Control method and system for temperature control instrument of switch cabinet

Abstract

The invention discloses a control method and system for a temperature control instrument of a switch cabinet, and relates to the technical field of intelligent electrical equipment thermal management, and the method comprises the steps: building a thermal effect periodic change function according to the obtained first data of the switch cabinet and a preset period length; constructing a load-heat effect rhythm regulation and control model based on the heat effect periodic change function and the obtained current load data; based on the regulation and control model, generating a switch cabinet heat effect prediction result through machine learning; generating a temperature control response strategy through fuzzy logic reasoning according to the prediction result and the real-time load state data; and regulating and controlling the operation state of the temperature control instrument control system according to the temperature control response strategy. The problem that a traditional switch cabinet temperature control instrument control method is difficult to generate a temperature control response strategy according to the heat effect is solved.

Description

Technical Field

[0001] This invention relates to the field of thermal management technology for intelligent electrical equipment, and more specifically, to a control method and system for a temperature control instrument for a switchgear. Background Technology

[0002] As a critical power distribution device in a power system, switchgear generates heat due to load current during long-term operation, leading to temperature increases. Excessive temperatures accelerate insulation aging, reduce mechanical strength, cause loose connections, and even result in equipment failure, severely impacting power supply reliability and equipment lifespan. Therefore, effective monitoring and intelligent control of the switchgear's internal temperature are crucial.

[0003] Traditional switchgear temperature control instruments mostly employ fixed threshold-based start-stop control or simple PID control methods. For example, when the temperature exceeds a certain set upper limit, the cooling device is activated, and it stops when the temperature drops to the lower limit. While these methods are simple and reliable, they have significant limitations: First, the thermal dynamics of switchgear exhibit significant time lag and nonlinearity. Temperature rise is not only related to the current load current but also influenced by a combination of factors such as historical load, ambient temperature and humidity, and airflow organization within the cabinet. Fixed threshold control methods struggle to adapt to complex and changing operating conditions, easily leading to control lag or frequent activation. Second, load current often exhibits periodic fluctuations (such as diurnal load variations and seasonal electricity consumption differences), but traditional methods fail to proactively learn and utilize these periodic rhythm characteristics for forward-looking regulation. Third, they fail to deeply integrate and analyze real-time load trends with environmental boundary conditions, resulting in insufficient intelligence and adaptive capabilities in the control strategy.

[0004] For example, the invention patent announcement CN102063056A discloses a digital temperature controller with intelligent setting and intelligent load adjustment functions. This includes research on intelligent setting and intelligent load adjustment functions in temperature control within the industrial control field, providing a design for a digital temperature controller with intelligent setting and intelligent load adjustment functions. Based on the correlation of multiple parameters, it automatically displays or hides user-set parameters; when setting a value, it intelligently determines whether the units, tens, and hundreds digits increase or decrease based on the duration of key presses; in the empty state before material feeding, it performs predetermined temperature control to obtain PID or fuzzy control parameters; and in the actual control process, during material feeding, it performs intelligent load adjustment based on the actual control situation to obtain PID or fuzzy control parameters corrected for the actual control process.

[0005] The above-disclosed technical solutions have at least the following technical problems: The system fails to consider the significant time-delay and nonlinear characteristics of load changes, heat accumulation, and temperature rise during the switchgear temperature rise process, and does not establish a correlation model between load and thermal effects. Parameter correction lags during sudden load changes, easily leading to temperature overshoot or untimely control. Secondly, it fails to extract and utilize the periodic rhythm characteristics accompanying the switchgear load current, adjusting parameters only based on real-time operating conditions. This lack of proactive control remains a "passive response" mode, making it difficult to avoid the risk of sudden temperature rises during peak load periods. Furthermore, its fuzzy control rules and parameter correction rely on preset logic and simple operating condition feedback, failing to optimize the rule base using historical switchgear operating data and lacking a time-series model-based thermal effect prediction mechanism. This makes it impossible to formulate control strategies in advance, easily leading to frequent start-stop of the cooling device in scenarios such as continuous load growth, affecting equipment lifespan and failing to meet the switchgear's high-precision and adaptive temperature control requirements.

[0006] To address the above problems, this invention proposes a solution. Summary of the Invention

[0007] To overcome the above-mentioned deficiencies of the prior art, embodiments of the present invention provide a control method and system for temperature control instruments in switchgear. By constructing a periodic change function of thermal effect, establishing a load-thermal effect rhythm control model, and integrating LSTM prediction and weighted fuzzy inference, the problem that traditional switchgear temperature control instrument control methods are difficult to generate temperature control response strategies based on thermal effects is solved.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A control method for a temperature control instrument in a switchgear, characterized by the following steps: constructing a thermal effect periodic variation function based on the acquired first data of the switchgear with a preset cycle length; constructing a load-thermal effect rhythm control model based on the thermal effect periodic variation function and the acquired current load data; generating a switchgear thermal effect prediction result through machine learning based on the control model; generating a temperature control response strategy through fuzzy logic reasoning based on the prediction result and real-time load status data; and controlling the operating state of the temperature control instrument control system according to the temperature control response strategy.

[0009] In a preferred embodiment, the period length is determined by extracting the period of thermal effect change through time series analysis.

[0010] In a preferred embodiment, the step of constructing a thermal effect periodic variation function based on the acquired first data of the switchgear with a preset period length specifically involves inputting the first data of the switchgear into a smooth fitting algorithm, using the period length as a constraint condition to perform data smooth fitting, and obtaining the thermal effect periodic variation function.

[0011] In a preferred embodiment, the step of constructing a load-thermal effect rhythm control model based on the thermal effect periodic variation function and current load data specifically involves: obtaining a load dynamic feature vector based on the current load data using a feature extraction algorithm; obtaining a joint feature sequence based on the load dynamic feature vector and the thermal effect periodic variation function using a multi-source feature fusion algorithm; and constructing a load-thermal effect rhythm control model based on the joint feature sequence using a time-series regression algorithm.

