A transformer temperature control system and method

Abstract

The present invention discloses a transformer temperature control system and method, which specifically relates to the field of dry-type transformer temperature control, and is used to solve the problems of fan speed optimization and temperature precision management. The method predicts the short-term temperature trend of the dry-type transformer early, identifies the similarity between the current temperature curve and historical data, draws on and verifies effective temperature control strategies, reduces calculations and experiments, and improves control efficiency. The similarity between the temperature change pattern and the historical pattern is analyzed to ensure accurate decisions are made during the temperature control process. When the similarity is high, the historical pattern is used for precise control. When the similarity is low, dynamic adjustment is relied on real-time data. Through comprehensive analysis of the periodic symmetry index and the periodic amplitude change rate, the resonant fluctuation control index is calculated to determine whether the temperature curve is suitable for peak shaving and valley filling, optimize fan speed regulation, reduce frequent speed regulation and mechanical wear, and ensure that the dry-type transformer is stable within a reasonable temperature range to ensure stable operation.

Description

Technical Field

[0001] The present invention relates to the field of dry-type transformer temperature control, and more specifically, to a transformer temperature control system and method. Background Art

[0002] During the operation of dry-type transformers, temperature management is a key factor in ensuring equipment safety and extending service life. Dry-type transformers rely on air for cooling, and temperature changes directly affect their operating performance. Traditional temperature control methods mainly rely on simple real-time monitoring and fan speed control strategies. However, this method is often limited in effect when dealing with complex temperature changes, and it is easy to cause the fan to start and stop frequently, increase mechanical wear, and shorten the service life of the equipment. In addition, traditional control methods lack in-depth analysis of historical data and cannot effectively utilize past operating experience, resulting in insufficient accuracy and efficiency of the control strategy, making it difficult to cope with complex temperature changes under different loads and environmental conditions. In this context, how to optimize the fan speed control strategy through effective prediction and analysis technology, achieve accurate temperature management and stable system operation, has become a technical problem that needs to be solved urgently.

[0003] In order to solve the above problems, a technical solution is now provided. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiment of the present invention provides a transformer temperature control system and method, which accurately predicts the short-term temperature trend of the dry-type transformer in the early stage, constructs the overall temperature change trend curve by integrating the data of each sensor, and performs similarity analysis, accurately identifies the similarity between the current temperature curve and the historical data, and draws on the temperature control strategy that has been verified to be effective in history to reduce unnecessary calculations and experiments, thereby improving the control efficiency. By analyzing the similarity between the current temperature change mode and the historical mode, it is possible to make accurate decisions in the temperature control process to avoid misjudgment. When the similarity is high, the historical mode is directly used for precise control. When the similarity is low, it relies on real-time data for dynamic adjustment to ensure the accuracy of temperature management and the stability of the system. Through the comprehensive analysis of the periodic symmetry index and the periodic amplitude change rate, the resonant fluctuation control index is calculated to accurately judge whether the fluctuation characteristics of the temperature curve are suitable for executing the peak shaving and valley filling strategy, thereby optimizing the fan speed control strategy, smoothing temperature fluctuations, reducing frequent fan speed regulation, reducing mechanical wear, and extending the service life of the fan. The real-time monitoring and feedback adjustment mechanism ensures that the temperature is controlled within a reasonable range, ensuring that the dry-type transformer maintains the stable operation of the system while optimizing fan management, and improving the overall operating efficiency, so as to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] S1, real-time collection of transformer temperature data, and storage after filtering;

[0007] S2, generates short-term temperature trend curve based on grayscale prediction model and matches historical data for similarity analysis;

[0008] S3, after identifying the similarity of the temperature curves, determine whether to implement the peak shaving and valley filling strategy, adjust the fan speed switching point, and smooth the temperature fluctuation;

[0009] S4: If the answer is yes, the fan speed is adjusted in real time according to the peak shaving and valley filling strategy; otherwise, the speed is adjusted in the traditional mode.

[0010] In a preferred embodiment, step S2 includes the following contents:

[0011] S2.1, first of all, for the internal layout of the transformer A temperature sensor collects temperature data. ,in Indicates the sensor number, Represents a time point, performs first-order difference processing, and calculates the temperature increment at each time point , remove the linear trend in the original time series;

[0012] S2.2, based on the differential data, the grayscale prediction model is applied, and the improved form of the GM(1,1) model is used for prediction; the construction of the grayscale model is based on the cumulative generation sequence, and the smooth data sequence is generated by accumulation; the cumulative generation sequence is calculated : ;

[0013] The prediction model based on the cumulative sequence is: ;

[0014] in, and are model parameters, estimated by the least squares method; by solving this differential equation, we get the solution of the cumulative generated sequence: ;

[0015] Then, the accumulated sequence is restored to obtain the predicted value of the temperature increment : ;

[0016] Finally, the original temperature series is restored by accumulation to obtain the predicted value of short-term temperature trend. : ;

[0017] Finally, a short-term temperature trend prediction curve for each sensor is generated;

[0018] S2.3, the predicted temperature values ​​of all sensors Summarize and construct the overall temperature trend curve of the transformer ;

[0019] S2.4, use the multidimensional dynamic time warping algorithm to perform similarity analysis; suppose the historical data set is , the current prediction data set is ;

[0020] The overall objective function of similarity calculation is: ;

[0021] : Path smoothness penalty coefficient, used to adjust the path continuity in similarity analysis;

[0022] : The distance between path nodes, controlling the path smoothness of similarity analysis.

[0023] In a preferred embodiment, S2.5, the similarity score calculated by MDTW is used to determine whether the current temperature trend is highly similar to a certain section in the historical data; if the similarity score is less than the corresponding threshold, it means that the current temperature trend is highly similar to a certain pattern in the historical data; at this time, it is considered that the current temperature change trend is very close to what happened in history and is marked as a short-term temperature trend curve.

