A method and system for predicting condensation of an electrical cabinet

Abstract

This invention discloses a condensation prediction and control method and system for electrical cabinets, relating to the field of electrical cabinet technology. The condensation prediction and control method includes the following steps: obtaining environmental characteristic data; establishing a condensation prediction model using the obtained environmental characteristic data to obtain condensation trends; updating key factor combinations as key factors affecting condensation inside explosion-proof electrical cabinets; and adjusting condensation prevention measures based on comparison results. This invention, by constructing a condensation prediction model, can predict and prevent condensation formation in a timely manner, reducing electrical faults caused by condensation and thus significantly improving the safe operation of electrical cabinets. The key influencing factors identified through trend analysis algorithms make maintenance work more targeted and efficient. Continuous comparison and updating of influencing factors ensures that condensation control measures can quickly respond to environmental changes and emerging risks.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electrical cabinets, in particular to a condensation prediction control method and system for an electrical cabinet. BACKGROUND

[0002] The internal space of the explosion-proof electrical cabinet is usually compact and narrow, and the spacing between electrical components is small. In particular, in humid seasons such as summer, the water vapor content in the air is high, and these water vapors will penetrate into the explosion-proof electrical cabinet, causing the internal humidity to increase. The increase in humidity makes it easy for insulating parts, live parts, and PE inner surfaces to form a water film. The formation of the water film not only accelerates the corrosion of the internal metal body, but also causes the oxidation of the contact surface, thereby further affecting the performance of the electrical cabinet. More seriously, the presence of the water film significantly reduces the electrical insulation creepage distance between electrical components, weakening the electrical insulation strength, which can cause flashover discharge, inter-phase short circuit, and ground short circuit failures, which can easily cause the explosion-proof electrical cabinet to trip. These problems not only threaten the safe and stable operation of the equipment, but also can cause serious safety hazards. Therefore, reasonable control of the humidity inside the explosion-proof electrical cabinet and taking effective moisture-proof and dehumidification measures are the key to ensuring its normal operation and prolonging the service life of the equipment. Through these measures, electrical failures caused by humidity can be significantly reduced, ensuring the safety and reliability of the electrical system.

[0003] In the prior art, it is inconvenient to predict the condensation state in the explosion-proof electrical cabinet, which makes it inconvenient to predict and prevent the formation of condensation in a timely manner, thereby increasing electrical failures caused by condensation, aggravating the aging of components in the electrical cabinet, reducing the service life, and making it inconvenient to understand the condensation trend, making it difficult to identify key factors affecting condensation, making maintenance work difficult, reducing the accuracy of maintenance work, and making it inconvenient to continuously compare and update the influencing factors, which cannot ensure that the condensation control measures can quickly respond to environmental changes and new risks, and thus cannot optimize the prevention and control measures, increasing the risk of accidents and potential damage caused by condensation.

[0004] In view of the problems in the related art, no effective solution has been proposed so far. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a condensation prediction control method and system for an electrical cabinet, which solves the problem that the existing condensation state in the explosion-proof electrical cabinet cannot be conveniently predicted, so that the formation of condensation cannot be conveniently predicted and prevented, thereby increasing electrical faults caused by condensation, aggravating the aging of components in the electrical cabinet, reducing the service life, and making it inconvenient to understand the condensation trend, so that the key factors affecting condensation cannot be conveniently identified, making the maintenance work difficult, reducing the accuracy of the maintenance work, and at the same time, it is inconvenient to continuously compare and update the influencing factors, so that the condensation control measures cannot quickly respond to environmental changes and newly emerging risks, thereby failing to optimize the prevention and control measures in a targeted manner, and increasing the risk of accidents and potential damage caused by condensation.

[0006] To achieve the above object, the present application is implemented by the following technical solutions:

[0007] According to one aspect of the present application, a condensation prediction control method for an electrical cabinet is provided, which comprises the following steps:

[0008] S1, obtaining real-time environmental data in the explosion-proof electrical cabinet and preprocessing the real-time environmental data to obtain environmental feature data;

[0009] S2, establishing a condensation prediction model using the obtained environmental feature data, and performing time series analysis on the environmental feature data in the current explosion-proof electrical cabinet, based on the time series analysis result, obtaining the condensation trend;

[0010] S3, based on the condensation trend, randomly generating an initial key factor combination and its weight, calculating the condensation trend fitting degree of each key factor combination, determining the current optimal key factor combination, presetting a conversion probability, if the conversion probability is greater than a random number, updating the weight for global search; otherwise, fine-tuning the weight in the neighborhood of the current optimal key factor combination for local optimization, evaluating the new key factor combination fitting degree, if it is better, updating the key factor combination as the key factor affecting the condensation in the explosion-proof electrical cabinet;

[0011] S4, comparing the identified key factors affecting the condensation in the explosion-proof electrical cabinet with known factors affecting the condensation in the explosion-proof electrical cabinet, and adjusting the condensation prevention measures according to the comparison result;

[0012] S2, establishing a condensation prediction model using the obtained environmental feature data, and performing time series analysis on the environmental feature data in the current explosion-proof electrical cabinet, based on the time series analysis result, obtaining the condensation trend comprises the following steps:

[0013] S21, obtaining the environmental feature data in the current explosion-proof electrical cabinet and recording the historical changes of the environmental feature data;

[0014] S22, time series analysis is performed on the environmental characteristic data in the current explosion-proof electrical cabinet, and the periodic characteristics of the environmental characteristic data are extracted;

[0015] S23, based on the periodic characteristics of the environmental characteristic data, a condensation prediction model is used to predict the condensation state in the electrical cabinet at future time, and a condensation trend is obtained.