[0012] In a preferred embodiment, the step of generating the switchgear thermal effect prediction result based on the control model and through machine learning specifically involves: constructing an LSTM input sequence based on the output of the load-thermal effect rhythm control model; obtaining a regular time series sample set based on the LSTM input sequence through data standardization and time step alignment; and obtaining the switchgear thermal effect prediction result based on the time series sample set using a long short-term memory network algorithm.

[0013] In a preferred embodiment, the step of constructing the LSTM input sequence based on the output of the load-thermal effect rhythm regulation model specifically involves: extracting the load dynamic feature vector and the thermal effect feature vector based on the output of the load-thermal effect rhythm regulation model; obtaining a fused feature matrix by concatenating the load dynamic feature vector and the thermal effect feature vector; and constructing the LSTM input sequence based on the fused feature matrix using a sliding window sampling method.

[0014] In a preferred embodiment, the step of generating a temperature control response strategy based on the prediction results and real-time load status data through fuzzy logic reasoning specifically involves: constructing a membership function based on the prediction results and real-time load status data to obtain a fuzzy set of temperature control status; generating a weighted fuzzy rule base based on the acquired historical data of the switchgear; performing reasoning based on the fuzzy set of temperature control status through the weighted fuzzy rule base to obtain a set of fuzzy temperature control strategies; and generating a temperature control response strategy based on the set of fuzzy temperature control strategies through defuzzification calculation.

[0015] In a preferred embodiment, generating a weighted fuzzy rule library based on historical switchgear data specifically involves: obtaining typical operating condition patterns through clustering algorithms based on historical load and temperature data; extracting initial fuzzy rules through association rule mining methods based on typical operating condition patterns; calculating the weight coefficients of each rule through a weight learning algorithm based on historical operating effect data; and generating a weighted fuzzy rule library by integrating the initial fuzzy rules and weight coefficients.

[0016] In a preferred embodiment, the step of extracting initial fuzzy rules based on typical operating condition patterns using association rule mining methods specifically involves: discretizing and fuzzily classifying the features in the typical operating condition patterns to form a set of operating condition feature items; based on the set of operating condition feature items, converting each operating condition sample into an operating condition sample record to construct an operating condition sample database; calculating the support of the set of operating condition feature items in the operating condition sample database, and filtering out frequent itemsets by setting a minimum support threshold; based on the frequent itemsets, generating an association rule set according to the division rules of input and output items; and mapping the operating condition items in the association rule set to corresponding fuzzy linguistic variables to form an initial fuzzy rule set with an "if-then" structure.

[0017] A control system for a temperature control instrument in a switchgear includes: a function construction module, which constructs a thermal effect periodic variation function based on the acquired first data of the switchgear and a preset period length; a model construction module, which constructs a load-thermal effect rhythm control model based on the thermal effect periodic variation function and the acquired current load data; a thermal effect prediction module, which generates a switchgear thermal effect prediction result through machine learning based on the control model; a strategy generation module, which generates a temperature control response strategy through fuzzy logic reasoning based on the prediction result and real-time load status data; and a control execution module, which regulates the operating state of the temperature control instrument control system according to the temperature control response strategy.

[0018] The technical effects and advantages of the control method and system for temperature control instruments in switchgear of the present invention are as follows: This invention discloses a control method and system for temperature control instruments in switchgear. It extracts periodic features of thermal effects from multi-source switchgear data using a smoothing fitting algorithm combined with time series analysis and constructs a periodic variation function to accurately capture the rhythmic pattern of thermal effects fluctuating with load. A multi-source feature fusion algorithm integrates load dynamic features and the periodic function of thermal effects to construct a load-thermal effect rhythmic control model to characterize their nonlinear dynamic relationship. Based on the LSTM algorithm, time-series prediction of the fused feature sequence is performed to achieve forward-looking trend prediction of thermal effects, overcoming the lag inherent in traditional "real-time response." Fuzzy logic reasoning is performed using a weighted fuzzy rule base generated from historical data to generate a temperature control strategy that matches real-time load and environmental conditions. This effectively overcomes the problems of traditional methods, such as ignoring periodicity, lack of correlation models, response lag, and poor rule adaptability, providing an efficient solution for thermal management of intelligent electrical equipment. It effectively solves the problem that traditional switchgear temperature control instruments struggle to generate temperature control response strategies based on thermal effects. Attached Figure Description

[0019] Figure 1 This is a flowchart of a control method for a temperature control instrument in a switchgear according to the present invention; Figure 2 This is a structural diagram of the control system for a temperature control instrument in a switchgear according to the present invention; Figure 3 A comparison chart showing the fitting results of the periodic function of the thermal effect and the temperature prediction results; Figure 4 This is a comparison chart of the temperature control effects of the present invention and traditional threshold control. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0022] Example 1, Figure 1 A control method for a temperature control instrument in a switchgear is provided, characterized by the following steps: S1, Based on the first data of the switchgear obtained, construct the thermal effect periodic change function with a preset cycle length; S2, a load-thermal effect rhythm control model is constructed based on the periodic variation function of the thermal effect and the acquired current load data; S3, based on the control model, generates prediction results of the thermal effect of the switchgear through machine learning; S4. Based on the prediction results and combined with real-time load status data, a temperature control response strategy is generated through fuzzy logic reasoning. S5 dynamically adjusts the operating status of the temperature control instrument control system according to the temperature control response strategy.

[0023] This invention discloses a control method and system for temperature control instruments in switchgear. It extracts periodic features of thermal effects from multi-source switchgear data using a smoothing fitting algorithm combined with time series analysis and constructs a periodic variation function to accurately capture the rhythmic pattern of thermal effects fluctuating with load. A multi-source feature fusion algorithm integrates load dynamic features and the periodic function of thermal effects to construct a load-thermal effect rhythmic control model to characterize their nonlinear dynamic relationship. Based on the LSTM algorithm, time-series prediction of the fused feature sequence is performed to achieve forward-looking trend prediction of thermal effects, overcoming the lag inherent in traditional "real-time response." Fuzzy logic reasoning is performed using a weighted fuzzy rule base generated from historical data to generate a temperature control strategy that matches real-time load and environmental conditions. This effectively overcomes the problems of traditional methods, such as ignoring periodicity, lack of correlation models, response lag, and poor rule adaptability, providing an efficient solution for thermal management of intelligent electrical equipment. It effectively solves the problem that traditional switchgear temperature control instruments struggle to generate temperature control response strategies based on thermal effects.