[0024] In a preferred embodiment, step S3 includes the following contents:

[0025] The logic for obtaining the periodic symmetry index is:

[0026] a1, short-term temperature trend curve Perform Fourier transform to identify the dominant frequency of the short-term temperature trend curve;

[0027] Determine the dominant frequency by analyzing the amplitude of the frequency domain signal , that is, the frequency corresponding to the maximum amplitude in the frequency domain signal;

[0028] a2. After determining the dominant frequency, extract the main period short-term temperature trend curve corresponding to this frequency from the short-term temperature trend curve ; Reconstruct the main periodic component using inverse Fourier transform;

[0029] a3, evaluate the symmetry within a cycle through the short-term temperature trend curve of the main cycle; The starting point is , calculate the symmetry difference of the temperature fluctuation within the period; the symmetry difference is the integral of the difference between the temperature peak and the corresponding trough within the period: ;

[0030] in:

[0031] It represents the difference in symmetry within the cycle and measures the symmetry between the peaks and troughs;

[0032] is the cycle start time;

[0033] a4, calculate the periodic symmetry index based on symmetry differences ; The symmetry difference of each period segment is normalized to obtain the periodic symmetry index of the entire short-term temperature trend curve: ;

[0034] in:

[0035] is the total number of cycles;

[0036] is the value of the maximum symmetry difference in all periodic segments, used for normalization;

[0037] For the The starting time of a cycle.

[0038] In a preferred embodiment, the acquisition logic of the period amplitude change rate is:

[0039] b1, first, according to step a1 short-term temperature trend curve The dominant frequency and its corresponding cycle :

[0040] b2. After determining the main cycle, segment the short-term temperature trend curve and identify the peak amplitude in each cycle segment according to each cycle. and trough amplitude ;

[0041] b3, calculate the amplitude difference between the peak amplitude and the trough amplitude in each cycle , the amplitude difference indicates the fluctuation amplitude of the short-term temperature trend curve in the corresponding period ;

[0042] b4, then, analyze the amplitude change trend of each cycle and calculate the cycle amplitude change rate The purpose is to quantify the amplitude change of the short-term temperature trend curve between each cycle; the cycle amplitude change rate is calculated as the speed of change of the amplitude difference between each cycle: .

[0043] In a preferred embodiment, the periodic symmetry index and the periodic amplitude change rate are comprehensively analyzed and calculated to obtain a resonance fluctuation control index; the resonance fluctuation control index is compared with a recognition threshold;

[0044] If the resonant fluctuation control index is greater than or equal to the identification threshold, it means that the fluctuation characteristics of the short-term temperature trend curve are suitable for executing the peak shaving and valley filling strategy, generating an actionable signal;

[0045] If the resonant fluctuation control index is less than the identification threshold, it means that the fluctuation characteristics of the short-term temperature trend curve are not suitable for executing the peak shaving and valley filling strategy, and an inoperable signal is generated.

[0046] In a preferred embodiment, the periodic symmetry index and the periodic amplitude change rate are comprehensively analyzed and calculated to obtain a resonance fluctuation control index; the resonance fluctuation control index is compared with a recognition threshold;

[0047] If the resonant fluctuation control index is greater than or equal to the identification threshold, it means that the fluctuation characteristics of the short-term temperature trend curve are suitable for executing the peak shaving and valley filling strategy, generating an actionable signal;

[0048] If the resonant fluctuation control index is less than the identification threshold, it means that the fluctuation characteristics of the short-term temperature trend curve are not suitable for executing the peak shaving and valley filling strategy, and an inoperable signal is generated.

[0049] In a preferred embodiment, after obtaining the operable signal, c1 analyzes the amplitude difference change trend between each cycle based on the amplitude difference calculated in step b3, and calculates the amplitude change rate : ;

[0050] in: Indicates The cycle and The rate of change of amplitude between cycles;

[0051] c2, based on the recognition results of the peaks and troughs and the amplitude change trend, adjust the switching point of the fan speed , adjust the fan speed switching point by the following formula: ;

[0052] in:

[0053] The fan speed switching time after adjustment;

[0054] To adjust the coefficient, set it according to actual needs to control the adjustment range of the switching point.

[0055] In a preferred embodiment, step S4 includes the following contents:

[0056] If an inoperable signal is received, the system will immediately switch to the traditional speed control mode. In the traditional mode, the fan speed will be adjusted directly based on the real-time monitored temperature data without considering the smoothing effect of the peak shaving and valley filling strategy on temperature fluctuations.

[0057] A transformer temperature control system, comprising: a data acquisition module, a trend prediction module, a similarity recognition module and a fan adjustment module;

[0058] The data acquisition module collects transformer temperature data in real time, stores it after filtering, and the processed data will be passed to the trend prediction module;

[0059] The trend prediction module generates a short-term temperature trend curve based on the grayscale prediction model, matches historical data for similarity analysis, and sends the similarity analysis results to the similarity recognition module;

[0060] After the similarity recognition module recognizes the similarity of the temperature curve, it determines whether to implement the peak shaving and valley filling strategy, adjusts the fan speed switching point, smoothes the temperature fluctuation, and sends the judgment result to the fan adjustment module;

[0061] If the fan adjustment module determines that it is, it will adjust the fan speed in real time according to the peak shaving and valley filling strategy, otherwise it will adjust the speed according to the traditional mode.

