit no-load losses, quantifying the transformer's ineffective energy consumption and resource idleness. It directly reflects the proportion of no-load operation time of the transformer per unit cycle. Although the transformer does not output effective power under no-load conditions, it still generates fixed energy consumption such as iron losses. The average operating interval refers to the average time interval between two effective operating operations of the transformer. It is used to assess the impact of operating modes on energy consumption, reflecting whether equipment operation is continuous and whether there is an intermittent operating state with frequent start-stop cycles, quantifying the stability and continuity of transformer operation. Combined with the number of transformer operating state switching operations, during state switching, the transformer will generate instantaneous additional losses due to sudden changes in current and voltage, thus quantifying the scale of additional losses caused by transformer state switching. Furthermore, this invention characterizes the degree of abnormal energy consumption losses in substation energy efficiency and the overall energy intensity by evaluating abnormal operating energy consumption values. This invention provides accurate quantitative basis for subsequent energy consumption assessment and anomaly diagnosis. It can identify hidden energy consumption scenarios that are easily overlooked by traditional methods, such as transformer "no-load standby" and "frequent start-stop", thereby improving the accuracy of energy consumption assessment.

[0020] In particular, this invention designs differentiated diagnostic and assessment paths based on the energy consumption levels of substations. For high energy consumption levels, it focuses on the dynamic energy efficiency loss status and equipment health risks during transformer operation. The maximum load deviation value refers to the maximum difference in operating power between different loads for several transformers under energized operation, directly reflecting the balance of load distribution and the suitability of rated power, quantifying the load matching degree and overload risk during transformer operation. The temperature rise per unit time during state switching refers to the temperature rise per unit time during state switching processes such as "no-load-load" and "load-stop," directly related to the instantaneous energy loss during switching, used to quantify the energy loss efficiency and potential equipment health hazards of the transformer during dynamic switching. Therefore, this invention covers the energy efficiency loss status of transformers in two key stages: "steady-state operation" and "dynamic switching," based on the aforementioned two characteristics. It reflects both long-term load matching issues and captures energy waste and potential equipment hazards during instantaneous switching, providing data support for determining whether a substation meets energy efficiency loss benchmarks. This invention, based on the actual operating scenarios of substations, proactively captures potential energy efficiency anomalies and reduces loss risks.

[0021] In particular, this invention quantifies the impact of voltage fluctuations on substation energy efficiency from both time and intensity dimensions. The maximum duration of a single fluctuation directly reflects the duration of continuous interference with equipment operation, quantifying the sustained impact of voltage fluctuations. The fluctuation amplitude range refers to the difference between the maximum and minimum values ​​of voltage fluctuations within a unit period, reflecting the severity of voltage fluctuations and quantifying their intensity, directly correlated with the intensity of instantaneous energy efficiency loss. Furthermore, by comprehensively reflecting the persistence and intensity of the impact of voltage fluctuations on the stable operation of substations through these two characteristics, energy efficiency anomaly characterization parameters are calculated, reflecting the scale of energy efficiency loss caused by voltage instability and the severity of substation energy efficiency loss caused by voltage fluctuations, indirectly reflecting the urgency of requiring operation and maintenance intervention. This provides data support for subsequently determining the corresponding energy efficiency anomaly level. This invention improves the accuracy of energy efficiency anomaly level determination, enabling refined assessment, precise early warning, and rapid response to energy efficiency anomalies. Attached Figure Description

[0022] Figure 1 A schematic diagram illustrating the steps of a method for constructing an energy efficiency anomaly prediction agent based on a large model, as described in an embodiment of the invention. Figure 2 A logic diagram for classifying the energy consumption levels of substations according to embodiments of the invention; Figure 3 A logic diagram for determining whether a substation meets the energy efficiency loss benchmark in an embodiment of the invention; Figure 4 This is a logic diagram for determining whether a substation has an abnormal energy efficiency tendency, as shown in the embodiment of the invention. Detailed Implementation