[0016] Further, real-time environmental data in the explosion-proof electrical cabinet is obtained, and the real-time environmental data is preprocessed to obtain environmental characteristic data including the following steps:

[0017] S11, obtaining each environmental data in the explosion-proof electrical cabinet, and calculating the average value of each environmental data;

[0018] S12, the average value of each environmental data is calculated to obtain the deviation value of each environmental data from the average value;

[0019] S13, the standard deviation value of each environmental data is calculated by the standard deviation formula;

[0020] S14, if the deviation value of each environmental data is greater than the standard deviation value, it is determined as abnormal data, and is eliminated to obtain an accurate data set;

[0021] S15, environmental feature extraction is performed on the accurate data set to obtain environmental characteristic data.

[0022] Further, based on the periodic characteristics of the environmental characteristic data, a condensation prediction model is used to predict the condensation state in the electrical cabinet at future time, and a condensation trend is obtained including the following steps:

[0023] S231, the time series stationarity of the environmental characteristic data is tested by using the test method;

[0024] S232, autocorrelation and partial autocorrelation analysis are performed on the stationary environmental characteristic data time series, and the best lag order of the environmental characteristic data is determined;

[0025] S233, the residual sequence in the environmental characteristic data is tested by using a statistical method;

[0026] S234, combining the stationarity test result of the time series, the determined best lag order and the residual sequence, the autoregressive order and the moving average order are selected, and a condensation prediction model is constructed;

[0027] S235, the condensation prediction model is used to predict the condensation state in the electrical cabinet at future time, and a condensation trend is obtained.

[0028] Further, the formula for testing the residual sequence in the environmental characteristic data by using a statistical method is:

[0029] ;

[0030] wherein, T is expressed as a test statistic;

[0031] M is expressed as a sample size;

[0032] is expressed as an autocorrelation coefficient value of the first S

[0033] L is expressed as a number of autocorrelation coefficients.

[0034] Further, based on the condensation trend, an initial key factor combination and its weight are randomly generated, the condensation trend fitting degree of each key factor combination is calculated, the current optimal key factor combination is determined, a conversion probability is preset, if the conversion probability is greater than a random number, the weight is updated to perform global search, otherwise the weight is fine-tuned in the neighborhood of the current optimal key factor combination to perform local optimization, the fitting degree of the new key factor combination is evaluated, if it is better, the key factor combination is updated as the key factor affecting the condensation in the explosion-proof electrical cabinet, and the method comprises the following steps.

[0035] S31, parameters of the trend analysis algorithm are initialized, a maximum iteration number and a preset conversion probability are set, an initial condensation key factor set is randomly generated, and initial influence weights of each factor are set;

[0036] S32, the fitting degree of each factor combination to the condensation trend is calculated, the current optimal key factor combination and the corresponding fitting degree value are identified, and a preliminary condensation influence model is established;

[0037] S33, if the conversion probability is greater than a random number, the weight of the key factor combination is adjusted by using a strategy adjustment algorithm to expand the global search range of the condensation key factor;

[0038] S34, if the conversion probability is less than or equal to the random number, local optimization is performed in the neighborhood of the current optimal key factor combination, and the weight of the key factor combination is fine-tuned by uniform distribution randomization;

[0039] S35, the fitting degree of the newly generated key factor is evaluated, if the new key factor combination is better than the current key factor combination, the key factor combination is updated, otherwise the original analysis result is retained to ensure continuous optimization of the condensation trend analysis;

[0040] S36, whether the maximum iteration number is reached is checked, if yes, the optimal key factor combination is output as the key factor affecting the condensation in the explosion-proof electrical cabinet, otherwise, the iterative analysis is continued.

[0041] ​Further, if the conversion probability is greater than the random number, a strategy adjustment algorithm is used to adjust the weight of the key factor combination to expand the global search range of the condensation key factor, including the following steps:

[0042] S331, if the conversion probability is greater than the random number, initialize the parameters of the strategy adjustment algorithm, set the maximum number of iterations and the constraint condition, randomly generate a key factor combination and its initial weight in the defined parameter space as the starting point of the current condensation trend analysis;

[0043] S332, randomly assign initial values to each key factor affecting the condensation state to form several groups of preliminary condensation trend analysis data, and evaluate and determine the current best key factor combination according to the pros and cons of the trend fitting degree;

[0044] S333, verify whether the newly generated key factor combination meets all the preset constraint conditions, if not, cancel the newly generated key factor combination, and try to evaluate other key factor combinations until a key factor combination that meets all the preset constraint conditions is obtained;

[0045] S334, repeatedly execute factor search and combination verification to continuously expand and optimize the key factor combination, and when the trend fitting degree reaches the local optimum, the current key factor combination is temporarily determined as the best key factor combination;

[0046] S335, introduce random factor disturbance to adjust the weight configuration of the temporarily determined current best key factor combination, and search again to obtain a new key factor combination;

[0047] S336, compare the trend fitting degrees of the new and old key factor combinations, if the new key factor combination is better, update it as the global optimal key factor combination, otherwise, output the current global optimal key factor combination.

