A method and system for controlling electric power equipment based on the Internet of Things

Abstract

The present invention provides an Internet of Things-based power equipment control method and system, which relates to the field of power equipment operation and maintenance technology. The method includes: obtaining historical equipment status data and taking out multiple groups of temperature adjustment reference data, extracting temperature adjustment characteristic values ​​and characteristic change values ​​of multiple parameters and calculating the temperature adjustment correlation factor of each parameter; determining multiple target parameters and extracting training data sets from multiple groups of temperature adjustment reference data, training to obtain a temperature adjustment analysis model; collecting target equipment status data of a target transformer and extracting the first target characteristic values ​​of multiple target parameters and the change parameters of each first target characteristic value, calculating multiple second target characteristic values, and generating a target score through a temperature adjustment analysis model; when the target score is greater than a preset warning threshold, the target transformer is temperature-regulated based on a preset temperature adjustment method. The present invention achieves the improvement of the stability of the transformer in a complex operating environment.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment operation and maintenance, and in particular to a power equipment control method and system based on the Internet of Things. Background Art

[0002] As a key power equipment in modern power systems, the operating status of transformers directly affects the safety and stability of the entire power network. The temperature of transformer windings is one of the important factors affecting its life and operational reliability. Traditional temperature control methods are usually based on temperature data collected in real time by temperature sensors. When the winding temperature exceeds the set threshold, the system automatically starts the cooling system for temperature adjustment. Although this method can protect the transformer to a certain extent, it also has obvious limitations. Relying on a single temperature parameter for temperature adjustment ignores other important factors that affect the temperature of the transformer, such as load current, ambient temperature, oil temperature and other related parameters. Temperature adjustment behavior is often the result of the combined effect of multiple complex operating conditions. Simply relying on real-time data from temperature sensors for temperature adjustment may lead to too late response or too many false alarms, and cannot predict and prevent potential failure risks. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes an Internet of Things-based power equipment control method and system, which collects multiple parameters of the transformer in real time through the Internet of Things technology, analyzes relevant data based on the temperature regulation analysis model, and issues early warning for potential overtemperature phenomena, so as to facilitate timely adjustment of the transformer winding temperature and improve the stability of the transformer operation in complex operating environments.

[0004] To achieve the above object, the present invention provides a first aspect of a power equipment control method based on the Internet of Things, comprising:

[0005] Acquire historical device status data of the target transformer collected by the IoT device, and extract multiple sets of temperature adjustment reference data from the historical device status data according to the historical temperature adjustment data of the target transformer;

[0006] Extracting temperature regulation characteristic values ​​and characteristic change values ​​of multiple parameters in each set of temperature regulation reference data, and calculating the temperature regulation correlation factor of each parameter in each set of temperature regulation reference data according to the temperature regulation characteristic values ​​and characteristic change values;

[0007] Determine multiple target parameters according to multiple temperature regulation correlation factors of each parameter, extract training data sets from multiple sets of temperature regulation reference data according to the multiple target parameters, obtain a temperature regulation analysis model through training the training data sets and determine a preset warning threshold;

[0008] Collect target device status data of the target transformer based on the Internet of Things device, and extract first target characteristic values ​​of multiple target parameters in the target device status data and change parameters of each first target characteristic value;

[0009] The second target characteristic value corresponding to each first target characteristic value is calculated based on the preset temperature warning duration and the change parameter, and the second target characteristic value is input into the temperature adjustment analysis model to generate a target score;

[0010] Determine whether the target score is greater than the preset warning threshold. If so, adjust the temperature of the target transformer based on the preset temperature adjustment method.

[0011] Preferably, the temperature regulation analysis model further includes:

[0012] The temperature regulation analysis model is constructed based on the following form:

[0013]

[0014] In the formula, is the comprehensive score of the temperature regulation analysis model output, For multiple input features, For multiple weights that need to be trained, is the number of target parameters;

[0015] Loss function of the temperature regulation analysis model for:

[0016]

[0017] In the formula, is the scoring threshold, is the first The comprehensive score of the group data, The temperature is adjusted. The comprehensive score of the group data, , is the loss adjustment factor;

[0018] During the training process of the temperature regulation analysis model, after the weights and scoring thresholds of the temperature regulation analysis model are initialized, the model is trained based on the training data set, and the gradient descent algorithm is used to optimize the multiple weights of the temperature regulation analysis model. After multiple iterations are performed until the preset convergence target is reached, the training of the temperature regulation analysis model is completed.