[0023] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0024] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0025] Please see Figure 1 The diagram illustrates the steps of a method for constructing an energy efficiency anomaly prediction agent based on a large model, according to an embodiment of the present invention. The method includes: Step S1: Collect and obtain the operating data of several transformers in the substation within a unit cycle, so as to extract the corresponding operating characteristics of each transformer. The operating characteristics include the average no-load time ratio and the average operating interval. Step S2: Combining the operational characteristics and the number of operational state switching, evaluate the abnormal energy consumption characterization value of the substation to classify the energy consumption level of the substation. Step S3: Based on the energy consumption level category, perform an energy efficiency anomaly assessment and analysis on the substation, including: For the transformers involved in the substation, the operating power time-domain variation curves are constructed. Based on the maximum load deviation value of each transformer in the energized state and the average temperature rise per unit time during state switching, it is determined whether the substation meets the energy efficiency loss benchmark, so as to determine whether the substation has an abnormal energy efficiency tendency. Based on the voltage fluctuation characteristics of the load served by the substation, the energy efficiency anomaly characterization parameters of the substation are calculated to determine the corresponding energy efficiency anomaly level and issue a corresponding early warning signal. Alternatively, based on the abnormal energy consumption characteristics, adjust the data collection and monitoring frequency for the substation; The voltage fluctuation characteristics include the maximum duration of a single fluctuation and the range of fluctuation amplitude.

[0026] Specifically, the operating data includes operating characteristics, number of operating state switching times, operating power, maximum load deviation, average temperature rise per unit time during state switching, and voltage fluctuation characteristics.

[0027] Specifically, the acquisition of operational data can be done using any of the existing devices or systems that can acquire the corresponding data. For example, the duration and range of voltage fluctuations can be directly acquired using the existing voltage monitoring equipment in the substation without the need for additional dedicated hardware, which will not be elaborated further here.

[0028] Specifically, the unit period can be set according to the applicable scenario. For example, for service factories, industrial parks, etc., the load fluctuates frequently with the production plan, and the transformer starts and stops and the state switches many times. Therefore, the unit period can be set to 1 hour.

[0029] It is understandable that the state switching of a transformer is not instantaneous, but involves a dynamic transition process. For example, during the "no-load to load" state switch, the current gradually increases from near zero, winding losses and iron losses gradually increase, and the temperature rises slowly. Therefore, the unit time setting needs to cover the critical temperature rise stage in the dynamic transition process. In this embodiment, the unit time is selected within the range [2min, 5min], and 2min is preferred in practice.

[0030] Specifically, the process of assessing the abnormal energy consumption characteristics of the substation includes: The sum of the ratio of the average idle time percentage to the average idle time percentage threshold and the ratio of the average working interval threshold to the average working interval is taken as the first operating energy consumption characteristic. The ratio of the number of times the operating state is switched to the threshold number of times the operating state is switched is used as the second operating energy consumption characteristic; The weighted summation of the first operating energy consumption feature and the second operating energy consumption feature is determined as the abnormal operating energy consumption characterization value.

[0031] Specifically, the operational characteristics correspond to the transformer's static idle loss and steady-state operating loss. The average no-load duration ratio directly determines the cumulative scale of fixed iron losses. As long as the equipment is in an no-load state, losses will continue to occur, which is a core component of the substation's long-term energy consumption. The average operating interval reflects operational stability; excessively short intervals lead to repeated start-up losses throughout the operating cycle, exhibiting a continuous and regular impact. Both have a more stable and higher proportion of influence on total energy consumption. The number of operating state switching times corresponds to the instantaneous losses from dynamic switching, and since state switching behavior is intermittent, the cumulative loss ratio is usually lower than that of static idle loss and steady-state operating loss. Therefore, the first operational energy consumption characteristic calculated based on the operational characteristics, namely the average no-load duration ratio and the average operating interval, is assigned a higher weighting coefficient, set to 0.6. Correspondingly, the weighting coefficient for the second operational energy consumption characteristic calculated based on the number of operating state switching times is set to 0.4.