[0048] Further, the formula for adjusting the weight configuration of the temporarily determined current best key factor combination by introducing random factor disturbance is:

[0049] ;

[0050] In the formula, R new represents the new key factor combination vector after random factor disturbance and weight configuration adjustment; R best represents the temporarily determined current best key factor combination vector; Δ a represents the random disturbance coefficient of the a th key factor; M a represents the random disturbance vector of the a th key factor; αa a weight adjustment coefficient representing the weight of the a Q a a weight adjustment vector representing the weight of the a ω a a weight vector representing the weight of the a

[0051] Further, the identified key factors affecting the condensation in the explosion-proof electrical cabinet are compared with known factors affecting the condensation in the explosion-proof electrical cabinet, and based on the comparison result, the condensation prevention measures are adjusted, including the following steps:

[0052] S41, set a data set of key factors affecting the condensation in the explosion-proof electrical cabinet, and preset the number of variation and crossover operations;

[0053] S42, perform a variation operation on the data set of key factors affecting the condensation in the explosion-proof electrical cabinet, determine the key factor data in the data set that has the greatest impact on the condensation by calculating the difference vector between each key factor data;

[0054] S43, perform a crossover operation on the new key factor data affecting the condensation in the explosion-proof electrical cabinet, and determine the key factor data combination that has the most significant impact on the condensation;

[0055] S44, use a clustering analysis method to perform new grouping on the varied key factor data, determine the cluster where each key factor data is located, and find the best correlation distance between different clusters;

[0056] S45, compare the key factor data obtained through variation and crossover with the preset target function, and select the key factor data with the lowest target function as the key factor data affecting the condensation in the explosion-proof electrical cabinet in the new key factor data set;

[0057] S46, judge whether the iteration termination condition is met, if yes, terminate the evolution, take the current best key factor data set as the key factor affecting the condensation in the explosion-proof electrical cabinet, and adjust the condensation prevention measures.

[0058] Further, the varied key factor data is grouped using a clustering analysis method to determine the cluster where each key factor data is located, and the best correlation distance between different clusters is found, including the following steps:

[0059] S441, randomly select several key factor data from the varied key factor data set as the center points of the initial clusters;

[0060] ​​​S442, allocate each unallocated key factor data to the cluster center point with the highest similarity;

[0061] S443, randomly select a key factor data in the key factor data set that does not belong to the cluster center as a potential new cluster center candidate;

[0062] S444, calculate the total cost change brought by replacing the current cluster center point with the randomly selected key factor data, and update the cluster center set;

[0063] S445, when the cluster center point no longer changes, determine the final cluster attribution of each key factor data, and calculate the optimal correlation distance between different clusters.

[0064] According to another aspect of the present application, there is also provided a condensation prediction control system of an electrical cabinet, comprising:

[0065] A data acquisition module is configured to acquire real-time environmental data in the explosion-proof electrical cabinet, and pre-process the real-time environmental data to obtain environmental feature data.

[0066] A model establishment module is configured to establish a condensation prediction model using the obtained environmental feature data, and perform time series analysis on the environmental feature data in the current explosion-proof electrical cabinet, and obtain a condensation trend based on the time series analysis result.

[0067] A trend analysis module is configured to randomly generate an initial key factor combination and its weight based on the condensation trend, calculate a condensation trend fitting degree of each key factor combination, determine a current optimal key factor combination, preset a conversion probability, and if the conversion probability is greater than a random number, update the weight to perform global search; otherwise, fine-tune the weight in a neighborhood of the current optimal key factor combination to perform local optimization, evaluate a new key factor combination fitting degree, and if the new key factor combination fitting degree is better, update the key factor combination as a key factor affecting condensation in the explosion-proof electrical cabinet.

[0068] An adjustment control module is configured to compare the identified key factor affecting condensation in the explosion-proof electrical cabinet with known factors affecting condensation in the explosion-proof electrical cabinet, and adjust condensation prevention measures according to a comparison result.

[0069] The data acquisition module is connected to the model establishment module and the trend analysis module, and the trend analysis module is connected to the adjustment control module.

[0070] The present application has the following advantages:

[0071] 1、The present application can predict and prevent the formation of condensation in time by real-time monitoring and preprocessing of environmental data in the explosion-proof electrical cabinet, reducing electrical failures caused by condensation, thereby significantly improving the safe operation of the electrical cabinet. The key influencing factors identified by the trend analysis algorithm can make the maintenance work more targeted and efficient. The continuous comparison and update of the influencing factors ensure that the condensation control measures can quickly respond to environmental changes and new risks. Through comparison and analysis, the most critical condensation influencing factors can be accurately identified, thereby optimizing the prevention and control measures. This not only improves the protection effect of the explosion-proof electrical cabinet, thereby effectively reducing accidental risks and potential damage caused by condensation, but also improves the accuracy of condensation prediction and control in the explosion-proof electrical cabinet, and helps to achieve more efficient and economic operation and management, ensuring the long-term stability of the equipment and system.

[0072] 2、The condensation prediction model can more accurately predict the future condensation state by time series analysis and periodic feature extraction of environmental characteristic data, effectively identify and utilize hidden patterns and periodic changes in the data, thereby improving the accuracy of the prediction. The optimal lag order determined by autocorrelation and partial autocorrelation plot analysis and the analysis of stationary test and residual sequence help to understand the internal structure and dynamic changes of the data, providing support for building a more effective condensation prediction model. Through accurate prediction and effective preventive measures, maintenance costs and equipment failures caused by condensation can be significantly reduced, equipment life can be extended, downtime can be reduced, and overall economic benefits can be improved.