[0019] Preferably, the temperature adjustment correlation factor of each parameter in each set of temperature adjustment reference data is calculated according to the temperature adjustment characteristic value and the characteristic change value, including:

[0020] The temperature regulation correlation factor is calculated using the following formula:

[0021]

[0022] In the formula, is the temperature regulation correlation factor, is the temperature regulation characteristic value, is the characteristic change value, is a constant term.

[0023] Preferably, a plurality of target parameters are determined according to a plurality of temperature adjustment correlation factors of each parameter, and a training data set is extracted from a plurality of sets of temperature adjustment reference data according to the plurality of target parameters, including:

[0024] Calculate the mean of the temperature adjustment correlation factors corresponding to each parameter in the multiple sets of temperature adjustment reference data, and select the parameters whose mean of the multiple temperature adjustment correlation factors is greater than the preset correlation threshold based on the preset correlation threshold and record them as target parameters;

[0025] The first reference data corresponding to each target parameter in the temperature adjustment reference data is extracted, multiple groups of first reference data are segmented to obtain multiple groups of second reference data, the mean value of each target parameter in the multiple groups of second reference data that do not contain temperature adjustment behavior is extracted, and multiple input features of the target parameter in a state where temperature adjustment is not performed are obtained, the actual value of each target parameter at the corresponding moment of the temperature adjustment behavior in the second reference data that contains the temperature adjustment behavior is extracted, and multiple input features of the target parameter in a state where temperature adjustment has been performed are obtained, and a training data set is constructed that contains multiple input features of the target parameter in a state where temperature adjustment has not been performed and multiple input features of the target parameter in a state where temperature adjustment has been performed.

[0026] Preferably, the second target characteristic value corresponding to each first target characteristic value is calculated based on the preset temperature warning duration and the variation parameter, and the second target characteristic value is input into the temperature adjustment analysis model to generate a target score, including:

[0027] According to the change parameters of multiple first target characteristic values ​​and the preset temperature warning duration, the growth value of each first target characteristic value within the preset temperature warning duration is calculated respectively, and the growth value of each first target characteristic value within the preset temperature warning duration is added to the corresponding first target characteristic value to obtain the second target characteristic value corresponding to each first target characteristic value. The temperature regulation analysis model analyzes the multiple second target characteristic values ​​and generates target scores corresponding to the multiple second target characteristic values.

[0028] Preferably, multiple groups of temperature adjustment reference data are extracted from the historical device status data according to the historical temperature adjustment data of the target transformer, including:

[0029] According to a plurality of temperature adjustment behaviors recorded in the historical temperature adjustment data of the target transformer, data within a preset time range before each temperature adjustment behavior is extracted to obtain a plurality of groups of temperature adjustment reference data.

[0030] A second aspect of the present invention provides an Internet of Things-based power equipment control system, which is used to implement the above-mentioned Internet of Things-based power equipment control method, including:

[0031] A data acquisition module is used to acquire historical device status data of a target transformer collected based on an IoT device, and extract multiple sets of temperature adjustment reference data from the historical device status data according to historical temperature adjustment data of the target transformer;

[0032] A correlation analysis module is used to extract temperature regulation characteristic values ​​and characteristic change values ​​of multiple parameters in each set of temperature regulation reference data, and calculate the temperature regulation correlation factor of each parameter in each set of temperature regulation reference data according to the temperature regulation characteristic values ​​and characteristic change values;

[0033] A model training module is used to determine multiple target parameters according to multiple temperature regulation correlation factors of each parameter, extract a training data set from multiple sets of temperature regulation reference data according to the multiple target parameters, obtain a temperature regulation analysis model through training the training data set, and determine a preset warning threshold;

[0034] A target data analysis module, used to collect target device status data of a target transformer based on an Internet of Things device, and extract first target characteristic values ​​of multiple target parameters in the target device status data and a change parameter of each first target characteristic value;

[0035] A state analysis module, used to calculate a second target characteristic value corresponding to each first target characteristic value based on a preset temperature warning duration and a change parameter, and input the second target characteristic value into a temperature regulation analysis model to generate a target score;

[0036] The equipment control module is used to determine whether the target score is greater than a preset warning threshold. If so, the temperature of the target transformer is adjusted based on a preset temperature adjustment method.