[0032] In this embodiment, the purpose of setting thresholds for the average no-load duration percentage, average operating interval, and number of operating state switching times is to characterize situations where the substation's energy efficiency exhibits a high degree of abnormal loss and a large overall energy consumption intensity. This is achieved by acquiring operating data corresponding to several historical unit cycles of the substation, and by calling the transformer's average no-load duration percentage data, average operating interval data, and number of operating state switching times data. The average no-load duration percentage, average operating interval, and average number of operating state switching times are then calculated and used as benchmark values ​​under normal conditions. Based on the purpose of setting the above three thresholds, the average no-load duration percentage, average operating interval, and average number of operating state switching times are determined. The load duration percentage threshold is determined as the product of the average idle time percentage and the percentage deviation coefficient. The average working interval threshold is determined as the product of the average working interval and the interval deviation coefficient. The number of operating state switching times threshold is determined as the product of the average number of operating state switching times and the switching deviation coefficient. The percentage deviation coefficient is selected within the interval [1.2, 1.3], preferably 1.2 in practice. The interval deviation coefficient is selected within the interval [0.9, 0.95], preferably 0.95 in practice. The switching deviation coefficient is selected within the interval [1.2, 1.4], preferably 1.2 in practice.

[0033] Specifically, this invention integrates three core loss sources: equipment idle time loss, operational stability loss, and additional loss from state switching. The average no-load time percentage serves as a core indicator for assessing implicit no-load loss, quantifying the transformer's ineffective energy consumption and resource idleness. It directly reflects the proportion of no-load operation time within a unit cycle. While the transformer does not output effective power under no-load conditions, it still generates fixed energy consumption such as iron losses. The longer the equipment idle time, the higher the proportion of ineffective energy consumption. The average operating interval refers to the average time interval between two effective operating operations of the transformer. It is used to assess the impact of operating modes on energy consumption, reflecting whether equipment operation is continuous and whether there are frequent start-stop intermittent operating states. It quantifies the stability and continuity of transformer operation. The shorter the interval, the more frequent the equipment operation is interrupted, easily leading to additional energy consumption during startup, such as a surge in copper losses; while the longer the interval, the more stable the operation and the lower the proportion of startup losses. This is combined with the number of transformer operating state switching times, i.e., the total number of times the transformer switches between different operating states such as "no-load-load" and "load-stop" within a unit cycle. During state switching operations, transformers experience instantaneous additional losses due to sudden changes in current and voltage, such as excitation losses and transient losses. This allows for the quantification of the scale of these additional losses caused by the transformer's state switching. Furthermore, this invention characterizes the degree of abnormal energy consumption and overall energy intensity of a substation by evaluating abnormal operating energy consumption indicators. Providing precise quantitative basis for subsequent energy consumption assessment and anomaly diagnosis, this invention can identify hidden energy consumption scenarios easily overlooked by traditional methods, such as transformer "no-load standby" and "frequent start-stop," thus improving the accuracy of energy consumption assessment.

[0034] Specifically, please refer to Figure 2 As shown, this is a logic diagram for classifying the energy consumption levels of substations according to an embodiment of the present invention. The classification of the energy consumption levels of the substations includes: If the abnormal energy consumption characteristic value of a substation is greater than or equal to the abnormal energy consumption characteristic threshold, the energy consumption level of the substation will be classified as a high energy consumption level category. If the abnormal energy consumption characteristic value of a substation is less than the abnormal energy consumption characteristic threshold, then the energy consumption level of the substation will be classified as a low energy consumption level category.

[0035] The abnormal operation energy consumption characterization threshold is predetermined. The abnormal operation energy consumption characterization value calculated when the average idle time ratio is equal to the average idle time ratio threshold, the average working interval threshold is equal to the average working interval threshold, and the number of operating state switching times is equal to the number of operating state switching times threshold is determined as the abnormal operation energy consumption characterization threshold.