[0062] The technical effects and advantages of a transformer temperature control system and method of the present invention are as follows:

[0063] 1. Through the processing of step S2, the short-term temperature trend of the dry-type transformer can be accurately predicted at an early stage, and the predicted data of each sensor can be integrated to construct a trend curve reflecting the overall temperature change of the transformer. The similarity analysis performed by the MDTW algorithm can accurately identify the similarity between the current temperature curve and the historical data, and draw on the temperature control strategies that have been proven effective in history, such as the peak-shaving and valley-filling strategies, so as to reduce unnecessary calculations and experiments and improve the control efficiency. More importantly, by analyzing the similarity between the current temperature change pattern and the historical pattern, it can help the identification and analysis of subsequent temperature changes based on mature, controllable and recognizable patterns, ensuring that accurate decisions can be made during the temperature control process to avoid misjudgment. When the similarity score is high, the historical pattern is directly used for control to achieve precise control; when the similarity is low, avoid blindly referring to inapplicable historical patterns, rely on real-time data for dynamic adjustment, ensure the accuracy of temperature management and the stability of the system, and optimize the fan speed control strategy, extend the fan service life, and improve the overall operation efficiency of the system on the premise of ensuring the safe operation of the dry-type transformer.

[0064] 2. By executing step S3, the temperature curve suitable for implementing the peak shaving and valley filling strategy can be accurately identified and screened to ensure that peak shaving and valley filling is implemented only under the most appropriate conditions. Through the comprehensive analysis of the periodic symmetry index and the periodic amplitude change rate, the calculated resonance fluctuation control index is compared with the identification threshold, which can accurately determine whether the fluctuation characteristics of the temperature curve are suitable for peak shaving and valley filling, thereby optimizing the fan speed regulation strategy. By adjusting the fan speed switching point, the temperature fluctuation is smoothed, the frequent speed regulation of the fan is effectively reduced, the mechanical wear is reduced, and the service life of the fan is extended. At the same time, the real-time monitoring and feedback adjustment mechanism ensures that the temperature is always controlled within a reasonable range, ensuring that the dry-type transformer maintains stable operation and efficient temperature management of the system while optimizing fan management. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 A schematic flow chart of a transformer temperature control method according to the present invention;

[0066] Figure 2 The present invention is a schematic flow chart of a transformer temperature control system. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] Example 1

[0069] Figure 1 The present invention provides a transformer temperature control method, comprising:

[0070] S1, real-time collection of transformer temperature data, and storage after filtering;

[0071] S2, generates short-term temperature trend curve based on grayscale prediction model and matches historical data for similarity analysis;

[0072] S3, after identifying the similarity of the temperature curves, determine whether to implement the peak shaving and valley filling strategy, adjust the fan speed switching point, and smooth the temperature fluctuation;

[0073] S4: If the answer is yes, the fan speed is adjusted in real time according to the peak shaving and valley filling strategy; otherwise, the speed is adjusted in the traditional mode.

[0074] During the operation of dry-type transformers, temperature management is a key factor in ensuring equipment safety and extending service life. Due to its structural characteristics, dry-type transformers do not use oil as a cooling medium, but rely on air cooling. Therefore, under high load or high ambient temperature, the temperature inside the transformer is prone to rise. If heat dissipation is not carried out in a timely and effective manner, it may cause the equipment to overheat, thereby affecting its performance and safety.

[0075] In order to control the temperature of the transformer, the system usually installs multiple temperature sensors to monitor the temperature changes inside the transformer in real time. These sensors are arranged in key parts of the transformer, such as near the windings and core, to capture the most representative temperature data. When the sensor detects a temperature rise, the system automatically starts the fan to dissipate heat. The fan accelerates the flow of surrounding air to take away heat and reduce the temperature of the transformer.

[0076] During the temperature control process, in order to ensure the efficient operation of the fan and the temperature stability of the dry-type transformer, the system will dynamically adjust the fan speed based on real-time temperature data and historical temperature change patterns. For example, when it is detected that the temperature change pattern is highly similar to the pattern that has appeared in history, the system can refer to the previous successful fan speed adjustment strategies, such as the peak-shaving and valley-filling strategy, and adjust the fan speed switching point appropriately to smooth temperature fluctuations, reduce the frequent start-up and shutdown of the fan, and thus extend the service life of the fan.

[0077] In addition, the system will rely on real-time temperature data to accurately adjust the speed when the temperature fluctuates greatly or the environmental conditions change drastically, ensuring that the dry-type transformer always operates within a safe temperature range to avoid equipment failure or performance degradation due to overheating. Through these refined temperature management measures, the fan can effectively dissipate heat, ensure the long-term stable operation of the dry-type transformer, and improve the safety and efficiency of the overall system.

[0078] Step S1 includes the following contents:

[0079] Deploy high-precision temperature sensors at key locations of dry-type transformers, including but not limited to windings, cores, heat dissipation channels and other areas prone to temperature changes. Temperature sensors must have high response speed and anti-interference capabilities to ensure that real-time collected temperature data accurately reflects the actual operating status of the transformer. The temperature data collection process is managed uniformly by the data acquisition system to ensure that the data of each sensor is recorded synchronously and to avoid time differences caused by collection delays. After the temperature data is collected, it is transmitted to the central data processing unit through the data transmission system. The transmission process should have a data verification mechanism to prevent data distortion caused by transmission errors.

[0080] In the data processing unit, data preprocessing is first performed, including denoising, outlier filtering, and data formatting. Noise removal uses an adaptive filtering algorithm to filter out noise signals that may be caused by external electromagnetic interference or mechanical vibration; outlier filtering sets a reasonable temperature range, marks temperature data that exceeds the range as outliers and removes them; data formatting unifies data from different sensors into a standardized format for subsequent calculation and analysis.