[0037] Preferably, for the state analysis module, a second target characteristic value corresponding to each first target characteristic value is calculated based on a preset temperature warning duration and a change parameter, and the second target characteristic value is input into a temperature regulation analysis model to generate a target score, including:

[0038] According to the change parameters of multiple first target characteristic values ​​and the preset temperature warning duration, the growth value of each first target characteristic value within the preset temperature warning duration is calculated respectively, and the growth value of each first target characteristic value within the preset temperature warning duration is added to the corresponding first target characteristic value to obtain the second target characteristic value corresponding to each first target characteristic value. The temperature regulation analysis model analyzes the multiple second target characteristic values ​​and generates target scores corresponding to the multiple second target characteristic values.

[0039] The present invention has the following beneficial effects:

[0040] The present invention realizes all-round monitoring by comprehensively analyzing multiple operating parameters of the transformer and combining characteristic values ​​and change rates. It also uses the trained temperature regulation analysis model to intelligently determine whether temperature regulation is needed. By predicting the future state change trend of the transformer, the present invention can identify potential overheating risks in advance, facilitate corresponding measures to be taken in advance, effectively extend the life of the transformer, reduce the occurrence of faults, improve the temperature regulation efficiency and the timeliness of early warning, and ensure the safe and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A schematic flow chart of a method for controlling power equipment based on the Internet of Things provided in an embodiment of the present invention.

[0042] Figure 2 A schematic structural diagram of an Internet of Things-based power equipment control system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to make those skilled in the art better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0044] See also Figure 1 , an embodiment of the present invention provides a power equipment control method based on the Internet of Things, comprising the following steps:

[0045] Step S1, acquiring historical device status data of a target transformer collected based on an Internet of Things device, and extracting multiple groups of temperature adjustment reference data from the historical device status data according to historical temperature adjustment data of the target transformer.

[0046] Among them, the historical equipment status data of the target transformer collected by the Internet of Things devices (such as temperature sensors, current sensors, oil temperature sensors, etc.) include but are not limited to the winding temperature, load current, ambient temperature, oil temperature and other parameters of the target transformer in different time periods. For these data, by analyzing the multiple temperature adjustment behaviors recorded in the historical temperature adjustment behavior data of the target transformer, multiple groups of temperature adjustment reference data are extracted according to the preset time range. The temperature adjustment behavior specifically involves the control behavior executed because the winding temperature is too high to reduce the winding temperature of the transformer. The time point when each temperature adjustment behavior occurs and the operating status of the transformer in a period of time before that, that is, within the preset time range, are identified, such as the status data within the first 10 minutes, thereby extracting multiple groups of temperature adjustment reference data associated with each temperature adjustment behavior, which are used to analyze the relevant change trends of various parameters before temperature adjustment of the target transformer.

[0047] Step S2: extracting temperature adjustment characteristic values ​​and characteristic change values ​​of multiple parameters in each set of temperature adjustment reference data, and calculating the temperature adjustment correlation factor of each parameter in each set of temperature adjustment reference data according to the temperature adjustment characteristic values ​​and characteristic change values.

[0048] Among them, for each set of temperature regulation reference data, the temperature regulation characteristic value of each parameter and the characteristic change value of these parameters are extracted therefrom. The temperature regulation characteristic value can be the instantaneous value of parameters such as load, output power, winding temperature, output current, etc. after the temperature regulation behavior occurs. Specifically, after the power system detects that one or more parameters have reached the critical value, it generates corresponding instructions to perform temperature regulation of the transformer. The instantaneous value of each parameter at the moment when it is determined that temperature regulation is required can be used as the temperature regulation characteristic value, and the characteristic change value indicates the change rate of these parameters in a period of time before the regulation behavior occurs, such as the change rate of each parameter in the first 10 minutes. Based on the temperature regulation characteristic value and the characteristic change value, the temperature regulation correlation factor of each parameter in each set of reference data is calculated. The temperature regulation correlation factor reflects the importance or correlation of each parameter in the temperature regulation behavior. By analyzing and calculating each regulation behavior, it is determined which parameters have a greater impact on temperature regulation.

[0049] In this embodiment, based on the temperature adjustment characteristic value and the characteristic change value, the temperature adjustment correlation factor is calculated using the following formula:

[0050]

[0051] In the formula, is the temperature regulation correlation factor, which is used to quantify the influence of a certain parameter on the temperature regulation behavior. is the temperature regulation characteristic value, which indicates the current state value of a key parameter during temperature regulation. is the characteristic change value, which indicates the change rate or trend of the parameter before the adjustment behavior occurs. is a constant term, and the temperature regulation correlation factor finally calculated has nonlinear response capability and can more sensitively capture those parameters that have a significant impact on the temperature regulation process.