[0036] Specifically, an energy efficiency anomaly assessment and analysis of the substation is conducted, including: If a substation is classified as a high-energy-consumption category, then the operating power time-domain variation curves of several transformers involved in the substation are constructed. Based on the maximum load deviation value of each transformer in the energized state and the average temperature rise per unit time during state switching, it is determined whether the substation meets the energy efficiency loss benchmark, so as to determine whether the substation has an abnormal energy efficiency tendency. Based on the voltage fluctuation characteristics of the load served by the substation, the energy efficiency anomaly characterization parameters of the substation are calculated to determine the corresponding energy efficiency anomaly level and issue a corresponding early warning signal. If a substation is classified as a low-energy-consumption category, the frequency of data collection and monitoring for that substation will be adjusted based on the abnormal energy consumption characteristics.

[0037] Specifically, the present invention adopts a differentiated energy efficiency anomaly assessment and analysis strategy, which can not only improve the identification accuracy of anomalies at high-risk sites, but also reduce the operation and maintenance costs of low-risk sites.

[0038] Specifically, please refer to Figure 3 As shown, this is a logic diagram for determining whether a substation meets the energy efficiency loss benchmark in an embodiment of the present invention. Determining whether the substation meets the energy efficiency loss benchmark includes: If the maximum load deviation is less than the maximum load deviation threshold, and the average temperature rise per unit time during state switching is less than the average temperature rise threshold, then the substation is determined to meet the energy efficiency loss benchmark.

[0039] In this embodiment, the purpose of setting the maximum load deviation threshold is to characterize the situation where the load matching degree of the transformer is low and the overload risk is high during operation. The purpose of setting the average temperature rise threshold is to characterize the situation where the instantaneous energy loss is large during state switching. By acquiring the operating data corresponding to several historical unit cycles of the substation, calling the maximum load deviation value data of the transformer and the average temperature rise data per unit time during state switching, the mean of the maximum load deviation and the mean of the average temperature rise are calculated and used as the benchmark values ​​under normal conditions. Based on the purpose of setting the above two thresholds, the maximum load deviation threshold is determined as the product of the mean of the maximum load deviation and the load deviation coefficient, and the average temperature rise threshold is determined as the product of the mean of the average temperature rise and the temperature rise deviation coefficient. The load deviation coefficient is selected in the interval [1.2, 1.4], preferably 1.2 in practice, and the temperature rise deviation coefficient is selected in the interval [1.3, 1.5], preferably 1.3 in practice.

[0040] Specifically, this invention designs differentiated diagnostic and assessment paths based on the energy consumption levels of substations. For high energy consumption levels, the focus is on the dynamic energy efficiency loss status and equipment health risks during transformer operation. The maximum load deviation value refers to the maximum difference between the operating power of several transformers under different loads when energized, directly reflecting the balance of load distribution and the degree of matching with rated power, quantifying the load matching degree and overload risk during transformer operation. A larger difference indicates more severe load fluctuations, such as exceeding the rated power in some time periods and falling far below the rated power in others, which can easily lead to a surge in overload losses or energy waste under low load conditions. The temperature rise per unit time during state switching refers to the temperature increase per unit time of the transformer during state switching processes such as "no-load-load" and "load-stop". It is directly related to the instantaneous energy loss during switching, such as excitation loss and transition loss. It is used to quantify the energy loss efficiency and potential health hazards of the transformer during dynamic switching. The larger the temperature rise, the more severe the energy loss during state switching, indicating that electrical energy is not effectively converted into output power but is instead dissipated as heat, resulting in more serious energy efficiency loss. At the same time, excessively rapid temperature rise may also reflect potential faults such as increased equipment contact resistance and coil aging. Therefore, this invention covers the energy efficiency loss status of transformers in two key stages, "steady-state operation" and "dynamic switching," based on the aforementioned two characteristics. It reflects both load matching problems in long-term operation and captures energy waste and potential equipment hazards during instantaneous switching, providing data support for determining whether a substation meets the energy efficiency loss benchmark. This invention, based on the actual operating scenario of substations, detects potential energy efficiency anomalies in advance, reducing loss risks.