[0073] 3、The present application realizes the accurate analysis of condensation trend by dynamically adjusting the weight of key factor combination. This method uses conversion probability mechanism to adaptively switch global search and local optimization strategy, effectively avoiding the limitation of traditional method easily falling into local optimum. By iteratively updating the key factor combination, the accuracy and adaptability of the condensation prediction model are significantly improved, providing a reliable technical basis for the condensation warning and active protection of the explosion-proof electrical cabinet, enhancing the safety and stability of equipment operation. BRIEF DESCRIPTION OF DRAWINGS

[0074] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0075] Figure 1 is a flow chart of a condensation prediction control method for an electrical cabinet according to an embodiment of the present application;

[0076] Figure 2 is a principle block diagram of a condensation prediction control system of an electrical cabinet according to an embodiment of the present application.

[0077] In the figure:

[0078] 1, data acquisition module; 2, model establishment module; 3, trend analysis module; 4, adjustment control module. DETAILED DESCRIPTION

[0079] The technical solutions in the embodiments of the present application will be clearly and completely described in combination with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments.

[0080] In the description of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more. In addition, the terms "first", "second", "third" and the like are only for the purpose of description, and cannot be understood as indicating or implying relative importance.

[0081] According to an embodiment of the present application, a condensation prediction control method and system of an electrical cabinet are provided.

[0082] The present application will be further described in combination with the accompanying drawings and specific embodiments, as shown, the condensation prediction control method of the electrical cabinet according to the embodiment of the present application, the condensation prediction control method of the electrical cabinet comprises the following steps: Figure 1

[0083] S1, obtaining real-time environmental data in the explosion-proof electrical cabinet, and preprocessing the real-time environmental data to obtain environmental feature data;

[0084] Specifically, the real-time environmental data is obtained by temperature sensor, humidity sensor, dew point sensor, air pressure sensor and the like, and the real-time environmental data includes temperature, humidity, air pressure, dew point and the like.

[0085] S2, using the obtained environmental feature data to establish a condensation prediction model, and performing time series analysis on the environmental feature data in the current explosion-proof electrical cabinet, and based on the time series analysis result, obtaining a condensation trend;

[0086] Specifically, the condensation state refers to the condensation of water vapor on the inner surface of the electrical cabinet, including the existence or nonexistence of condensation, the degree of condensation, the distribution of condensation, the duration of condensation and the like.

[0087] Specifically, the condensation trend refers to the prediction of the condensation state changing with time, including the time point of condensation formation, the development speed and intensity of condensation, the subsidence of condensation and the like.

[0088] ​S3, based on the condensation trend, randomly generate an initial key factor combination and its weight, calculate the condensation trend fitting degree of each key factor combination, determine the current optimal key factor combination, preset the conversion probability, if the conversion probability is greater than the random number, update the weight to perform global search; otherwise, fine-tune the weight in the neighborhood of the current optimal key factor combination to perform local optimization, evaluate the fitting degree of the new key factor combination, if better, update the key factor combination as the key factor affecting the condensation in the explosion-proof electrical cabinet;

[0089] Specifically, the key factors affecting the condensation in the explosion-proof electrical cabinet include temperature difference, humidity, ventilation and airflow, temperature fluctuation, equipment operation heat production, external environmental factors, and the sealing of the electrical cabinet.

[0090] S4, compare the identified key factors affecting the condensation in the explosion-proof electrical cabinet with known factors affecting the condensation in the explosion-proof electrical cabinet, and adjust the condensation prevention measures according to the comparison result;

[0091] Specifically, adjusting the condensation prevention measures includes improving ventilation, temperature and humidity control, improving sealing, heat insulation and temperature insulation materials, regular inspection and maintenance, etc.

[0092] A condensation prediction model is established using the obtained environmental characteristic data, and time series analysis is performed on the current environmental characteristic data in the explosion-proof electrical cabinet, and based on the time series analysis result, the condensation trend is obtained, including the following steps:

[0093] S21, obtain the environmental characteristic data in the current explosion-proof electrical cabinet, and record the historical changes of the environmental characteristic data;

[0094] It needs to be explained that, for example, in an explosion-proof electrical cabinet in a chemical plant, temperature and humidity sensors are installed to monitor the internal environment. These sensors record data every 15 minutes and send the data to the central monitoring system. This data includes real-time temperature, humidity, dew point and external environmental conditions in the cabinet.

[0095] S22, perform time series analysis on the environmental characteristic data in the current explosion-proof electrical cabinet, and extract the periodic characteristics of the environmental characteristic data;

[0096] It needs to be explained that after collecting enough historical data, time series analysis techniques such as autoregressive moving average (ARMA) model or Fourier transform are used to analyze the periodic changes of the data. For example, the analysis shows that the temperature is generally lower in winter every year, and the humidity change shows a clear periodicity of day and night, which may be due to the change of the plant operation mode.

[0097] S23, based on the periodic characteristics of the environmental characteristic data, the condensation prediction model is used to predict the condensation state in the electrical cabinet at future time, and the condensation trend is obtained.