[0052] Step S3, determining multiple target parameters according to multiple temperature adjustment correlation factors of each parameter, extracting a training data set from multiple groups of temperature adjustment reference data according to the multiple target parameters, obtaining a temperature adjustment analysis model through training the training data set and determining a preset warning threshold.

[0053] Among them, the multiple temperature adjustment correlation factors calculated in step S2 are analyzed and screened to determine multiple target parameters that have a greater impact on the temperature adjustment behavior. According to these target parameters, a training data set is extracted from multiple groups of temperature adjustment reference data, and a temperature adjustment analysis model is trained using the extracted training data set. The temperature adjustment analysis model can be used to analyze the input characteristic parameters to evaluate the operating status of the transformer, specifically to generate a score for the transformer, which is used to characterize the possibility of temperature adjustment under the current state.

[0054] Step S4: collecting target device status data of the target transformer based on the Internet of Things device, and extracting first target characteristic values ​​of multiple target parameters in the target device status data and change parameters of each first target characteristic value.

[0055] Among them, in the actual operation of the target transformer, the target device status data of the target transformer can be collected in real time through the Internet of Things device, including the monitoring values ​​of various parameters. After determining the real-time value of the previously selected target parameter from these status data, the first target characteristic value of each target parameter is extracted, specifically the current instantaneous value, and the change parameter corresponding to the first target characteristic value in the previous period of time, specifically the rate of change of the parameter in this period of time, for example, the rate of change of the winding temperature represents the temperature rise per minute, and the rate of change of the load current represents the increase of the current per minute, etc.

[0056] Step S5: Calculate the second target characteristic value corresponding to each first target characteristic value based on the preset temperature warning duration and the change parameter, and input the second target characteristic value into the temperature adjustment analysis model to generate a target score.

[0057] Among them, the preset temperature warning time is the time required for early warning, such as 5 minutes. According to the change rate of the parameter in a period of time before the first target characteristic value of each parameter, in order to improve the prediction accuracy, those skilled in the art can select the change rate of the parameter in a shorter time according to actual needs, for example, the change rate of the parameter is calculated based on the data in the previous minute as a reference, and the growth value of each first target characteristic value in the preset temperature warning time is calculated according to the change parameters of multiple first target characteristic values ​​and the preset temperature warning time, and the growth value of each first target characteristic value in the preset temperature warning time is added to the corresponding first target characteristic value to obtain the second target characteristic value corresponding to each first target characteristic value. Exemplarily, the current growth rate of the winding temperature of the target transformer is 2°C / minute. Under the condition that the preset time required for early warning is 5 minutes, the growth value of the winding temperature is expected to be 10°C during this period. The calculated growth value is added to the current first target characteristic value to obtain the second target characteristic value, and the obtained second target characteristic value is used as input and substituted into the temperature regulation analysis model for analysis. The temperature regulation analysis model will calculate the corresponding target score based on these second target characteristic values.

[0058] Step S6: determine whether the target score is greater than a preset warning threshold, and if so, adjust the temperature of the target transformer based on a preset temperature adjustment method.

[0059] Among them, the target score output by the model is compared with the preset warning threshold. If the target score is greater than the preset warning threshold, it means that the state of the system may be abnormal after the preset temperature warning time. In such a state, there is a high probability that the winding temperature is too high, so as to prompt that temperature adjustment may be required. In this case, the target transformer is subjected to corresponding temperature adjustment operations based on the preset temperature adjustment method. For example, the cooling system is started, the load is reduced, and other preventive measures are taken to ensure the safe operation of the transformer and prevent the occurrence of faults. Those skilled in the art can select appropriate preventive measures to adjust the temperature of the target transformer according to actual needs to avoid excessively high winding temperature to affect the stability of equipment operation.

[0060] As one implementation manner, in step S3, multiple target parameters are determined according to multiple temperature adjustment correlation factors of each parameter, and a training data set is extracted from multiple groups of temperature adjustment reference data according to the multiple target parameters, specifically including:

[0061] The mean of the temperature adjustment correlation factors corresponding to each parameter in multiple sets of temperature adjustment reference data is calculated, and based on a preset correlation threshold, parameters whose mean of multiple temperature adjustment correlation factors is greater than the preset correlation threshold are screened out and recorded as target parameters.