[0024] S1. Based on the first data of the switchgear obtained, construct a periodic change function of thermal effect with a preset cycle length.

[0025] In this embodiment, the first data of the switch cabinet includes temperature data, humidity data, and environmental parameters; the current load data includes the real-time current value of the circuit, the rate of change of current, the rated current, and the load duration.

[0026] In this embodiment, the period length is determined by extracting the period of thermal effect change through time series analysis.

[0027] In this embodiment, the step of constructing a thermal effect periodic variation function based on the acquired first data of the switchgear with a preset period length specifically involves inputting the first data of the switchgear into a smooth fitting algorithm, using the period length as a constraint condition, and performing data smooth fitting to obtain the thermal effect periodic variation function.

[0028] The period length of the thermal effect change is extracted from the first data of the switchgear to provide a constraint basis for subsequent fitting: The periodicity of data is analyzed by calculating the autocorrelation coefficient—for any time lag, the similarity between the data at that lag point and the original data is compared.

[0029] The lag times with high correlation coefficients are selected, and the most significant period length is found through cluster analysis to determine the main period.

[0030] Calculate the maximum temperature difference within a main cycle, and retain a certain tolerance range as the amplitude constraint for subsequent fitting; calculate the offset of this time relative to the start of the cycle, as the reference for the cycle phase.

[0031] The smooth fitting algorithm parameter configuration is based on the periodic characteristics, setting the algorithm parameters to balance data fitting accuracy and curve smoothness: The size of the smoothing window is determined based on the length of the main period, usually taking 1 / 8 of the main period, to ensure that the window can cover the local fluctuation characteristics within the period.

[0032] Data at different positions within the window are assigned different weights—the data at the center has the highest weight, which gradually decreases towards both sides, and the weight changes follow a Gaussian distribution.

[0033] Regularization parameters are determined through cross-validation. Specifically, the data is divided into a training set (80%) and a validation set (20%), multiple parameter values ​​are tested, and the parameter that minimizes the fitting error on the validation set is selected.

[0034] The formula for the smooth fitting algorithm is as follows:

[0035] in, The objective function value for the thermal effect period is... The size of the thermal effect periodic smoothing window is determined based on the length of the main thermal effect period. For the index of data points within the window, For the first in the window Gaussian weights for each data point For the first in the window The actual value of each original data point. The smoothed fit value is the value at the center of the window. The thermal effect periodicity regularization parameter, For the first in the window The smoothed fit value at each position, For the first in the window The smoothed fit value at each position, For the first in the window The smoothed fit value at each position.

[0036] The periodic variation function of the thermal effect is expressed in the form of a sine function to quantify the periodic variation pattern of the thermal effect, as shown in the following formula:

[0037] in, It is a periodic characteristic function of the change in thermal effect; The amplitude of the periodic characteristic of the change in thermal effect; Angular frequency; This is the phase offset. The mean term representing the periodic characteristic of the change in thermal effect. It is a sine function.

[0038] S2, a load-thermal effect rhythm control model is constructed based on the thermal effect periodic variation function and the acquired current load data.

[0039] In this embodiment, the construction of the load-thermal effect rhythm control model based on the periodic variation function of the thermal effect and the current load data specifically includes: Based on current load data, a dynamic feature vector of the load is obtained through a feature extraction algorithm; Based on the load dynamic feature vector and the thermal effect periodic change function, a joint feature sequence is obtained through a multi-source feature fusion algorithm. Based on the joint feature sequence, a load-heat effect rhythm regulation model is constructed using a time-series regression algorithm.

[0040] Extracting dynamic feature vectors of the load. Based on current load data, feature extraction algorithms are used to transform it into structured feature vectors. Specifically, the statistical and dynamic characteristics of the current are calculated from the time domain perspective, and the spectral characteristics of load fluctuations may also be extracted through frequency domain analysis, ultimately forming a high-dimensional vector that comprehensively reflects the real-time status and changing trends of the load.

[0041] The feature extraction algorithm refers to transforming the original time-series signal into a structured feature vector based on current load data through multi-dimensional analysis methods. Specifically, it includes: extracting statistical features such as mean, variance, maximum value, peak factor, rate of change, and proportion of sustained high load in the time domain to characterize the current load level and transient fluctuations; identifying load mutations, impact peaks, and their durations at the event level to reflect the load process characteristics; calculating the power spectral density and harmonic distribution in the frequency domain using Fourier transform to capture periodic fluctuations and random jitter characteristics; and forming a high-dimensional load dynamic feature vector that can comprehensively reflect the load intensity and dynamic trends after normalization and standardization.

[0042] Achieve multi-source feature fusion. The aforementioned load dynamic feature vector is jointly processed with the thermal effect periodic variation function obtained in step S1: the thermal effect periodic function is discretized into time-series features, and then the two types of features are integrated into a joint feature sequence through a multi-source feature fusion algorithm. For example, if the load feature vector at a certain moment is [500A, 0.2A / s, 1 cycle, 80%], and the output of the thermal effect periodic function at the same time is [35℃, in the rising phase of the cycle], the fusion will form a sequence containing "load-thermal effect" collaborative information, reflecting both the current load intensity and its position in the thermal effect cycle, thereby capturing the dynamic correlation between the two.