[0098] It needs to be explained that based on the analysis results, the future condensation state is predicted by a machine learning model, such as random forest or neural network, etc. This model will input historical and current environmental feature data and output the condensation state prediction at a certain time in the future, such as predicting the condensation risk in the next 24 hours.

[0099] Preferably, real-time environmental data in the explosion-proof electrical cabinet is obtained, and the real-time environmental data is preprocessed to obtain environmental feature data, including the following steps:

[0100] S11, obtaining each environmental data in the explosion-proof electrical cabinet, and calculating the average value of each environmental data;

[0101] Specifically, the formula for calculating the average value of each environmental data is:

[0102] ;

[0103] In the formula, represents the average value of each environmental data collected by the sensor;

[0104] represents the environmental data collected by the i-th sensor; j

[0105] n represents the number of sensors.

[0106] It needs to be explained that, taking the explosion-proof electrical cabinet of a chemical plant as an example, for example, an explosion-proof electrical cabinet of a chemical plant equipped with multiple sensors, these sensors continuously monitor temperature, humidity and dew point parameters. Every 15 minutes, the readings of these sensors are automatically recorded and uploaded to the central processing system. The processing system averages the collected data to obtain the average value of each parameter in 15 minutes.

[0107] S12, calculating the deviation value of each environmental data from the average value using the average value of each environmental data;

[0108] Specifically, the formula for calculating the deviation value of each environmental data from the average value is:

[0109] ;

[0110] In the formula, represents the deviation value of each environmental data collected by the sensor from the average value;

[0111] represents the environmental data collected by the i-th sensor; j ​​

[0112] denotes the average value of the environmental data collected by the sensor.

[0113] It should be explained that the system compares the environmental data at each time point with the corresponding average value to calculate the deviation value. For example, if the temperature reading at a certain time point is higher than the average temperature within 15 minutes, this deviation value will be recorded.

[0114] S13, calculate the standard deviation value of each environmental data by the standard deviation formula;

[0115] Specifically, the standard deviation formula is:

[0116] ;

[0117] In the formula, denotes the standard deviation value of the environmental data collected by the sensor.

[0118] denotes the average value of the environmental data collected by the sensor.

[0119] denotes the environmental data collected by the i-th sensor; j

[0120] n denotes the number of sensors.

[0121] It should be explained that the system uses the standard deviation formula to calculate the standard deviation of each environmental data, which represents the variation degree of each data point relative to the average value. For example, if the standard deviation of a certain parameter is large, it indicates that the parameter changes greatly during the monitoring period.

[0122] S14, if the deviation value of each environmental data is greater than the standard deviation value, it is determined as abnormal data and is removed to obtain an accurate data set;

[0123] It should be explained that the system compares the deviation value of each data point with the corresponding standard deviation. If the deviation value of a certain data point exceeds a certain multiple (for example, 2 times) of the standard deviation, it is usually considered as abnormal. These abnormal data will be removed from the data set to ensure the accuracy of subsequent analysis.

[0124] S15, extract environmental feature data from the accurate data set.

[0125] It should be explained that for the accurate data set that has been filtered and calibrated, the system further extracts key environmental features. These features will be used to establish a condensation prediction model to help predict future environmental changes and condensation risks.​

[0126] Preferably, based on the periodicity of the environmental feature data, the condensation state in the electrical cabinet at a future time is predicted using a condensation prediction model, and the condensation trend is obtained by the following steps:

[0127] S231, test the time series stationarity of the environmental feature data using a test method (i.e. Q-statistic test);

[0128] It needs to be explained that, taking the explosion-proof electrical cabinet of a chemical plant as an example, for example, continuous temperature and humidity data of several months are collected from the sensors of the explosion-proof electrical cabinet. The Q-statistic test is used to test the stationarity of these time series. If the data is non-stationary, the data needs to be differenced or transformed, such as logarithmic transformation, to achieve stationarity.

[0129] Specifically, the Q-statistic test is a statistical test used in time series analysis, mainly used to test whether the autocorrelation of the time series is significantly different from zero, i.e. to test the randomness of the time series. This test is particularly suitable for determining whether a time series is a white noise sequence. If the time series is white noise, there is no autocorrelation between its time points, meaning that the data points in the sequence are randomly and independently distributed.

[0130] S232, autocorrelation and partial autocorrelation analysis of the stationarized environmental feature data time series, and determination of the best lag order of the environmental feature data;

[0131] It needs to be explained that, for the stationarized data, autocorrelation function (ACF) and partial autocorrelation function (PACF) are drawn to identify the autocorrelation structure in the data, in order to determine the best lag order in the condensation prediction model. For example, if the partial autocorrelation function is truncated at the first lag, and the autocorrelation function gradually decays, it indicates that the condensation prediction model may be appropriate.

[0132] S233, test the residual sequence in the environmental feature data using a statistical method;

[0133] It needs to be explained that, a statistical method (such as white noise test) is used to analyze and confirm the residual sequence used to establish the condensation prediction model. Ensuring that the residual sequence is close to white noise indicates that the condensation prediction model has fully extracted the information in the data.

[0134] S234, select the autoregressive order and the moving average order to construct the condensation prediction model, combined with the stationarity test result of the time series, the determined best lag order and the residual sequence;

[0135] It needs to be explained that based on the determined optimal lag order and the tested residual series, the appropriate autoregressive order (AR) and moving average order (MA) are selected to construct the condensation prediction model (i.e. autoregressive integrated moving average model).