[0062] Among them, the mean value of the temperature regulation correlation factor is used to reflect the average influence of different parameters on temperature regulation under different operating conditions. By presetting the correlation threshold, multiple parameters with greater influence are screened out from multiple parameters and recorded as target parameters to obtain multiple parameters with a strong correlation with the temperature regulation behavior.

[0063] For the extraction of training data sets, firstly extract the first reference data corresponding to each target parameter in the temperature adjustment reference data, specifically the data corresponding to multiple target parameters in the temperature adjustment reference data, and then perform data segmentation on multiple groups of first reference data to obtain multiple groups of second reference data. The role of segmentation is to subdivide the first reference data into smaller time windows or data under specific operating conditions, so that the state changes before and after the temperature adjustment behavior can be captured more finely. At the same time, the continuous data is discretized, and the mean of each target parameter in multiple groups of second reference data that do not contain temperature adjustment behavior is extracted to obtain multiple input features of the target parameter in the state without temperature adjustment, and the actual value of each target parameter in the second reference data containing temperature adjustment behavior at the corresponding time of temperature adjustment behavior is extracted. For example, for the data of the first 10 minutes of the temperature adjustment behavior, the analysis is performed at intervals of 0.5 minutes, and the mean of each target parameter in every 0.5 minutes is calculated. For the multiple groups of second reference data obtained by segmentation, the last group of data contains temperature adjustment behavior, and the remaining data represents the performance of the target parameter in the state without temperature adjustment.

[0064] Finally, a training data set is constructed that includes multiple input features of the target parameters without temperature adjustment and multiple input features of the target parameters with temperature adjustment. It contains the target parameter performance in various states, including both normal state features and abnormal (needing adjustment) state features. Through these comprehensive data, the model can learn how to judge the current operating state through input features.

[0065] As one implementation method, the temperature adjustment analysis model in step S3 is constructed based on the following form:

[0066]

[0067] In the formula, is the comprehensive score of the temperature regulation analysis model output, are multiple input features, representing the characteristic values ​​of multiple target parameters (such as winding temperature, load current, oil temperature, etc.). are multiple weights that need to be trained, reflecting the importance of each target parameter in the comprehensive score. is the number of target parameters, which indicates the dimension of the input features considered in the model. The core of the temperature regulation analysis model is the weighted linear combination model, and the comprehensive score of its output is used to evaluate the temperature regulation requirements of the transformer in the current state.

[0068] In order to optimize the temperature regulation analysis model so that it can accurately distinguish between states that require temperature regulation and states that do not require regulation in practical applications, the loss function of the temperature regulation analysis model is The design is as follows:

[0069]

[0070] In the formula, is the scoring threshold, which is used to distinguish between the state that requires temperature adjustment and the normal state. is the first The comprehensive score of the group data, The temperature is adjusted. The comprehensive score of the group data, , is the loss adjustment coefficient, which controls the contribution of the unadjusted state and the adjusted state to the overall loss in the loss function. The design goal of the above loss function is that for the state without temperature adjustment, that is, the multiple states before the temperature adjustment behavior occurs, if the score predicted by the model exceeds the threshold, it means that the model has generated a false positive, and the loss function will penalize the part exceeding the threshold, so that the model can reduce the occurrence of such false positives during the training process; for the state with temperature adjustment, that is, the state when the temperature adjustment behavior occurs, if the score predicted by the model is lower than the threshold, it means that the model has failed to correctly identify the state that needs to be adjusted, and the loss function will penalize the part below the threshold, so that the model can more accurately identify the state that needs to be adjusted during training.

[0071] During the training process of the temperature regulation analysis model, the weights and scoring thresholds of the temperature regulation analysis model are first initialized. The initial values ​​of these weights can be random or set based on the statistical characteristics of historical data. Similarly, the initial thresholds can be set based on the empirical values ​​in the historical data to ensure that the model has a reasonable benchmark in the initial stage.