[0041] Specifically, please refer to Figure 4 As shown, this is a logic diagram for determining whether a substation has an abnormal energy efficiency trend according to an embodiment of the present invention. Determining whether the substation has an abnormal energy efficiency trend includes: If a substation does not meet the energy efficiency loss benchmark, it is determined that the substation has an abnormal energy efficiency tendency. If a substation meets the energy efficiency loss benchmark, it is determined that the substation does not have an abnormal energy efficiency tendency.

[0042] Specifically, the process of calculating the energy efficiency anomaly characterization parameters of the substation includes: The ratio of the maximum duration of a single fluctuation to the threshold of the maximum duration of a single fluctuation is used as the first energy efficiency anomaly feature; The ratio of the fluctuation range to the fluctuation range threshold is used as the second energy efficiency anomaly feature. The sum of the first energy efficiency anomaly feature and the second energy efficiency anomaly feature is used as the energy efficiency anomaly degree characterization parameter.

[0043] In this embodiment, the purpose of setting the maximum duration threshold and the fluctuation range threshold for a single fluctuation is to characterize the severity of substation energy efficiency loss due to load voltage fluctuations. By acquiring the operating data corresponding to several historical unit cycles of the substation, calling the maximum duration data and fluctuation range data of a single fluctuation corresponding to the load served by the substation, the average maximum duration of a single fluctuation and the average fluctuation range are calculated and used as the benchmark values ​​under normal conditions. Based on the purpose of setting the above two thresholds, the maximum duration threshold for a single fluctuation is determined as the product of the average maximum duration of a single fluctuation and the duration deviation coefficient, and the fluctuation range threshold is determined as the product of the average fluctuation range and the fluctuation deviation coefficient. The duration deviation coefficient is selected within the interval [1.2, 1.4], preferably 1.2 in practice, and the fluctuation deviation coefficient is selected within the interval [1.3, 1.4], preferably 1.3 in practice.

[0044] Specifically, this invention quantifies the impact of voltage fluctuations on substation energy efficiency from both time and intensity dimensions. The maximum duration of a single fluctuation directly reflects the duration of continuous interference with equipment operation, quantifying the sustained impact of voltage fluctuations. The longer the duration, the more persistent the impact of voltage instability, which can easily lead to a decrease in operating efficiency, such as reduced motor efficiency, increased transformer iron losses, and a larger cumulative scale of implicit energy efficiency losses. At the same time, long-term fluctuations may also exacerbate equipment fatigue, indirectly affecting long-term energy efficiency stability. The fluctuation amplitude range refers to the difference between the maximum and minimum values ​​of voltage fluctuations within a unit period. It reflects the severity of voltage fluctuations and quantifies the intensity of voltage fluctuation impact, directly related to the intensity of instantaneous energy efficiency losses. The larger the fluctuation amplitude range, the more significant the deviation of the voltage from the stable state, and the stronger the impact on the power conversion efficiency of the equipment. For example, excessively high voltage can lead to a surge in equipment overheating losses, while excessively low voltage can lead to insufficient output power and an increase in energy consumption ratio. Furthermore, by comprehensively reflecting the persistence and intensity of the impact of voltage fluctuations on the stable operation of substations through the above two characteristics, the energy efficiency anomaly characterization parameters are calculated to reflect the scale of energy efficiency losses caused by voltage instability and the severity of substation energy efficiency losses caused by voltage fluctuations, indirectly reflecting the urgency of requiring operation and maintenance intervention. This provides data support for subsequently determining the corresponding energy efficiency anomaly level. This invention improves the accuracy of energy efficiency anomaly level determination, enabling refined assessment, accurate early warning, and rapid response to energy efficiency anomalies.