[0043] The multi-source feature fusion algorithm refers to jointly processing the load dynamic feature vector and the output of the thermal effect periodic variation function to form a joint feature sequence. Specifically, it includes: first, aligning the two types of data based on timestamps to ensure fusion at the same time; second, discretizing the thermal effect periodic variation function into structured features such as temperature value, periodic phase, and periodic segments; then, concatenating the above periodic features with the load features into an extended vector, and explicitly expressing typical coupling relationships such as high load with periodic peak and low load with periodic valley by constructing cross features; finally, organizing the extended feature vector into a joint feature sequence in chronological order to provide an input basis for time series modeling.

[0044] A time-series regression model was constructed. Using joint feature sequences as training data, a time-series regression algorithm was employed for modeling: the model input was the joint feature sequence over a certain period, and the output was the actual value of the thermal effect for the corresponding period. The model parameters were optimized through iterative training, enabling it to learn the nonlinear mapping relationship between dynamic load changes and the periodic pattern of the thermal effect. The resulting load-thermal effect rhythm regulation model can dynamically output thermal effect correlation results matching the periodic pattern based on load characteristics, providing a core correlation basis for subsequent thermal effect prediction.

[0045] The aforementioned time-series regression algorithm refers to the process of establishing a time-series prediction model by utilizing the relationship between the joint feature sequence and the actual thermal effect response. Specifically, it includes: using a sliding window to extract a fixed-length joint feature sequence as input, and using measured temperature, temperature rise rate, or thermal effect index as the output target; establishing a nonlinear mapping relationship between input and output through an autoregressive model, a state-space regression model, or a machine learning-based regression method; iteratively optimizing parameters during training to minimize the error between the predicted value and the actual thermal effect, enabling the model to accurately learn the coupling characteristics between load dynamics and the periodic law of thermal effect; and finally, the obtained time-series regression model can output a predicted value of thermal effect matching the operating condition when new load and periodic characteristics are input, providing core support for load-thermal effect rhythm regulation.

[0046] S3, based on the control model, generates prediction results of the thermal effect of the switchgear through machine learning.

[0047] In this embodiment, the step of generating the switchgear thermal effect prediction result based on the control model and through machine learning specifically involves: Based on the output of the load-heat effect rhythm regulation model, an LSTM input sequence is constructed; Based on the LSTM input sequence, a regular time series sample set is obtained through data standardization and time step alignment. Based on the time series sample set, the prediction results of the thermal effect of the switchgear are obtained through the Long Short-Term Memory Network algorithm.

[0048] Constructing the LSTM input sequence: Based on the output of the load-thermal effect rhythm regulation model, two core features are extracted: a load dynamic feature vector and a thermal effect feature vector. These two vectors are then fused into a fusion feature matrix containing collaborative information about the load and thermal effect through feature concatenation. A sliding window sampling method is then used to generate the LSTM input sequence: a window size is set, and the window slides in time steps; the feature matrix within each window constitutes an input sample, aiming to capture the short-term temporal dependence of the thermal effect on load changes.

[0049] Secondly, data standardization and time step alignment are performed. Since the dimensions of different features in the fused feature matrix vary significantly, standardization is needed to eliminate the influence of dimensions—commonly Z-score standardization or Min-Max standardization—to prevent a single feature from dominating model learning due to its large numerical range. Simultaneously, because there may be slight differences in sensor acquisition frequencies, linear interpolation or nearest-neighbor interpolation is used to unify the time step size, ensuring strict feature alignment at each time point in the input sequence. This ultimately results in a well-organized time series sample set, with each sample containing both the "input sequence" and the "target output."

[0050] Finally, the prediction results are generated using the LSTM algorithm. An LSTM network is constructed based on a regular sample set: the input layer dimension is consistent with the dimension of the fused features; 2-3 hidden layers are used, each containing 32-128 memory units, selectively retaining key temporal information through a gating mechanism; the output layer consists of one neuron, corresponding to the predicted thermal effect value at a future time. During training, mean squared error is used as the loss function, and the network parameters are iteratively adjusted using the Adam optimizer. Simultaneously, 80% of the samples are divided into a training set and 20% into a validation set, with early stopping used to prevent overfitting. After the model training is complete, inputting a new fused feature sequence will output the predicted thermal effect of the switchgear, providing a forward-looking basis for subsequent temperature control strategy generation.

[0051] In this embodiment, the construction of the LSTM input sequence based on the output of the load-heat effect rhythm regulation model specifically involves: Based on the output of the load-thermal effect rhythm control model, extract the load dynamic feature vector and the thermal effect feature vector; Based on the load dynamic feature vector and the thermal effect feature vector, a fused feature matrix is ​​obtained by feature concatenation; The LSTM input sequence is constructed based on the fused feature matrix using the sliding window sampling method.

[0052] When constructing the LSTM input sequence based on the load-thermal effect rhythm regulation model output, it is necessary to proceed step by step according to the logic of feature extraction, feature fusion and sequence generation to ensure that the input sequence can accurately carry the temporal correlation information of load and thermal effect.

[0053] Extract the load dynamic feature vector and the thermal effect feature vector. The output of the load-thermal effect rhythm control model already includes the correlation analysis results between the load and the thermal effect. Two types of core features need to be separated from it: On the one hand, extract the load dynamic feature vector. This vector needs to cover the real-time status and changing trend of the load, specifically including quantitative dimensions such as the real-time current value of the circuit, the rate of change of current, the proportion of rated current, and the duration of the load, so as to fully characterize the dynamic characteristics of the load. On the other hand, extract the thermal effect feature vector. This vector is based on the thermal effect periodic change function and includes the reference thermal effect of the switchgear, the thermal effect periodic phase, the humidity correction coefficient, etc., to accurately reflect the periodicity of the thermal effect and environmental influencing factors.

[0054] A fused feature matrix is ​​obtained through feature concatenation. Since the load dynamic feature vector and the thermal effect feature vector both correspond to the same time dimension, the two types of vectors must first be strictly aligned according to the timestamp to ensure that the load feature at a certain moment matches the thermal effect feature at the same time. Then, column concatenation is used to merge the load dynamic feature vector and the thermal effect feature vector at the same moment into a multi-dimensional feature row vector. Finally, the feature row vectors at all moments are arranged in chronological order to form a fused feature matrix. Each row of the matrix corresponds to a "load-thermal effect" collaborative feature at a certain time point, and each column corresponds to a type of feature, realizing the deep integration of multi-source features of load and thermal effect.