[0072] In the process of model training based on the training data set, the gradient descent algorithm is used to optimize the multiple weights of the temperature regulation analysis model, so that the value of the loss function is gradually reduced. When the loss value tends to be stable, for example, the loss function converges or reaches the preset stop condition, the model training is completed. This process can be carried out through multiple iterations until the loss function converges to a stable value. The model adjusts the weight parameters and the scoring threshold in each iteration to gradually improve the prediction accuracy of the model. Through multiple iterations, the model can gradually learn the complex relationship between each input feature and accurately determine whether the current system state requires temperature regulation. The trained model can evaluate the operating status of the transformer in real time, and decide whether temperature regulation is needed based on the comparison result between the comprehensive score and the scoring threshold. In the actual application process, the predicted value after a period of time is determined based on the estimated actual value and growth value of different parameters, and then the operating status of the transformer under the predicted value is analyzed based on the model to determine whether temperature regulation is needed in this state, so as to take corresponding preventive measures in advance for possible over-temperature phenomena, thereby effectively ensuring the safe and stable operation of power equipment.

[0073] The embodiment of the present invention provides a power equipment control method based on the Internet of Things. By comprehensively analyzing multiple operating parameters of the transformer, combined with the characteristic value of the current state and the change rate of the parameter, the transformer state is fully monitored; the temperature adjustment analysis model obtained by training the training data set can evaluate the operating state of the transformer in real time, and intelligently judge whether temperature adjustment is needed according to the comprehensive score and the preset score threshold; the second target characteristic value is calculated by the preset temperature warning time and the change parameter to characterize the temperature change trend in the future period; the second target characteristic value is analyzed by the temperature adjustment analysis model to generate a target score, which can evaluate the operating state of the transformer in real time, and compare the comprehensive score with the preset score threshold, identify the potential overheating risk in advance, and intelligently judge whether temperature adjustment is needed, so as to identify the potential overheating risk in advance and take appropriate adjustment measures. This prediction mechanism effectively reduces the response time of the system and improves the timeliness of the early warning. Compared with the traditional simple threshold-based adjustment method, the method can effectively extend the service life of the transformer, reduce the probability of failure, significantly improve the efficiency and accuracy of transformer temperature adjustment, and ensure the safe and stable operation of the power system.

[0074] See also Figure 2 The present invention provides an Internet of Things-based power equipment control system, which is used to implement the above-mentioned Internet of Things-based power equipment control method, including:

[0075] A data acquisition module is used to acquire historical device status data of a target transformer collected based on an IoT device, and extract multiple sets of temperature adjustment reference data from the historical device status data according to historical temperature adjustment data of the target transformer;

[0076] A correlation analysis module is used to extract temperature regulation characteristic values ​​and characteristic change values ​​of multiple parameters in each set of temperature regulation reference data, and calculate the temperature regulation correlation factor of each parameter in each set of temperature regulation reference data according to the temperature regulation characteristic values ​​and characteristic change values;

[0077] A model training module is used to determine multiple target parameters according to multiple temperature regulation correlation factors of each parameter, extract a training data set from multiple sets of temperature regulation reference data according to the multiple target parameters, obtain a temperature regulation analysis model through training the training data set, and determine a preset warning threshold;

[0078] Wherein, a plurality of target parameters are determined according to a plurality of temperature adjustment correlation factors of each parameter, and a training data set is extracted from a plurality of sets of temperature adjustment reference data according to the plurality of target parameters, including:

[0079] Calculate the mean of the temperature adjustment correlation factors corresponding to each parameter in the multiple sets of temperature adjustment reference data, and select the parameters whose mean of the multiple temperature adjustment correlation factors is greater than the preset correlation threshold based on the preset correlation threshold and record them as target parameters;

[0080] The first reference data corresponding to each target parameter in the temperature adjustment reference data is extracted, multiple groups of first reference data are segmented to obtain multiple groups of second reference data, the mean value of each target parameter in the multiple groups of second reference data that do not contain temperature adjustment behavior is extracted, and multiple input features of the target parameter in a state where temperature adjustment is not performed are obtained, the actual value of each target parameter at the corresponding moment of the temperature adjustment behavior in the second reference data that contains the temperature adjustment behavior is extracted, and multiple input features of the target parameter in a state where temperature adjustment has been performed are obtained, and a training data set is constructed that contains multiple input features of the target parameter in a state where temperature adjustment has not been performed and multiple input features of the target parameter in a state where temperature adjustment has been performed.