[0045] Specifically, the process of determining the corresponding energy efficiency anomaly level includes: The correspondence between the energy efficiency anomaly level and the predetermined range of energy efficiency anomaly degree characterization parameters is pre-defined; Determine the energy efficiency anomaly characterization parameter range to which the corresponding energy efficiency anomaly characterization parameter of the substation belongs; The energy efficiency anomaly level corresponding to the range of energy efficiency anomaly characterization parameters is taken as the energy efficiency anomaly level of the substation. Among them, the energy efficiency anomaly level corresponds one-to-one with the range of parameters representing the degree of energy efficiency anomaly.

[0046] In this embodiment, the energy efficiency anomaly level of the substation is determined according to the following method: The parameters characterizing the degree of energy efficiency anomaly are divided into three preset intervals, and three levels of energy efficiency anomaly are set. If the energy efficiency anomaly characterization parameter corresponding to the substation is within the first preset range (0, E0), then it corresponds to the first energy efficiency anomaly level. If the energy efficiency anomaly characterization parameter corresponding to the substation is within the second preset range [E0, 1.3E0), then it corresponds to the second energy efficiency anomaly level. If the energy efficiency anomaly characterization parameter corresponding to the substation is within the third preset range [1.3E0,+∞), then it corresponds to the third energy efficiency anomaly level.

[0047] The threshold E0 for the energy efficiency anomaly characterization parameter is predetermined. The energy efficiency anomaly characterization parameter calculated under the condition that the maximum duration of a single fluctuation is equal to the threshold of the maximum duration of a single fluctuation, and the fluctuation amplitude range is equal to the threshold of the fluctuation amplitude range, is determined as the energy efficiency anomaly characterization parameter threshold.

[0048] Specifically, the warning signals correspond one-to-one with the energy efficiency anomaly levels.

[0049] In this embodiment, the correspondence between the warning signal and the energy efficiency anomaly level is set in the following manner: Warning signal 1 corresponds to the first energy efficiency anomaly level. Warning signal 1 includes: a green text prompt from the SCADA system and a message reminder sent to the operation and maintenance platform. Warning signal 2 corresponds to the second energy efficiency anomaly level. Warning signal 2 includes: a yellow flashing pop-up window in the SCADA system, intermittent flashing of the yellow light in the control room, and a low-voltage buzzer, pushing information to relevant operation and maintenance personnel; Warning signal 3 corresponds to the third energy efficiency anomaly level. Warning signal 3 includes: a red full-screen pop-up window in the SCADA system, a constantly lit red light in the control room, a continuous high-voltage buzzer, and multi-terminal information push notifications via platform / SMS / telephone.

[0050] Specifically, the warning signals corresponding to the energy efficiency anomaly level are directly related to voltage fluctuation characteristics, allowing relevant maintenance personnel to quickly locate the root cause of the anomaly. For example, if the duration exceeds the standard, it may be due to the lag in the response of the voltage regulator; if the amplitude exceeds the standard, it may be due to sudden load changes or line faults. There is no need to check each device one by one, which shortens the problem handling cycle and reduces the energy efficiency loss and power supply stability risks caused by the continuous anomaly. This will not be elaborated further.

[0051] Specifically, adjusting the data acquisition and monitoring frequency for the substation includes: The frequency of data acquisition and monitoring for substations is increased, and the increase in the frequency of data acquisition and monitoring is positively correlated with the abnormal energy consumption characterization value.

[0052] In this embodiment, optionally, The abnormal operating energy consumption value is compared with the preset first and second abnormal operating energy consumption comparison thresholds. When the abnormal energy consumption indicator value is greater than the second abnormal energy consumption indicator comparison threshold, the increase in the acquisition and monitoring frequency is determined as the first increase, which is set to 0.5 times the initial acquisition and monitoring frequency. When the abnormal energy consumption indicator value is greater than or equal to the first abnormal energy consumption indicator comparison threshold and less than or equal to the second abnormal energy consumption indicator comparison threshold, the increase in the acquisition and monitoring frequency is determined to be the second increase, which is set to be 0.3 times the initial acquisition and monitoring frequency. When the abnormal energy consumption indicator value is less than the first abnormal energy consumption indicator comparison threshold, the increase in the acquisition and monitoring frequency is determined to be the third increase, which is set to be 0.2 times the initial acquisition and monitoring frequency. The first threshold for comparing abnormal operating energy consumption is 1.1 times the threshold for abnormal operating energy consumption, and the second threshold for comparing abnormal operating energy consumption is 1.3 times the threshold for abnormal operating energy consumption.