[0055] The LSTM input sequence is constructed using a sliding window sampling method based on the fused feature matrix. Considering that LSTM needs to learn patterns from historical time-series information, reasonable sliding window parameters must be set: the window size is determined based on the data acquisition frequency and prediction requirements, and the sliding step size is usually consistent with the data acquisition cycle. The window then slides along the time axis. After each slide, the fused features from the 12 time points contained within the window constitute an LSTM input sample. Through continuous sliding, the entire fused feature matrix is ​​transformed into multiple consecutive input samples, forming the input sequence required by the LSTM. This sequence effectively captures the short-term temporal dependency between load changes and the evolution of thermal effects, laying the foundation for the subsequent LSTM model to learn and predict patterns.

[0056] In the prediction process, the predicted values ​​of the thermal effect at future times are first output based on the control model, as shown in the following formula:

[0057] in, The first output of the model A switch cabinet in the future Predicted temperature at any given time; The switch cabinet number monitored by the sensors in the switch cabinet; For load-heat effect rhythm regulation model; For a moment The multimodal sensing data input vector; To predict the span of the time window.

[0058] A continuous forecast curve is plotted based on the predicted change values. The temperature change curve reflects the temperature evolution trend of each region within the forecast period (such as the next 2 hours, 4 hours or 24 hours), and is used to characterize the potential overheating risk or abnormal changes in the region.

[0059] S4 generates a temperature control response strategy based on the prediction results and real-time load status data through fuzzy logic reasoning.

[0060] In this embodiment, the step of generating a temperature control response strategy based on the prediction results and real-time load status data through fuzzy logic reasoning specifically involves: Based on the prediction results and real-time load status data, a membership function is constructed to obtain the fuzzy set of temperature control status; Based on the acquired historical data of the switchgear, a weighted fuzzy rule base is generated; Based on the fuzzy set of temperature control status, a set of fuzzy temperature control strategies is obtained through reasoning using a weighted fuzzy rule base; Based on a set of fuzzy temperature control strategies, a temperature control response strategy is generated through defuzzification calculation.

[0061] Fuzzification of input variables. The fuzzification interface transforms precise numerical inputs into fuzzy sets to suit the "language-based reasoning" characteristics of fuzzy logic.

[0062] The contextual naming of linguistic variables defines linguistic variables that fit the temperature control scenario for each input variable. For example, "real-time load rate" corresponds to "load intensity" and "predicted temperature deviation" corresponds to "temperature deviation degree", making the fuzzy results more in line with human language description habits for temperature control scenarios.

[0063] The reasonable division of fuzzy subsets divides "load intensity" into "low load", "medium load" and "high load".

[0064] "Temperature deviation" (derived from prediction results) is divided into "negative deviation (predicted temperature is lower than the target)", "zero deviation (predicted temperature is close to the target)" and "positive deviation (predicted temperature is higher than the target)".

[0065] In the selection of membership function type, the triangular function is adopted in the temperature control scenario. It is defined by three parameters: "left endpoint (a), vertex (b), and right endpoint (c)," which can intuitively reflect the degree to which the precise value belongs to a certain fuzzy subset.

[0066] The formula for the membership function of the triangle is:

[0067] in, Fuzzy subset of real-time data from switchgear Enter the exact value. For the type of fuzzy subset, Fuzzy subset of real-time data from switchgear membership degree Let be the left endpoint of the trigonometric function. The vertex of the trigonometric function, is the right endpoint of the trigonometric function.

[0068] The fuzzy set of input variables can be represented as:

[0069] in, For a fuzzy set of input variables, Fuzzy subset of real-time data from switchgear Enter the exact value. Fuzzy subset of real-time data from switchgear membership degree For the type of fuzzy subset, The general type of fuzzy subsets.

[0070] Generation of a weighted fuzzy rule base. The rule base construction requires combining historical data and introducing a weighting mechanism: First, based on historical load and temperature data, typical operating conditions are divided using clustering algorithms such as K-means to reduce rule redundancy; second, for each operating condition, initial fuzzy rules are extracted using the Apriori algorithm; third, based on historical operating performance data, a weight learning algorithm is used to assign weight coefficients to each rule; finally, these are integrated to form a weighted fuzzy rule base, where rules with higher weights have higher priority in inference.

[0071] Then comes fuzzy inference based on a rule base. Using the fuzzy set of input variables as a basis, triggering conditions are matched in a weighted rule base; the strength of each triggering rule is calculated; subsequently, the outputs of all triggering rules are aggregated and superimposed according to their strength to obtain a set of fuzzy temperature control strategies, reflecting the applicability of different strategies.

[0072] In this embodiment, the defuzzification calculation specifically includes: Receive a set of fuzzy temperature control strategies obtained based on fuzzy inference, wherein the set of fuzzy temperature control strategies is characterized by a membership function to represent the applicability of different temperature control action intensities; The centroid coordinates of the fuzzy temperature control strategy set are calculated using the centroid method, and the abscissa value of the centroid coordinates is mapped to the precise initial value of the temperature control command. Based on the physical characteristics of the switchgear cooling equipment, the initial value of the precise temperature control command is standardized and discretized to generate a specific temperature control response strategy that can be directly issued.

[0073] In this embodiment, generating a weighted fuzzy rule base based on historical switchgear data specifically involves: Based on historical load and temperature data, typical operating conditions are obtained through clustering algorithms; Based on typical operating conditions, initial fuzzy rules are extracted using association rule mining methods. Based on historical performance data, the weight coefficients of each rule are calculated using a weight learning algorithm. Based on the initial fuzzy rules and weight coefficients, a weighted fuzzy rule library is generated by integration.