[0081] For temperature regulation analysis models, also included:

[0082] The temperature regulation analysis model is constructed based on the following form:

[0083]

[0084] In the formula, is the comprehensive score of the temperature regulation analysis model output, For multiple input features, For multiple weights that need to be trained, is the number of target parameters;

[0085] Loss function of the temperature regulation analysis model for:

[0086]

[0087] In the formula, is the scoring threshold, is the first The comprehensive score of the group data, The temperature is adjusted. The comprehensive score of the group data, , is the loss adjustment factor;

[0088] During the training process of the temperature regulation analysis model, after the weights and scoring thresholds of the temperature regulation analysis model are initialized, the model is trained based on the training data set, and the gradient descent algorithm is used to optimize the multiple weights of the temperature regulation analysis model. After multiple iterations are performed until the preset convergence target is reached, the training of the temperature regulation analysis model is completed.

[0089] A target data analysis module, used to collect target device status data of a target transformer based on an Internet of Things device, and extract first target characteristic values ​​of multiple target parameters in the target device status data and a change parameter of each first target characteristic value;

[0090] A state analysis module, used to calculate a second target characteristic value corresponding to each first target characteristic value based on a preset temperature warning duration and a change parameter, and input the second target characteristic value into a temperature regulation analysis model to generate a target score;

[0091] The second target characteristic value corresponding to each first target characteristic value is calculated based on the preset temperature warning duration and the change parameter, and the second target characteristic value is input into the temperature adjustment analysis model to generate a target score, including:

[0092] According to the change parameters of multiple first target characteristic values ​​and the preset temperature warning duration, the growth value of each first target characteristic value within the preset temperature warning duration is calculated respectively, and the growth value of each first target characteristic value within the preset temperature warning duration is added to the corresponding first target characteristic value to obtain the second target characteristic value corresponding to each first target characteristic value. The temperature regulation analysis model analyzes the multiple second target characteristic values ​​and generates target scores corresponding to the multiple second target characteristic values.

[0093] The equipment control module is used to determine whether the target score is greater than a preset warning threshold. If so, the temperature of the target transformer is adjusted based on a preset temperature adjustment method.

[0094] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A power equipment control method based on the Internet of Things, characterized in that: include: Acquire historical device status data of the target transformer collected by the IoT device, and extract multiple sets of temperature adjustment reference data from the historical device status data according to the historical temperature adjustment data of the target transformer; Extracting temperature regulation characteristic values ​​and characteristic change values ​​of multiple parameters in each set of temperature regulation reference data, and calculating the temperature regulation correlation factor of each parameter in each set of temperature regulation reference data according to the temperature regulation characteristic values ​​and characteristic change values; Determine multiple target parameters according to multiple temperature regulation correlation factors of each parameter, extract training data sets from multiple sets of temperature regulation reference data according to the multiple target parameters, obtain a temperature regulation analysis model through training the training data sets and determine a preset warning threshold; Collect target device status data of the target transformer based on the Internet of Things device, and extract first target characteristic values ​​of multiple target parameters in the target device status data and change parameters of each first target characteristic value; The second target characteristic value corresponding to each first target characteristic value is calculated based on the preset temperature warning duration and the change parameter, and the second target characteristic value is input into the temperature adjustment analysis model to generate a target score; Determine whether the target score is greater than a preset warning threshold, and if so, adjust the temperature of the target transformer based on a preset temperature adjustment method; For temperature regulation analysis models, also included: The temperature regulation analysis model is constructed based on the following form: In the formula, is the comprehensive score of the temperature regulation analysis model output, For multiple input features, For multiple weights that need to be trained, is the number of target parameters; Loss function of the temperature regulation analysis model for: In the formula, is the scoring threshold, is the first The comprehensive score of the group data, The temperature is adjusted. The comprehensive score of the group data, , is the loss adjustment factor; During the training of the temperature regulation analysis model, after the weights and scoring thresholds of the temperature regulation analysis model are initialized, the model is trained based on the training data set, and the gradient descent algorithm is used to optimize multiple weights of the temperature regulation analysis model. After multiple iterations are performed until the preset convergence target is reached, the training of the temperature regulation analysis model is completed; A plurality of target parameters are determined according to a plurality of temperature regulation correlation factors of each parameter, and a training data set is extracted from a plurality of sets of temperature regulation reference data according to the plurality of target parameters, including: Calculate the mean of the temperature adjustment correlation factors corresponding to each parameter in the multiple sets of temperature adjustment reference data, and select the parameters whose mean of the multiple temperature adjustment correlation factors is greater than the preset correlation threshold based on the preset correlation threshold and record them as target parameters; The first reference data corresponding to each target parameter in the temperature adjustment reference data is extracted, multiple groups of first reference data are segmented to obtain multiple groups of second reference data, the mean value of each target parameter in the multiple groups of second reference data that do not contain temperature adjustment behavior is extracted, and multiple input features of the target parameter in a state where temperature adjustment is not performed are obtained, the actual value of each target parameter at the corresponding moment of the temperature adjustment behavior in the second reference data that contains the temperature adjustment behavior is extracted, and multiple input features of the target parameter in a state where temperature adjustment has been performed are obtained, and a training data set is constructed that contains multiple input features of the target parameter in a state where temperature adjustment has not been performed and multiple input features of the target parameter in a state where temperature adjustment has been performed.