[0081] The processed temperature data is stored in the central database in sequence. The database needs to have efficient data retrieval and update capabilities to support subsequent temperature prediction and fan control. The data storage structure is organized according to time series and spatial location, so that in subsequent analysis, the temperature data at a specific time or specific area can be quickly located and extracted. After storage is completed, the data processing unit will update the latest data records in real time and provide basic data for the next grayscale prediction model calculation.

[0082] In the temperature control process of dry-type transformers, accurate prediction and analysis of future temperature change trends are crucial for optimizing fan speed regulation and ensuring stable system operation. The dynamic change of temperature is not only affected by the current operating state, but also closely related to the historical operating mode. To this end, by differential processing of the temperature data of multiple sensors inside the transformer and the application of the grayscale prediction model, it is possible to capture the subtle fluctuations of temperature changes and generate a short-term temperature trend prediction curve. These prediction results are integrated to construct the overall temperature trend curve of the transformer, laying the foundation for subsequent similarity analysis. The multidimensional dynamic time warping (MDTW) algorithm plays a key role in this process. By matching the current temperature trend with historical data, it accurately evaluates whether the current temperature change is highly similar to the historical pattern. With the help of this similarity analysis, effective temperature control strategies in history, such as peak shaving and valley filling strategies, can be used to optimize fan speed regulation and improve the accuracy and efficiency of temperature management.

[0083] Step S2 includes the following contents:

[0084] S2.1, first of all, for the internal layout of the transformer A temperature sensor collects temperature data. ,in Indicates the sensor number, Represents a time point, performs first-order difference processing, and calculates the temperature increment at each time point , eliminating the linear trend in the original time series and making the data more suitable for processing by the grayscale prediction model.

[0085] This difference calculation can help better capture subtle changes in temperature and is suitable for non-stationary time series in short-term forecasts.

[0086] S2.2, based on the differential data, the grayscale prediction model (GM) is applied, and the improved form of the GM (1,1) model is used for prediction. The construction of the grayscale model is based on the cumulative generated sequence (AGO), which generates a smooth data sequence by accumulation to make it conform to the characteristics of the exponential model. Calculate the cumulative generated sequence : ;

[0087] The prediction model based on the cumulative sequence is: ;

[0088] in, and is the model parameter, estimated by the least squares method. By solving this differential equation, we get the solution of the cumulative generated sequence: ;

[0089] Then, the accumulated sequence is restored to obtain the predicted value of the temperature increment : ;

[0090] Finally, the original temperature series is restored by accumulation to obtain the predicted value of short-term temperature trend. : ;

[0091] Through the above steps, the short-term temperature trend prediction curve of each sensor is generated.

[0092] S2.3, the predicted temperature values ​​of all sensors Summarize and construct the overall temperature trend curve of the transformer This curve comprehensively considers the temperature changes of key parts inside the transformer and reflects the future trend of overall temperature changes.

[0093] ;

[0094] in, It is a multidimensional function that integrates the predicted values ​​of each sensor to generate an overall temperature trend curve.

[0095] S2.4, in order to evaluate the similarity between the current overall temperature trend curve and historical data, the multidimensional dynamic time warping (MDTW) algorithm is used for similarity analysis. Assume that the historical data set is , the current prediction data set is .

[0096] The overall objective function of similarity calculation is: ;

[0097] : Path smoothness penalty coefficient, used to adjust the path continuity in similarity analysis.

[0098] : The distance between path nodes, controlling the path smoothness of similarity analysis.

[0099] Among them, the first term represents the temperature difference of all sensors at each time point, and the second term is the path smoothness penalty to ensure the rationality of the path in the similarity analysis.

[0100] Through the MDTW algorithm, the data from all sensors are integrated into a multidimensional similarity score that reflects the overall similarity of the current temperature trend with the historical pattern.

[0101] S2.5, the similarity score calculated by MDTW is used to determine whether the current temperature trend is highly similar to a certain section of historical data. If the similarity score is less than the corresponding threshold, it means that the current temperature trend is highly similar to a certain pattern in the historical data. At this time, it is considered that the current temperature change trend is very close to what happened in history. The system can refer to the control strategy in the historical data and execute similar temperature control measures (such as peak shaving and valley filling strategy) to optimize the current fan speed regulation and temperature management, marked as a short-term temperature trend curve;

[0102] On the contrary, if the similarity score is greater than or equal to the corresponding threshold, it means that the similarity between the current temperature trend and the historical data is low, or almost non-existent. At this time, the current temperature change pattern is quite different from the historically recorded data, and the system cannot be controlled directly according to the strategy in the historical data. It may need to rely on real-time data and standard dynamic control methods to adjust the temperature and avoid wrong control decisions caused by referring to historical patterns.

[0103] Through the processing of step S2, the short-term temperature trend of the dry-type transformer can be accurately predicted at an early stage, and the predicted data of each sensor can be integrated to construct a trend curve reflecting the overall temperature change of the transformer. The similarity analysis performed by the MDTW algorithm can accurately identify the similarity between the current temperature curve and the historical data, and draw on the temperature control strategies that have been proven effective in history, such as the peak-shaving and valley-filling strategies, so as to reduce unnecessary calculations and experiments and improve the control efficiency. More importantly, by analyzing the similarity between the current temperature change pattern and the historical pattern, it can help the identification and analysis of subsequent temperature changes based on mature, controllable and recognizable patterns, ensuring that accurate decisions can be made during the temperature control process to avoid misjudgment. When the similarity score is high, the historical pattern is directly used for control to achieve precise control; when the similarity is low, avoid blindly referring to inapplicable historical patterns, rely on real-time data for dynamic adjustment, ensure the accuracy of temperature management and the stability of the system, and optimize the fan speed control strategy, extend the fan service life, and improve the overall operation efficiency of the system on the premise of ensuring the safe operation of the dry-type transformer.