[0074] Typical operating condition patterns are extracted using clustering algorithms based on historical load and temperature data. The core of clustering is to group similar operating states from massive amounts of historical data into one category, reducing redundancy and highlighting key patterns. Load and temperature dimensions that reflect operating states are selected from historical data, and clustering algorithms such as K-means are used for grouping. The elbow method is used to determine the optimal number of clusters. The algorithm groups data points into several clusters based on feature similarity, with each cluster representing a typical operating condition. These typical operating conditions remove data noise, providing a clear scenario foundation for subsequent rule extraction.

[0075] Initial fuzzy rules are extracted based on typical working conditions through association rule mining.

[0076] In this embodiment, the step of extracting initial fuzzy rules based on typical operating conditions using association rule mining methods specifically involves: The features in typical operating conditions are discretized and fuzzy-classified to form a set of operating condition feature items. Based on the set of operating condition feature items, each operating condition sample is converted into an operating condition sample record, and an operating condition sample database is constructed. Calculate the support of the set of working condition feature items in the working condition sample database, and filter out frequent itemsets by setting a minimum support threshold. Based on frequent itemsets, an association rule set is generated according to the partitioning rules of input and output items; The working condition items in the association rule set are mapped to the corresponding fuzzy linguistic variables to form an initial fuzzy rule set with an "if-then" structure.

[0077] Preferably, the characteristic data in typical operating conditions are first preprocessed. This characteristic data includes temperature data, current data, temperature rise rate, humidity data, and environmental parameters. The preprocessing steps include removing outlier data and completing missing data to ensure the completeness and accuracy of the input data. Further, the aforementioned characteristic data is segmented, dividing continuous values ​​into several intervals. Fuzzy linguistic variables are assigned to these intervals based on operating characteristics. For example, temperature is divided into low, medium, and high temperatures; current into light load, rated, and overload; and temperature rise rate into decreasing, stable, and increasing rates. This method forms a set of operating condition characteristic items, providing a foundation for subsequent fuzzy rule establishment.

[0078] Furthermore, based on the set of operating condition feature items, each operating condition sample is transformed into an operating condition sample record, with each sample containing several operating condition items. By collecting and transforming a large number of operating condition samples, an operating condition sample database is constructed. This operating condition sample database can comprehensively reflect the characteristic combination relationships of the switchgear under different operating conditions.

[0079] Preferably, in the operating condition sample database, the frequency of occurrence of operating condition feature item sets is statistically analyzed, support is calculated, and frequent itemsets are filtered out by setting a minimum support threshold. Filtering frequent itemsets ensures that the subsequently generated rules have a certain degree of representativeness and reliability, avoiding interference from low-frequency events on the effectiveness of the rules.

[0080] Furthermore, based on the frequent itemsets, candidate association rules are generated according to the partitioning of input and output items. For example, the input item could be a combination of high current load and rising temperature, and the output item could be rising temperature. Subsequently, confidence, lift, and relevance indices are calculated for the candidate rules and compared with preset thresholds to obtain an effective set of association rules. Simultaneously, redundancy resolution and conflict detection are performed on the effective rules, deleting duplicate or contradictory rules to ensure the simplicity and accuracy of the rule set.

[0081] Preferably, the operating condition items in the simplified association rule set are mapped to corresponding fuzzy linguistic variables, and their membership function relationships are combined to form an initial fuzzy rule set. For example, if the current is high and the rate of temperature change is increasing, then the temperature is high. This initial fuzzy rule set provides a direct basis for the subsequent establishment of a weighted fuzzy rule base.

[0082] It should be noted that the purpose of association rule mining is to discover the potential correlation between operating condition characteristics and effective temperature control strategies by analyzing the features of typical operating conditions. Specifically, association rule mining can not only reveal the coupling patterns between operating parameters such as current, temperature, and humidity, but also derive corresponding temperature control strategy patterns based on this, thus providing data support for the automatic generation of fuzzy rules. Compared with traditional methods that rely on manual experience to set rules, this approach can extract more universal and adaptable rule sets from a larger-scale data sample, ensuring the matching and forward-looking nature of temperature control strategies with actual operating conditions.

[0083] Weight coefficients for each rule are calculated using a weighted learning algorithm based on historical performance data. The weights quantify the actual effectiveness of a rule, giving higher priority to rules with better performance in inference. Historical performance data includes key metrics for each rule after past execution, such as cooling efficiency, energy cost, and equipment wear and tear. The weighted learning algorithm calculates weights by first assigning weights to each performance metric, then calculating a comprehensive performance score for each rule, and finally normalizing the scores to obtain the rule weight coefficient. Higher weights indicate better overall performance of the rule in historical applications.

[0084] Finally, the initial fuzzy rules and weight coefficients are integrated to generate a weighted fuzzy rule base. Each initial rule is bound to its corresponding weight coefficient, forming a structured entry of "rule content + weight value," such as "Rule 1: If the load is heavy and the temperature is high, the strategy is full fan speed + heatsink on," and "Rule 2: If the load is medium and the temperature fluctuates, the strategy is medium fan speed + 10-minute monitoring." A rule indexing mechanism is also established to ensure rapid matching of relevant rules during inference. The resulting weighted fuzzy rule base includes decision logic extracted from historical data and reflects the actual effect differences of rules through weights, providing a decision-making basis that balances experience and efficiency for subsequent fuzzy inference.

[0085] S5 adjusts the operating status of the temperature control instrument control system according to the temperature control response strategy.

[0086] In this embodiment, adjusting the operating state of the temperature control instrument control system according to the temperature control response strategy specifically means: Based on the temperature control response strategy, specific control commands and target parameters are generated through a preset parsing protocol; Based on the control commands and target parameters, the various execution modules in the temperature control instrument control system are adjusted.