2. The method for controlling power equipment based on the Internet of Things according to claim 1, characterized in that: The temperature regulation correlation factor of each parameter in each set of temperature regulation reference data is calculated based on the temperature regulation characteristic value and characteristic change value, including: The temperature regulation correlation factor is calculated using the following formula: In the formula, is the temperature regulation correlation factor, is the temperature regulation characteristic value, is the characteristic change value, is a constant term.

3. The method for controlling power equipment based on the Internet of Things according to claim 1, characterized in that: The second target characteristic value corresponding to each first target characteristic value is calculated based on the preset temperature warning duration and the change parameter, and the second target characteristic value is input into the temperature adjustment analysis model to generate a target score, including: According to the change parameters of multiple first target characteristic values ​​and the preset temperature warning duration, the growth value of each first target characteristic value within the preset temperature warning duration is calculated respectively, and the growth value of each first target characteristic value within the preset temperature warning duration is added to the corresponding first target characteristic value to obtain the second target characteristic value corresponding to each first target characteristic value. The temperature regulation analysis model analyzes the multiple second target characteristic values ​​and generates target scores corresponding to the multiple second target characteristic values.

4. The method for controlling power equipment based on the Internet of Things according to claim 1, characterized in that: According to the historical temperature adjustment data of the target transformer, multiple sets of temperature adjustment reference data are extracted from the historical equipment status data, including: According to a plurality of temperature adjustment behaviors recorded in the historical temperature adjustment data of the target transformer, data within a preset time range before each temperature adjustment behavior is extracted to obtain a plurality of groups of temperature adjustment reference data.

5. An Internet of Things-based power equipment control system, characterized in that: The system is used to implement the power equipment control method based on the Internet of Things as described in any one of claims 1 to 4 above, comprising: A data acquisition module is used to acquire historical device status data of a target transformer collected based on an IoT device, and extract multiple sets of temperature adjustment reference data from the historical device status data according to historical temperature adjustment data of the target transformer; A correlation analysis module is used to extract temperature regulation characteristic values ​​and characteristic change values ​​of multiple parameters in each set of temperature regulation reference data, and calculate the temperature regulation correlation factor of each parameter in each set of temperature regulation reference data according to the temperature regulation characteristic values ​​and characteristic change values; A model training module is used to determine multiple target parameters according to multiple temperature regulation correlation factors of each parameter, extract a training data set from multiple sets of temperature regulation reference data according to the multiple target parameters, obtain a temperature regulation analysis model through training the training data set, and determine a preset warning threshold; A target data analysis module, used to collect target device status data of a target transformer based on an Internet of Things device, and extract first target characteristic values ​​of multiple target parameters in the target device status data and a change parameter of each first target characteristic value; A state analysis module, used to calculate a second target characteristic value corresponding to each first target characteristic value based on a preset temperature warning duration and a change parameter, and input the second target characteristic value into a temperature regulation analysis model to generate a target score; The equipment control module is used to determine whether the target score is greater than a preset warning threshold. If so, the temperature of the target transformer is adjusted based on a preset temperature adjustment method.

6. The power equipment control system based on the Internet of Things according to claim 5 is characterized in that: For the state analysis module, the second target characteristic value corresponding to each first target characteristic value is calculated based on the preset temperature warning duration and the change parameter, and the second target characteristic value is input into the temperature adjustment analysis model to generate a target score, including: According to the change parameters of multiple first target characteristic values ​​and the preset temperature warning duration, the growth value of each first target characteristic value within the preset temperature warning duration is calculated respectively, and the growth value of each first target characteristic value within the preset temperature warning duration is added to the corresponding first target characteristic value to obtain the second target characteristic value corresponding to each first target characteristic value. The temperature regulation analysis model analyzes the multiple second target characteristic values ​​and generates target scores corresponding to the multiple second target characteristic values.

Images