[0104] Step S3 includes the following contents:

[0105] The purpose of identifying the similarity of temperature curves is to determine whether the current temperature change trend is similar to the pattern that has appeared in history. If the current temperature curve is highly similar to a certain section in the historical data, it means that the current temperature change pattern may have occurred in history, and there may have been a successful temperature control strategy in the historical data. This provides a basis for executing similar control strategies, reducing unnecessary calculations and experiments, and directly applying proven effective solutions to improve control efficiency and accuracy.

[0106] The peak-shaving strategy aims to smooth temperature fluctuations by adjusting the switching point of the fan speed. The execution of this strategy mainly depends on the matching degree of the current temperature trend with similar situations in history:

[0107] Similar signal: The current temperature curve is highly similar to the historical data, indicating that the current temperature change may have similar peaks and valleys. If no intervention is taken, it may cause the fan to adjust speed frequently, thereby increasing mechanical wear and energy consumption. In this case, the peak-shaving and valley-filling strategy is implemented. By properly adjusting the switching point of the fan speed, possible temperature fluctuations can be alleviated in advance, making the fan run more smoothly and extending its service life.

[0108] Dissimilar signal: The current temperature curve has a low similarity with historical data, which means that the current situation is quite different from the historical pattern. At this time, the control strategy in the historical data may not be applicable to the current situation, and blindly implementing the peak shaving and valley filling strategy may cause the temperature control to fail. Therefore, in the case of low similarity, dynamic control should rely on real-time data rather than historical patterns to ensure the adaptability of the system and the accuracy of control.

[0109] In the temperature control process of dry-type transformers, the fluctuation characteristics of the temperature curve directly affect the fan speed control strategy and the stability of the system. As a means of optimizing fan speed control, the effectiveness of the peak shaving and valley filling strategy depends on the periodic fluctuation characteristics of the temperature curve. In order to accurately identify the temperature curve suitable for peak shaving and valley filling, the key indicator of the periodic symmetry index is introduced. The periodic symmetry index quantifies the symmetry of the peak and trough of the temperature curve in each cycle, helps to determine whether the fluctuation of the curve is balanced and symmetrical, and thus judges the feasibility of the peak shaving and valley filling strategy. The higher the index, the more symmetrical the curve fluctuation is, and the fan speed control is easier to achieve stable temperature control; the lower the index, the more uneven the fluctuation is, and the peak shaving and valley filling strategy is difficult to apply effectively. Through the evaluation of the periodic symmetry index, more efficient strategy optimization can be achieved in transformer temperature management to ensure the stable operation of the system.

[0110] The fluctuation characteristics of the temperature curve directly affect the effectiveness of the peak shaving and valley filling strategy. The periodic symmetry index and the periodic amplitude change rate are key indicators. By quantitatively analyzing the symmetry and stability of the temperature curve in periodic fluctuations, it is possible to comprehensively evaluate whether the curve is suitable for implementing the peak shaving and valley filling strategy. The curve with a high periodic symmetry index indicates that the symmetry between the peak and the trough is good, which means that the temperature fluctuation is relatively regular and easy to smooth by adjusting the fan speed; while the curve with a low periodic amplitude change rate shows that the amplitude change between each cycle is small, the temperature fluctuation is more stable, and the peak shaving and valley filling strategy can more effectively reduce the difference between the temperature peak and valley value. Through the comprehensive analysis of these two indicators, those temperature curves with stability and regularity can be accurately screened out to ensure that the implementation of the peak shaving and valley filling strategy can significantly optimize the fan speed regulation, extend the fan service life, and keep the temperature of the dry-type transformer within a safe range, thereby improving the overall operating efficiency and stability of the system.

[0111] The logic for obtaining the periodic symmetry index is:

[0112] a1, short-term temperature trend curve Perform Fourier transform to identify the dominant frequency of the short-term temperature trend curve. Fourier transform converts the time domain signal into a frequency domain signal, which is expressed as: ;

[0113] in:

[0114] Represents the frequency domain signal, that is, the frequency The corresponding signal amplitude.

[0115] Is an imaginary unit.

[0116] It is the short-term temperature trend curve.

[0117] Determine the dominant frequency by analyzing the amplitude of the frequency domain signal , which is the frequency corresponding to the maximum amplitude in the frequency domain signal. The period corresponding to the dominant frequency is: ;

[0118] a2. After determining the dominant frequency, extract the main period short-term temperature trend curve corresponding to this frequency from the short-term temperature trend curve , which reflects the most significant periodic fluctuation in the short-term temperature trend curve. The main periodic component is reconstructed using inverse Fourier transform: ;

[0119] in:

[0120] It is the short-term temperature trend curve of the main period.

[0121] represents the inverse Fourier transform.

[0122] a3, evaluate the symmetry within a cycle through the short-term temperature trend curve of the main cycle. The starting point is , calculate the symmetry difference of the temperature fluctuation within the period. The symmetry difference is the integral of the difference between the temperature peak and the corresponding trough within the period: ;

[0123] in:

[0124] It represents the difference in symmetry within a cycle and measures the symmetry between peaks and troughs.

[0125] The start time of the cycle.

[0126] a4, calculate the periodic symmetry index based on symmetry differences The symmetry difference of each period segment is normalized to obtain the periodic symmetry index of the entire short-term temperature trend curve: ;

[0127] in:

[0128] is the total number of cycles.

[0129] is the value of the maximum symmetry difference within all periodic segments, used for normalization.

[0130] For the The starting time of a cycle.