[0087] The specific control commands and target parameters are generated through a preset parsing protocol, specifically as follows: Input the control requirements and target indicators in the temperature control response strategy into the protocol parsing module; Based on the execution module corresponding to the strategy, determine the module address and generate the corresponding instruction frame header; The instruction code is determined based on the control type, such as heat dissipation control, heating control, or sensor sampling control; Map the target index values ​​in the temperature control response strategy to the corresponding parameter fields; Perform parameter range validation on the instruction content and generate a complete frame structure; A check field is generated using CRC checksum to ensure the integrity of the command during transmission. Finally, control commands conforming to the parsing protocol format are generated and sent to the corresponding execution module along with the target parameters.

[0088] The system adjusts each execution module according to control commands and target parameters to achieve multi-module coordinated action and dynamic feedback. The core execution modules of the temperature control instrument control system include a heat dissipation module, a heating module, and a monitoring module. After receiving the corresponding command, each module completes the action through drive circuits and control logic. If multiple modules are involved in linkage, the main control unit will issue commands according to a preset timing sequence, while simultaneously receiving feedback data from the monitoring module in real time. If the actual cooling effect is found to be less than expected, the system will re-analyze the strategy to generate supplementary commands until the temperature stabilizes within the target range, forming a closed-loop control of command execution, data feedback, and dynamic adjustment to ensure that the temperature control system adapts to real-time changes in load and thermal effect.

[0089] During the closed-loop control process, the weights of the fuzzy rules are continuously optimized based on the temperature control execution feedback results. The weight update formula is as follows:

[0090] in, Apply fuzzy rule weighting coefficients to the updated temperature control; For the first Temperature control executes fuzzy rules at any time The weights; Adaptive learning rate; This is the adjustment factor calculated from the feedback error function.

[0091] Example 2: To verify the applicability and effectiveness of the method of the present invention, a reduced calculation system of a provincial power grid was selected as the research object, and the method described in Example 1 was used for analysis.

[0092] 1) Description of the case study scenario This scaled-down simulation system comprises 10 main nodes and 15 transmission lines. The operating scenarios cover three typical load periods: off-peak (0–6h), flat (6–18h), and peak (18–24h). The collected characteristic data include: node current, voltage amplitude, and cabinet temperature and humidity. Table 1 shows some examples of operating parameters. Table 1

[0093] The system data fully reflects the patterns of load and temperature changes over the daily cycle, providing a typical test scenario for verifying thermal effect prediction and temperature control strategy generation.

[0094] 2) Model building and thermal effect prediction Based on step S1 of Example 1, the periodic characteristics of the thermal effect are extracted from the nodal temperature and humidity data and environmental parameters. In this example, the period amplitude is set to 5°C, the reference temperature to 30°C, and the phase shift to 0.3 rad. The main period is determined using autocorrelation analysis, and the periodic variation function is obtained using a smooth fitting algorithm.

[0095] Based on step S2 of Example 1, the load dynamic feature vector is extracted from the node current, voltage, and cabinet temperature data in Table 2. This vector is then fused with the thermal effect periodic function using multi-source feature fusion to form a joint feature sequence. A load-thermal effect rhythm control model is constructed using a time-series regression algorithm, enabling dynamic prediction of future thermal effects.

[0096] Based on step S3 of Example 1, a fused feature matrix is ​​constructed and an LSTM input sequence is generated to predict the cabinet temperature for the next 24 hours. Table 2 shows the prediction results: Table 2

[0097] Error analysis showed that the maximum deviation between the predicted and measured values ​​did not exceed 0.3℃, verifying the high accuracy and foresight of the model.

[0098] Figure 3 This paper presents a comparison between the fitting results of the periodic function of the thermal effect and the temperature prediction results, aiming to verify the fitting accuracy of the periodic function of the thermal effect and the forward-looking capability of the LSTM prediction model. The graph shows a high degree of agreement between the predicted and measured values, with small errors, indicating that the model can accurately capture the periodic temperature variation trend of the switchgear and possesses good predictive performance.

[0099] 3) Generation and verification of temperature control response strategy Based on predicted temperature and real-time load data, a fuzzy set of temperature control status (low load / medium load / high load, temperature deviation, etc.) is constructed, and fuzzy inference is performed using a weighted fuzzy rule base. The centroid method is used to defuzzify the data, generating a specific temperature control response strategy. For example, see Table 3: Table 3

[0100] The strategy is sent to the temperature control instrument execution module to monitor temperature changes inside the cabinet. The results are displayed as follows: During peak load (18h) periods, the cabinet temperature is stabilized at the target value of 36±0.5℃ through a strategy of full-speed fan and auxiliary heat sink. During low load periods, the temperature control system automatically reduces the fan speed to achieve both energy saving and temperature control goals. In the closed-loop temperature control execution, the rule weights are dynamically updated based on feedback errors to optimize the adaptability of the strategy.

[0101] Figure 4 The temperature control performance of the intelligent temperature control method proposed in this invention and the traditional fixed threshold control method was compared in actual operation. The figures show the difference in temperature control between the two methods: the method of this invention provides smooth control with minimal temperature fluctuations; the traditional method exhibits large fluctuations and poses a risk of overheating. This comparison highlights the advantages of this invention: higher control precision, lower energy consumption, and fewer equipment operation cycles, thereby extending equipment lifespan and improving system reliability.

[0102] To quantify the advantages of this invention, the method of this invention is compared with the traditional threshold control method (strong cooling activated when temperature > 70℃: fan 100%, cooling chip 100%; shut off when temperature < 65℃) under the same computational example. The performance comparison between the present invention and the traditional method is shown in Table 4. Table 4

[0103] Control effect: The method of this invention keeps the cabinet temperature stably below 68.5℃ throughout the peak period without overshoot. In contrast, the traditional method, under the influence of load inertia, causes the temperature to rise continuously and trigger the 70℃ threshold multiple times, resulting in a maximum temperature of 72.5℃, which poses a risk of overheating.

[0104] Economy and equipment lifespan: The method of this invention, due to its early and smooth intervention, avoids frequent start-stop and long-term operation of the actuator at 100% power, thus effectively extending the equipment lifespan.

[0105] Quality control: The standard deviation of temperature fluctuation under the control of this invention is 1.2℃, which is much lower than the 3.5℃ of the traditional method, and the control process is more stable.