[0136] S235, using the constructed condensation prediction model to predict the condensation state in the electrical cabinet at future time, obtaining the condensation trend.

[0137] It needs to be explained that using the constructed condensation prediction model, the environmental feature data is input into the condensation prediction model to predict the condensation state in the explosion-proof electrical cabinet within a certain period of time in the future. The prediction result can help the factory managers to adjust the environmental control system to prevent the damage caused by condensation.

[0138] The specific autoregressive integrated moving average (ARIMA) model is a statistical model widely used in time series data prediction, especially suitable for data with trends or seasonality. ARIMA model combines autoregressive (AR), moving average (MA) and difference integration (I) to describe the changes of time series data, so as to predict future values.

[0139] Preferably, the formula for testing the residual series in the environmental feature data by statistical method is:

[0140] ;

[0141] In the formula, T is the test statistic;

[0142] M is the sample size;

[0143] Specifically, the sample size refers to the total number of data points used for prediction in the process of establishing the condensation prediction model.

[0144] is the autocorrelation coefficient value of the S th residual series;

[0145] L is the number of autocorrelation coefficients.

[0146] Preferably, based on the condensation trend, an initial key factor combination and its weight are randomly generated, the condensation trend fitting degree of each key factor combination is calculated, the current optimal key factor combination is determined, the conversion probability is preset, if the conversion probability is greater than the random number, the weight is updated for global search; otherwise, the weight is fine-tuned for local optimization in the neighborhood of the current optimal key factor combination, the new key factor combination fitting degree is evaluated, if it is better, the key factor combination is updated as the key factor affecting the condensation in the explosion-proof electrical cabinet, including the following steps:

[0147] S31, initialize parameters of the trend analysis algorithm, set the maximum number of iterations and the preset conversion probability, randomly generate an initial set of condensation key factors, and set the initial influence weight of each factor;

[0148] S32, calculate the fitting degree of each factor combination to the condensation trend, identify the current optimal key factor combination and its corresponding fitting degree value, and establish a preliminary condensation influence model;

[0149] S33, if the conversion probability is greater than the random number, adjust the weight of the key factor combination using the strategy adjustment algorithm to expand the global search range of the condensation key factors;

[0150] S34, if the conversion probability is less than or equal to the random number, perform local optimization within the neighborhood of the current optimal key factor combination, and adjust the weight of the key factor combination by uniformly distributed random;

[0151] S35, evaluate the fitting degree of the newly generated key factor, if the new key factor combination is better than the current key factor combination, update the key factor combination; otherwise, keep the original analysis result to ensure continuous optimization of the condensation trend analysis;

[0152] S36, check if the maximum number of iterations is reached, if so, output the optimal key factor combination as the key factor affecting the condensation in the explosion-proof electrical cabinet, otherwise, continue the iterative analysis.

[0153] Specifically, the system is initialized, the basic running parameters of the algorithm are set, including the maximum number of iterations and the conversion probability threshold, on this basis, multiple sets of initial key factor combinations are randomly generated, and each factor is assigned an initial weight value, establishing a preliminary basis for condensation trend analysis. After entering the iterative optimization phase, the system calculates the condensation trend fitting degree of each factor combination, and selects the combination with the best fitting degree as the current baseline model. In each iteration process, the system will generate a random number and compare it with the preset conversion probability to determine: when the global search condition is met, the strategy adjustment algorithm is used to update the factor weight greatly to expand the search range and explore new potential advantage area; when the global search condition is not met, small-scale fine adjustment is performed in the neighborhood of the current optimal combination, and the accuracy of the solution is improved through local optimization. After generating a new key factor combination, the system will immediately evaluate its condensation trend fitting degree. If the fitting degree of the new combination is better than the current optimal solution, update the optimal combination record in time; otherwise, keep the original optimization result to ensure continuous improvement of the analysis process. At the same time, the system will verify the new combination for constraint conditions to ensure that it meets various physical limitations and operating requirements. After sufficient iteration, the algorithm automatically outputs the optimal key factor combination as the key factor affecting the condensation in the explosion-proof electrical cabinet.

[0154] Specifically, the trend analysis algorithm is a flower pollination algorithm, which is a kind of swarm intelligence optimization algorithm simulating the pollination behavior of plants. The algorithm realizes the optimization target by simulating the global search of cross-pollination and the local development of self-pollination. In the application, the algorithm is applied to the identification of the key factors of condensation: the conversion probability mechanism controls the adaptive switching between global weight update and local fine-tuning, and through iterative evaluation of the fitting degree of factor combination, the key factor set is gradually optimized, and finally an accurate condensation prediction model is established, providing a reliable basis for the condensation warning of the explosion-proof electrical cabinet.