[0131] The larger the value of the periodic symmetry index, the more symmetrical the periodic fluctuation of the short-term temperature trend curve is, and it is suitable for implementing the peak shaving and valley filling strategy. On the contrary, if the value of the periodic symmetry index is small, it means that the fluctuation symmetry of the short-term temperature trend curve is poor, and it is not suitable for implementing the peak shaving and valley filling strategy.

[0132] The cyclical symmetry index is used to represent and reflect the symmetry characteristics of the short-term temperature trend curve in the cyclical fluctuations. It mainly measures the symmetry between the peaks and troughs of the short-term temperature trend curve in each cycle. The larger the cyclical symmetry index, the more symmetrical the cyclical fluctuations of the short-term temperature trend curve, that is, in each cycle, the changes in the peaks and troughs are relatively balanced, which is suitable for implementing the peak-cutting and valley-filling strategy to smooth temperature fluctuations. On the contrary, the smaller the cyclical symmetry index, the poorer the fluctuation symmetry of the short-term temperature trend curve, the uneven changes between the peaks and troughs in the cycle, and the unsuitable for implementing the peak-cutting and valley-filling strategy, because it is difficult to achieve temperature balance through simple speed adjustment.

[0133] In the temperature control process of dry-type transformers, the fluctuation characteristics of the temperature curve directly affect the fan speed regulation strategy and the overall stability of the system. The peak shaving and valley filling strategy aims to smooth temperature fluctuations by adjusting the fan speed, thereby achieving more stable temperature control. The cycle amplitude change rate is an important indicator to measure the degree of change of peaks and troughs within the short-term temperature trend curve cycle, which can effectively reflect the fluctuation stability of the short-term temperature trend curve in different cycles. By evaluating the cycle amplitude change rate, it is possible to accurately determine whether the short-term temperature trend curve has the conditions for executing the peak shaving and valley filling strategy. If the amplitude change rate is low, it means that the short-term temperature trend curve fluctuates more evenly, and the peak shaving and valley filling strategy can be effectively applied; while a higher amplitude change rate indicates that the short-term temperature trend curve fluctuates violently, and the peak shaving and valley filling strategy may be difficult to smooth the fluctuations. Therefore, the cycle amplitude change rate plays an important role in optimizing the fan speed regulation strategy and improving the system operation efficiency, and is a key indicator to ensure the stability of transformer temperature management.

[0134] The logic for obtaining the cycle amplitude change rate is:

[0135] b1, first, according to step a1 short-term temperature trend curve The dominant frequency and its corresponding cycle :

[0136] b2. After determining the main cycle, segment the short-term temperature trend curve and identify the peak amplitude in each cycle segment according to each cycle. and trough amplitude The calculations of the peak amplitude and trough amplitude are:

[0137] ;

[0138] ;

[0139] in:

[0140] For the The starting time of a cycle.

[0141] For the The peak amplitude within a cycle.

[0142] For the The trough amplitude within a cycle.

[0143] b3, calculate the amplitude difference between the peak amplitude and the trough amplitude in each cycle , the difference represents the fluctuation range of the short-term temperature trend curve within the period : ;

[0144] This difference reflects the amplitude of the temperature change within each cycle.

[0145] The cycle amplitude change rate is used to represent and reflect the stability and consistency of the amplitude change of the short-term temperature trend curve in each cycle. It mainly measures the difference in amplitude between the peak and the trough of the short-term temperature trend curve in different cycles. The larger the cycle amplitude change rate, the more dramatic the amplitude change between cycles, the unstable temperature fluctuation, and the uneven fluctuation characteristics of the curve in different cycles, which is not suitable for implementing the peak shaving and valley filling strategy; on the contrary, the smaller the cycle amplitude change rate, the more stable the amplitude change between cycles, the relatively uniform temperature fluctuation, and the high consistency of the fluctuation characteristics of the curve in different cycles, which is suitable for implementing the peak shaving and valley filling strategy to smooth the temperature fluctuation.

[0146] The resonant fluctuation control index is obtained by comprehensively analyzing and calculating the periodic symmetry index and the periodic amplitude change rate. For example, the product normalization method is used to calculate: ;

[0147] in, is the resonant fluctuation control index; is a normalization coefficient used to balance the dimension difference between the periodic symmetry index and the periodic amplitude change rate.

[0148] The resonant fluctuation control index is used to represent and reflect the adaptability and operability of the short-term temperature trend curve in the peak shaving and valley filling strategy. By combining the periodic symmetry index and the periodic amplitude change rate, the resonant fluctuation control index can quantify the symmetry and stability of the short-term temperature trend curve in the periodic fluctuation, so as to evaluate whether the curve is suitable for optimization through the peak shaving and valley filling strategy. Specifically, the resonant fluctuation control index reflects the fluctuation characteristics of the short-term temperature trend curve, especially the degree of coordination between the symmetry of the periodic fluctuation and the stability of the amplitude change. The higher the index, the more suitable the fluctuation characteristics of the short-term temperature trend curve are for peak shaving and valley filling, and the expected effect of the control is more ideal; the lower the index, the more suitable the fluctuation characteristics of the short-term temperature trend curve are for peak shaving and valley filling, and the control is more difficult or the effect is not ideal. Through the resonant fluctuation control index, the peak shaving and valley filling strategy can be judged and executed more accurately, the fan speed regulation and temperature management can be optimized, and the stable operation of the system can be ensured.

[0149] Comparing the resonant fluctuation regulation index with the identification threshold;

[0150] If the resonant fluctuation control index is greater than or equal to the identification threshold, it means that the fluctuation characteristics of the short-term temperature trend curve are suitable for the peak shaving and valley filling strategy. At this time, the periodic symmetry and stability of the short-term temperature trend curve have reached the standard that can be effectively controlled by peak shaving and valley filling, and the expected control effect is good. The system can use the peak shaving and valley filling strategy to smooth temperature fluctuations, optimize fan speed regulation and temperature management, and generate actionable signals.