[0106] Example 3, Figure 2 The present invention provides a control system for a temperature control instrument in a switchgear, comprising a function construction module, a model construction module, a thermal effect prediction module, a strategy generation module, and a control execution module, wherein the modules are interconnected. The function construction module constructs a thermal effect periodic change function based on the first data of the switchgear obtained and a preset period length. The model building module constructs a load-thermal effect rhythm control model based on the periodic variation function of the thermal effect and the acquired current load data. The thermal effect prediction module, based on the control model, generates prediction results of the thermal effect of the switchgear through machine learning; The strategy generation module generates a temperature control response strategy based on the prediction results, combined with the current load level and environmental boundary conditions, through fuzzy logic reasoning. The control execution module dynamically adjusts the operating status of the temperature control instrument control system according to the temperature control response strategy.

[0107] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0108] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0109] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0110] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0111] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0112] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A control method for a temperature control instrument in a switchgear, characterized in that, Includes the following steps: Based on the first data of the switchgear obtained, a periodic change function of thermal effect is constructed with a preset period length; A load-thermal effect rhythm control model is constructed based on the periodic variation function of the thermal effect and the acquired current load data; Based on the control model, machine learning is used to generate prediction results of the thermal effect of the switchgear. Based on the prediction results and combined with real-time load status data, a temperature control response strategy is generated through fuzzy logic reasoning. Based on the temperature control response strategy, adjust the operating status of the temperature control instrument control system.

2. The control method for the temperature control instrument of the switchgear according to claim 1, characterized in that, The period length is determined by extracting the period of thermal effect changes through time series analysis.

3. The control method for the temperature control instrument of the switchgear according to claim 2, characterized in that, The step of constructing a thermal effect periodic variation function based on the first data of the switchgear with a preset period length involves inputting the first data of the switchgear into a smooth fitting algorithm, using the period length as a constraint, and performing data smooth fitting to obtain the thermal effect periodic variation function.

4. The control method for the temperature control instrument of the switchgear according to claim 3, characterized in that, The load-thermal effect rhythm control model constructed based on the periodic variation function of the thermal effect and current load data is as follows: Based on current load data, a dynamic feature vector of the load is obtained through a feature extraction algorithm; Based on the load dynamic feature vector and the thermal effect periodic change function, a joint feature sequence is obtained through a multi-source feature fusion algorithm. Based on the joint feature sequence, a load-heat effect rhythm regulation model is constructed using a time-series regression algorithm.

5. The control method for the temperature control instrument of the switchgear according to claim 4, characterized in that, The method of generating prediction results for the thermal effects of switchgear based on the control model and through machine learning is as follows: Based on the output of the load-heat effect rhythm regulation model, an LSTM input sequence is constructed; Based on the LSTM input sequence, a regular time series sample set is obtained through data standardization and time step alignment. Based on the time series sample set, the prediction results of the thermal effect of the switchgear are obtained through the Long Short-Term Memory Network algorithm.

6. The control method for the temperature control instrument of the switchgear according to claim 5, characterized in that, The LSTM input sequence is constructed based on the output of the load-heat effect rhythm regulation model, specifically as follows: Based on the output of the load-thermal effect rhythm control model, extract the load dynamic feature vector and the thermal effect feature vector; Based on the load dynamic feature vector and the thermal effect feature vector, a fused feature matrix is ​​obtained by feature concatenation; The LSTM input sequence is constructed based on the fused feature matrix using the sliding window sampling method.

7. The control method for the temperature control instrument of the switchgear according to claim 6, characterized in that, The process of generating a temperature control response strategy based on the prediction results and real-time load status data through fuzzy logic reasoning is as follows: Based on the prediction results and real-time load status data, a membership function is constructed to obtain the fuzzy set of temperature control status; Based on the acquired historical data of the switchgear, a weighted fuzzy rule base is generated; Based on the fuzzy set of temperature control status, a set of fuzzy temperature control strategies is obtained through reasoning using a weighted fuzzy rule base; Based on a set of fuzzy temperature control strategies, a temperature control response strategy is generated through defuzzification calculation.

8. The control method for the temperature control instrument of the switchgear according to claim 7, characterized in that, The step of generating a weighted fuzzy rule base based on the acquired historical data of the switchgear is as follows: Based on historical load and temperature data, typical operating conditions are obtained through clustering algorithms; Based on typical operating conditions, initial fuzzy rules are extracted using association rule mining methods. Based on historical performance data, the weight coefficients of each rule are calculated using a weight learning algorithm. Based on the initial fuzzy rules and weight coefficients, a weighted fuzzy rule library is generated by integration.

9. The control method for the temperature control instrument of the switchgear according to claim 8, characterized in that, The initial fuzzy rules are extracted based on typical operating conditions using association rule mining methods, specifically as follows: The features in typical operating conditions are discretized and fuzzy-classified to form a set of operating condition feature items. Based on the set of operating condition feature items, each operating condition sample is converted into an operating condition sample record, and an operating condition sample database is constructed. Calculate the support of the set of working condition feature items in the working condition sample database, and filter out frequent itemsets by setting a minimum support threshold. Based on frequent itemsets, an association rule set is generated according to the partitioning rules of input and output items; The working condition items in the association rule set are mapped to the corresponding fuzzy linguistic variables to form an initial fuzzy rule set with an "if-then" structure.

10. A system for controlling a temperature control instrument for a switchgear as described in any one of claims 1-9, characterized in that, include: The function construction module constructs a thermal effect periodic change function based on the first data of the switchgear obtained and a preset period length. The model building module constructs a load-thermal effect rhythm control model based on the periodic variation function of the thermal effect and the acquired current load data. The thermal effect prediction module, based on the control model, generates prediction results of the thermal effect of the switchgear through machine learning; The strategy generation module generates a temperature control response strategy based on the prediction results and real-time load status data through fuzzy logic reasoning. The control execution module regulates the operating status of the temperature control instrument control system according to the temperature control response strategy.

Images