[0155] Preferably, if the conversion probability is greater than the random number, the weight of the key factor combination is adjusted by using the strategy adjustment algorithm to expand the global search range of the key factors of condensation, including the following steps:

[0156] S331, if the conversion probability is greater than the random number, the parameters of the strategy adjustment algorithm are initialized, the maximum number of iterations and the constraint conditions are set, and a key factor combination and its initial weight are randomly generated in the defined parameter space as the starting point of the current condensation trend analysis;

[0157] S332, randomly assign initial values to each key factor affecting the condensation state to form several groups of preliminary condensation trend analysis data, and evaluate and determine the current best key factor combination according to the pros and cons of the trend fitting degree;

[0158] S333, verify whether the newly generated key factor combination meets all the preset constraint conditions, if not, cancel the newly generated key factor combination, and try to evaluate other key factor combinations until a key factor combination that meets all the preset constraint conditions is obtained;

[0159] S334, repeatedly execute factor search and combination verification to continuously expand and optimize the key factor combination, and when the trend fitting degree reaches the local optimum, the current key factor combination is temporarily determined as the best key factor combination;

[0160] S335, introduce random factor disturbance to adjust the weight configuration of the temporarily determined current best key factor combination, and search again to obtain a new key factor combination;

[0161] S336, compare the trend fitting degrees of the new and old key factor combinations, if the new key factor combination is better, update it to the global optimal key factor combination, otherwise, output the current global optimal key factor combination.

[0162] Specifically, the system is initialized, the maximum number of iterations of the algorithm and the threshold of the conversion probability are set. A plurality of sets of initial key factor combinations are randomly generated in the parameter space, and initial weight values are assigned to each factor. The iteration optimization stage is entered. The condensation trend fitting degree of each set of factor combinations is calculated, and the combination with the optimal fitting degree is selected as the current benchmark. In each iteration, a random number is compared with the conversion probability: if the global search condition is met, the strategy adjustment algorithm is used to adjust the factor weights and expand the search range; otherwise, local fine adjustment is performed in the neighborhood of the current optimal combination. In the strategy adjustment process, the greedy algorithm is used for fast optimization. A new key factor combination is generated, and then it is verified whether it meets the various constraint conditions. The condensation trend fitting degree of the combination that meets the requirements is calculated, and if it is better than the current optimal solution, it is immediately updated. By repeatedly executing this process, the key factor combinations are continuously expanded and optimized. To avoid falling into local optimization, when it is detected that the fitting degree has not improved for multiple iterations, a random disturbance mechanism is introduced. By randomly adjusting the weights of some factors in the current optimal combination, the local balance state is broken, and the search process is restarted. After each new combination is obtained, it is compared with the current global optimal solution, and the combination with the better fitting degree is retained. After sufficient iterations, the algorithm outputs the current global optimal key factor combination.

[0163] Specifically, the strategy adjustment algorithm is a greedy algorithm, which is a local optimization algorithm that selects the current optimal solution at each step. The core of this algorithm is to achieve efficient optimization through fast local search. In the present application, this algorithm is used for strategy adjustment of the weights of the key factors: by verifying and selecting the current optimal key factor combination, and prioritizing the combination with higher fitting degree in each iteration, combined with random disturbance to jump out of local optimization, the weights of the key factors are quickly optimized, and the accuracy and computational efficiency of the condensation trend analysis are effectively improved.

[0164] Preferably, a random factor disturbance is introduced, and the formula for adjusting the weight configuration of the provisional current best key factor combination is:

[0165] ;

[0166] In the formula, R new represents the new key factor combination vector after random factor disturbance and weight configuration adjustment; R best represents the provisional current best key factor combination vector; Δ a represents the random disturbance coefficient of the a th key factor; M a represents the random disturbance vector of the a th key factor; α a represents the random disturbance vector of the aWeight adjustment coefficients for key factors; Q a Indicates the first a Weight adjustment vectors for key factors; ω a Indicates the first a The weight vector of each key factor in the current optimal combination of key factors.

[0167] Preferably, comparing the identified key factors affecting condensation inside the explosion-proof electrical cabinet with known factors affecting condensation inside the explosion-proof electrical cabinet, and adjusting the condensation prevention measures based on the comparison results, includes the following steps:

[0168] S41. Set the dataset of key factors affecting condensation inside the explosion-proof electrical cabinet, and preset the number of mutation and cross operations;

[0169] It should be explained that, taking the explosion-proof electrical cabinet in a chemical plant as an example, key factors affecting condensation are collected and defined within the cabinet, forming the initial dataset. The number of iterations and the frequency of mutation and crossover operations in the genetic algorithm are set to ensure sufficient opportunities to find the optimal solution.

[0170] S42. Perform a mutation operation on the dataset of key factors affecting condensation inside the explosion-proof electrical cabinet. By calculating the difference vector between the data of each key factor, determine the key factor data in the dataset that has the greatest impact on condensation.

[0171] It should be explained that mutation operations are performed on the initial dataset, for example, by adding or changing the weights or values ​​of certain factors to simulate the impact of environmental changes. Difference vectors are calculated between the key factor data to identify which factors have the greatest impact on condensation.

[0172] S43. Cross-operate on the data of key factors affecting condensation inside the new explosion-proof electrical cabinet, and determine the combination of key factor data that has the most significant impact on condensation.

[0173] It should be explained that cross-referencing the mutated key factor data and combining it with the characteristics of different datasets generates new data combinations. This can reveal new, advantageous combinations that may more effectively reflect condensation control in actual operations.

[0174] S44. Use cluster analysis to regroup the mutated key factor data, determine the cluster to which each key factor data belongs, and find the best association distance between different clusters.

[0175] It should be explained that cluster analysis is used to regroup the cross-referenced data and identify datasets with similar characteristics. By analyzing the optimal association distance between different clusters, the optimal group for each key factor's data is determined.