[0151] If the resonant fluctuation control index is less than the identification threshold, it means that the fluctuation characteristics of the short-term temperature trend curve are not suitable for the peak shaving and valley filling strategy. At this time, the short-term temperature trend curve has deficiencies in periodic symmetry or amplitude change stability. It is difficult to obtain the ideal control effect through the peak shaving and valley filling strategy. It may be necessary to rely on other temperature management methods or dynamic speed control modes to maintain the stability of the system and generate an inoperable signal.

[0152] c1. After obtaining the operational signal, based on the amplitude difference calculated in step b3, analyze the amplitude difference change trend between each cycle and calculate the amplitude change rate : ;

[0153] in: Indicates The cycle and The rate of change of amplitude between cycles.

[0154] The purpose of analyzing the amplitude change rate is to evaluate the fluctuation stability of the temperature curve in different periods and determine the timing of implementing the peak shaving and valley filling strategy.

[0155] c2, based on the recognition results of the peaks and troughs and the amplitude change trend, adjust the switching point of the fan speed , in order to achieve the effect of peak shaving and valley filling. The fan speed switching point is adjusted by the following formula: ;

[0156] in:

[0157] The fan speed is switched at the adjusted time.

[0158] To adjust the coefficient, set it according to actual needs to control the adjustment range of the switching point.

[0159] When implementing the peak shaving and valley filling strategy, the fan speed switching point is adjusted mainly based on the fluctuation intensity and change trend of the temperature curve, and the switching time is advanced or delayed to smooth the temperature fluctuation.

[0160] c3, after adjusting the fan speed switching point, real-time monitoring of the changes in the short-term temperature trend curve and the fan speed The actual response of the peak load shaving strategy is evaluated. Real-time adjustment is made through the following feedback formula: ; ;

[0161] in:

[0162] is the error between the actual temperature and the predicted temperature.

[0163] The fan speed adjustment amount.

[0164] It is the feedback gain coefficient, which determines the size of the adjustment.

[0165] By executing step S3, the temperature curve suitable for implementing the peak shaving and valley filling strategy can be accurately identified and screened to ensure that peak shaving and valley filling is implemented only under the most appropriate conditions. Through the comprehensive analysis of the periodic symmetry index and the periodic amplitude change rate, the calculated resonance fluctuation control index is compared with the identification threshold, which can accurately determine whether the fluctuation characteristics of the temperature curve are suitable for peak shaving and valley filling, thereby optimizing the fan speed regulation strategy. By adjusting the fan speed switching point, the temperature fluctuation is smoothed, the frequent speed regulation of the fan is effectively reduced, the mechanical wear is reduced, and the service life of the fan is extended. At the same time, the real-time monitoring and feedback adjustment mechanism ensures that the temperature is always controlled within a reasonable range, ensuring that the dry-type transformer maintains stable operation and efficient temperature management of the system while optimizing fan management.

[0166] Step S4 includes the following contents:

[0167] If an inoperable signal is obtained, the traditional speed control mode will be switched immediately to ensure the timeliness and accuracy of temperature control. In the traditional mode, the fan speed will be adjusted directly according to the real-time monitored temperature data, without considering the smoothing effect of the peak shaving and valley filling strategy on temperature fluctuations. Specifically, the traditional mode will directly adjust the fan speed through dynamic feedback of real-time temperature data to quickly respond to temperature changes and ensure that the temperature of the dry-type transformer is always maintained within a safe operating range. The purpose of this is to rely on the real-time temperature feedback mechanism to quickly adjust the speed in the case of complex or unstable temperature fluctuations, so as to avoid the failure or delay of temperature control that may be caused by the peak shaving and valley filling strategy under unsuitable conditions, and ensure the safety of the system and the stability of operation.

[0168] Example 2

[0169] Figure 2 The present invention provides a transformer temperature control system, comprising: a data acquisition module, a trend prediction module, a similarity recognition module and a fan adjustment module;

[0170] The data acquisition module collects transformer temperature data in real time, stores it after filtering, and the processed data will be passed to the trend prediction module;

[0171] The trend prediction module generates a short-term temperature trend curve based on the grayscale prediction model, matches historical data for similarity analysis, and sends the similarity analysis results to the similarity recognition module;

[0172] After the similarity recognition module recognizes the similarity of the temperature curve, it determines whether to implement the peak shaving and valley filling strategy, adjusts the fan speed switching point, smoothes the temperature fluctuation, and sends the judgment result to the fan adjustment module;

[0173] If the fan adjustment module determines that it is, it will adjust the fan speed in real time according to the peak shaving and valley filling strategy, otherwise it will adjust the speed according to the traditional mode.