[0053] Understandably, the purpose of increasing the data collection and monitoring frequency is to dynamically adapt to changes in energy consumption, accurately capture abnormal trends in high-energy-consumption states, and avoid wasting operational resources in low-risk states. Furthermore, the amount of increase in the data collection and monitoring frequency can be adjusted by those skilled in the art under different circumstances. Those skilled in the art can determine the initial acquisition and monitoring frequency based on the scale of the substation equipment. For example, for a 220kV industrial hub substation with many devices and large load fluctuations, high-frequency capture of anomalies is required, so the initial acquisition and monitoring frequency is set to 10 minutes / time. This will not be elaborated further.

[0054] Specifically, this invention constructs a dynamic acquisition and monitoring frequency control mechanism that is positively correlated with the abnormal energy consumption characteristic value. The abnormal energy consumption characteristic value is used as a continuous variable for frequency control, enabling gradient and fine-grained adjustment of the acquisition and monitoring frequency. Simultaneously, it balances monitoring timeliness and resource economy, improving resource utilization efficiency and achieving dynamic allocation of operation and maintenance resources.

[0055] Specifically, the process of constructing the time-domain variation curve of operating power includes: Construct a rectangular coordinate system with time as the horizontal axis and operating power as the vertical axis; The coordinates of the operating power at each moment are marked in the rectangular coordinate system; Connect the coordinate points with a smooth curve to obtain the time-domain variation curve of the operating power.

[0056] Specifically, there are no restrictions on the method for constructing the operating power time-domain variation curve. For example, the operating power time-domain variation curve can be fitted using Matlab correlation fitting software, which will not be elaborated further.

[0057] Understandably, the operating power time-domain variation curve visualizes the operating power data at each moment with time as the axis, which can clearly show the power fluctuation pattern of the transformer.

[0058] If the energy efficiency anomaly prediction intelligent agent construction method based on a large model of the present invention is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0059] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model, characterized in that, include: The operation data of several transformers in the substation within a unit cycle are collected to extract the corresponding operating characteristics of each transformer. The operating characteristics include the average no-load time ratio and the average operating interval. Based on the operational characteristics and the number of operational state switching, the abnormal energy consumption characteristics of the substation are evaluated to classify the energy consumption level of the substation. Based on the energy consumption level category, an energy efficiency anomaly assessment and analysis is performed on the substation, including: For the transformers involved in the substation, the operating power time-domain variation curves are constructed. Based on the maximum load deviation value of each transformer in the energized state and the average temperature rise per unit time during state switching, it is determined whether the substation meets the energy efficiency loss benchmark, so as to determine whether the substation has an abnormal energy efficiency tendency. Based on the voltage fluctuation characteristics of the load served by the substation, the energy efficiency anomaly characterization parameters of the substation are calculated to determine the corresponding energy efficiency anomaly level and issue a corresponding early warning signal. Alternatively, based on the abnormal energy consumption characteristics, adjust the data collection and monitoring frequency for the substation; The voltage fluctuation characteristics include the maximum duration of a single fluctuation and the range of fluctuation amplitude.

2. The method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model according to claim 1, characterized in that, The process of assessing the abnormal operating energy consumption characteristics of the substation includes: The sum of the ratio of the average idle time percentage to the average idle time percentage threshold and the ratio of the average working interval threshold to the average working interval is taken as the first operating energy consumption characteristic. The ratio of the number of times the operating state is switched to the threshold number of times the operating state is switched is used as the second operating energy consumption characteristic; The weighted summation of the first operating energy consumption feature and the second operating energy consumption feature is determined as the abnormal operating energy consumption characterization value.