[0174] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0175] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0176] It should be noted that, in this article, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including a..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0177] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A transformer temperature control method, characterized in that: Includes steps: S1, real-time collection of transformer temperature data, and storage after filtering; S2, generates short-term temperature trend curve based on grayscale prediction model and matches historical data for similarity analysis; S3, after identifying the similarity of the temperature curves, determine whether to implement the peak shaving and valley filling strategy, adjust the fan speed switching point, and smooth the temperature fluctuation; S4, if the answer is yes, adjust the fan speed in real time according to the peak-shaving and valley-filling strategy, otherwise adjust the speed according to the traditional mode; Step S2 includes the following contents: S2.1, first of all, for the internal layout of the transformer A temperature sensor collects temperature data. ,in Indicates the sensor number, Represents a time point, performs first-order difference processing, and calculates the temperature increment at each time point , remove the linear trend in the original time series; S2.2, based on the differential data, the grayscale prediction model is applied, and the improved form of the GM(1,1) model is used for prediction; the construction of the grayscale model is based on the cumulative generation sequence, and the smooth data sequence is generated by accumulation; the cumulative generation sequence is calculated : ; The prediction model based on the cumulative sequence is: ; in, and are model parameters, estimated by the least squares method; by solving this differential equation, we get the solution of the cumulative generated sequence: ; Then, the accumulated sequence is restored to obtain the predicted value of the temperature increment : ; Finally, the original temperature series is restored by accumulation to obtain the predicted value of short-term temperature trend. : ; Finally, a short-term temperature trend prediction curve for each sensor is generated; S2.3, the predicted temperature values ​​of all sensors Summarize and construct the overall temperature trend curve of the transformer ; S2.4, use the multidimensional dynamic time warping algorithm to perform similarity analysis; suppose the historical data set is , the current prediction data set is ; The overall objective function of similarity calculation is: ; : Path smoothness penalty coefficient, used to adjust the path continuity in similarity analysis; : The distance between path nodes, controlling the path smoothness of similarity analysis; S2.5, the similarity score calculated by MDTW is used to determine whether the current temperature trend is highly similar to a certain section of the historical data; if the similarity score is less than the corresponding threshold, it means that the current temperature trend is highly similar to a certain section of the historical data; at this time, it is considered that the current temperature change trend is very close to what happened in history and is marked as a short-term temperature trend curve; Step S3 includes the following contents: The logic for obtaining the periodic symmetry index is: a1, short-term temperature trend curve Perform Fourier transform to identify the dominant frequency of the short-term temperature trend curve; Determine the dominant frequency by analyzing the amplitude of the frequency domain signal , that is, the frequency corresponding to the maximum amplitude in the frequency domain signal; a2. After determining the dominant frequency, extract the main period short-term temperature trend curve corresponding to this frequency from the short-term temperature trend curve ; Reconstruct the main periodic component using inverse Fourier transform; a3, evaluate the symmetry within a cycle through the short-term temperature trend curve of the main cycle; The starting point is , calculate the symmetry difference of the temperature fluctuation within the period; the symmetry difference is the integral of the difference between the temperature peak and the corresponding trough within the period: ; in: It represents the difference in symmetry within the cycle and measures the symmetry between the peaks and troughs; is the cycle start time; a4, calculate the periodic symmetry index based on symmetry differences ; The symmetry difference of each period segment is normalized to obtain the periodic symmetry index of the entire short-term temperature trend curve: ; in: is the total number of cycles; is the value of the maximum symmetry difference in all periodic segments, used for normalization; For the The starting time of each cycle; The logic for obtaining the cycle amplitude change rate is: b1, first, according to step a1 short-term temperature trend curve The dominant frequency and its corresponding cycle : b2. After determining the main cycle, segment the short-term temperature trend curve and identify the peak amplitude in each cycle segment according to each cycle. and trough amplitude ; b3, calculate the amplitude difference between the peak amplitude and the trough amplitude in each cycle , the amplitude difference indicates the fluctuation amplitude of the short-term temperature trend curve in the corresponding period : ; b4, then, analyze the amplitude change trend of each cycle and calculate the cycle amplitude change rate The purpose is to quantify the amplitude change of the short-term temperature trend curve between each cycle; the cycle amplitude change rate is calculated as the speed of change of the amplitude difference between each cycle: ; The resonant fluctuation control index is calculated by comprehensive analysis of the periodic symmetry index and the periodic amplitude change rate; the resonant fluctuation control index is compared with the identification threshold; If the resonant fluctuation control index is greater than or equal to the identification threshold, it means that the fluctuation characteristics of the short-term temperature trend curve are suitable for executing the peak shaving and valley filling strategy, generating an actionable signal; If the resonant fluctuation control index is less than the identification threshold, it means that the fluctuation characteristics of the short-term temperature trend curve are not suitable for executing the peak shaving and valley filling strategy, and an inoperable signal is generated.

2. A transformer temperature control method according to claim 1, characterized in that: c1. After obtaining the operational signal, based on the amplitude difference calculated in step b3, analyze the amplitude difference change trend between each cycle and calculate the amplitude change rate : ; in: Indicates The cycle and The rate of change of amplitude between cycles; c2, based on the recognition results of the peaks and troughs and the amplitude change trend, adjust the switching point of the fan speed , adjust the fan speed switching point by the following formula: ; in: The fan speed switching time after adjustment; To adjust the coefficient, set it according to actual needs to control the adjustment range of the switching point.

3. A transformer temperature control method according to claim 2, characterized in that: Step S4 includes the following contents: If an inoperable signal is received, the system will immediately switch to the traditional speed control mode. In the traditional mode, the fan speed will be adjusted directly based on the real-time monitored temperature data without considering the smoothing effect of the peak shaving and valley filling strategy on temperature fluctuations.

4. A transformer temperature control system, used to implement a transformer temperature control method according to any one of claims 1 to 3, characterized in that: include: Data acquisition module, trend prediction module, similarity recognition module and fan adjustment module; The data acquisition module collects transformer temperature data in real time, stores it after filtering, and the processed data will be passed to the trend prediction module; The trend prediction module generates a short-term temperature trend curve based on the grayscale prediction model, matches historical data for similarity analysis, and sends the similarity analysis results to the similarity recognition module; After the similarity recognition module recognizes the similarity of the temperature curve, it determines whether to implement the peak shaving and valley filling strategy, adjusts the fan speed switching point, smoothes the temperature fluctuation, and sends the judgment result to the fan adjustment module; If the fan adjustment module determines that it is, it will adjust the fan speed in real time according to the peak shaving and valley filling strategy, otherwise it will adjust the speed according to the traditional mode.

Images