3. The method for constructing an energy efficiency anomaly prediction agent based on a large model according to claim 2, characterized in that, The energy consumption levels of the substations are categorized as follows: If the abnormal energy consumption characteristic value of a substation is greater than or equal to the abnormal energy consumption characteristic threshold, then the energy consumption level of the substation will be classified as a high energy consumption level category. If the abnormal energy consumption characteristic value of a substation is less than the abnormal energy consumption characteristic threshold, then the energy consumption level of the substation will be classified as a low energy consumption level category.

4. The method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model according to claim 3, characterized in that, An energy efficiency anomaly assessment and analysis was conducted on the substation, including: If a substation is classified as a high-energy-consumption category, then the operating power time-domain variation curves of several transformers involved in the substation are constructed. Based on the maximum load deviation value of each transformer in the energized state and the average temperature rise per unit time during state switching, it is determined whether the substation meets the energy efficiency loss benchmark, so as to determine whether the substation has an abnormal energy efficiency tendency. Based on the voltage fluctuation characteristics of the load served by the substation, the energy efficiency anomaly characterization parameters of the substation are calculated to determine the corresponding energy efficiency anomaly level and issue a corresponding early warning signal. If a substation is classified as a low-energy-consumption category, the collection and monitoring frequency for that substation will be adjusted based on the abnormal operating energy consumption characteristic value.

5. The method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model according to claim 1, characterized in that, Determining whether the substation meets the energy efficiency loss benchmark includes: If the maximum load deviation is less than the maximum load deviation threshold, and the average temperature rise per unit time during state switching is less than the average temperature rise threshold, then the substation is determined to meet the energy efficiency loss benchmark.

6. The method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model according to claim 5, characterized in that, Determining whether the substation exhibits an abnormal energy efficiency trend includes: If a substation does not meet the energy efficiency loss benchmark, it is determined that the substation has an abnormal energy efficiency tendency.

7. The method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model according to claim 1, characterized in that, The process of calculating the energy efficiency anomaly characterization parameters of the substation includes: The ratio of the maximum duration of a single fluctuation to the threshold of the maximum duration of a single fluctuation is used as the first energy efficiency anomaly feature; The ratio of the fluctuation range to the fluctuation range threshold is used as the second energy efficiency anomaly feature. The sum of the first energy efficiency anomaly feature and the second energy efficiency anomaly feature is used as the energy efficiency anomaly degree characterization parameter.

8. The method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model according to claim 7, characterized in that, The process of determining the corresponding energy efficiency anomaly level includes: The correspondence between the energy efficiency anomaly level and the predetermined range of energy efficiency anomaly degree characterization parameters is pre-defined; Determine the range of energy efficiency anomaly characterization parameters to which the corresponding energy efficiency anomaly characterization parameters for the substation belong; The energy efficiency anomaly level corresponding to the range of energy efficiency anomaly characterization parameters is taken as the energy efficiency anomaly level of the substation. Among them, the energy efficiency anomaly level corresponds one-to-one with the range of parameters representing the degree of energy efficiency anomaly.

9. The method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model according to claim 1, characterized in that, Adjusting the data acquisition and monitoring frequency for the substation, including: Increase the frequency of data acquisition and monitoring for substations, and the increase in the frequency of data acquisition and monitoring is positively correlated with the abnormal energy consumption characterization value.

10. The method for constructing an intelligent agent for predicting energy efficiency anomalies based on a large model according to claim 1, characterized in that, The process of constructing the time-domain variation curve of operating power includes: Construct a rectangular coordinate system with time as the horizontal axis and operating power as the vertical axis; The coordinates of the operating power at each moment are marked in the rectangular coordinate system; Connect the coordinate points with a smooth curve to obtain the time-domain variation curve of the operating power.

Images