[0176] S45, comparing the key factor data obtained through mutation and crossover with the preset objective function, and selecting the key factor data with the lowest objective function as the key factor data affecting the condensation in the explosion-proof electrical cabinet in the new key factor data set;

[0177] It should be explained that the key factor data obtained through mutation and crossover is compared with the preset objective function, and the key factor data that makes the objective function lowest is selected as the new key data set.

[0178] S46, judging whether the iteration termination condition is met, if yes, the evolution is terminated, and the current best key factor data set is selected as the key factor affecting the condensation in the explosion-proof electrical cabinet, and the condensation prevention measures are adjusted.

[0179] It should be explained that the iteration is repeated until the iteration termination condition is met. Finally, the optimal key factor data set obtained is used to adjust and optimize the condensation prevention measures.

[0180] Specifically, the identified key factors affecting the condensation in the explosion-proof electrical cabinet are compared with the known factors affecting the condensation in the explosion-proof electrical cabinet based on the K-medoids clustering algorithm of differential evolution. The idea of the K-medoids clustering algorithm of differential evolution is that a certain scale of population is randomly selected from the population, and the scale of the population is set as a fixed value. At the same time, the maximum scale limit of the population is set. The differential evolution algorithm (DE) is used to perform mutation and crossover operations on the population to determine the optimal mutation individual and experimental individual in the population, so as to avoid the population from falling into local optimum and enhance the local search ability of the algorithm in the clustering space region. According to the K-medoids clustering algorithm, new clustering analysis is performed on the optimized experimental individuals to determine the clustering of each individual and the optimal distance between them. Through this method, the individual with the lowest objective function value can be found, which is used as a member of the new population, and the optimal solution is finally output. The global search ability of the differential evolution algorithm (DE) and the local search efficiency of the K-medoids clustering algorithm in a specific space region are combined to achieve more efficient clustering analysis and feature extraction.

[0181] Specifically, the differential evolution algorithm (DE) is a population-based evolutionary algorithm with the characteristics of remembering the optimal solution of individuals and sharing information within the population, that is, the solution of the optimization problem is realized through the cooperation and competition between individuals in the population, and its essence is a greedy genetic algorithm with the idea of preserving the optimal solution based on real number coding.

[0182] Preferably, the new grouping of the mutated key factor data is performed using a clustering analysis method, the final cluster attribution of each key factor data is determined, and the optimal association distance between different clusters is calculated, including the following steps:

[0183] S441. Randomly select several key factor data from the mutated key factor data set as the initial cluster center points;

[0184] It should be explained that, taking the explosion-proof electrical cabinet of a chemical plant as an example, in the key factor data set of the explosion-proof electrical cabinet of the chemical plant, such as temperature, humidity, ventilation conditions, etc., several data points are randomly selected as the initial centers of the clusters. These center points represent various possible combinations of environmental conditions.

[0185] S442. Assign each unassigned key factor data to the cluster center point with the highest similarity;

[0186] It should be explained that for each unassigned data point in the data set, the similarity between it and each cluster center point is calculated, and it is assigned to the nearest cluster center. In this way, it can be ensured that each data point belongs to the cluster with the most matching characteristics.

[0187] S443. Randomly select a key factor data in the key factor data set that does not belong to the cluster center as a potential new cluster center candidate;

[0188] It should be explained that, under the condition that the existing cluster center is fixed, a data point in the key factor data set that does not belong to the current cluster center is randomly selected as a new cluster center candidate to explore possible improvement space.

[0189] S444. Calculate the total cost change caused by replacing the current cluster center point with the randomly selected key factor data, and update the cluster center set;

[0190] It should be explained that the total cost change caused by selecting the new candidate data point as the cluster center is calculated, and the cluster center set is updated accordingly. This process may involve accepting a better cluster center or keeping the original center.

[0191] S445. When the cluster center points no longer change, determine the final cluster attribution of each key factor data, and calculate the optimal association distance between different clusters.

[0192] It should be explained that when the cluster center points are stable and no longer change, the final cluster attribution of each key factor data is confirmed. The optimal association distance between different clusters is calculated, which helps to understand the relationship and interaction between different combinations of environmental factors.

[0193] According to another embodiment of the present application, as shown in Figure 2 There is also provided a condensation prediction control system of an electrical cabinet, comprising:

[0194] A data acquisition module 1 is configured to acquire real-time environmental data in the explosion-proof electrical cabinet, pre-process the real-time environmental data, and obtain environmental feature data.

[0195] A model establishment module 2 is configured to establish a condensation prediction model using the obtained environmental feature data, perform time series analysis on the environmental feature data in the current explosion-proof electrical cabinet, and obtain a condensation trend based on the time series analysis result.

[0196] A trend analysis module 3 is configured to randomly generate an initial key factor combination and its weight based on the condensation trend, calculate a condensation trend fitting degree of each key factor combination, determine a current optimal key factor combination, preset a conversion probability, and if the conversion probability is greater than a random number, update the weight to perform a global search; otherwise, fine-tune the weight in a neighborhood of the current optimal key factor combination to perform a local optimization, evaluate a new key factor combination fitting degree, and if the new key factor combination fitting degree is better, update the key factor combination as a key factor affecting condensation in the explosion-proof electrical cabinet.

[0197] An adjustment control module 4 is configured to compare the identified key factor affecting condensation in the explosion-proof electrical cabinet with known factors affecting condensation in the explosion-proof electrical cabinet, and adjust a condensation prevention measure according to a comparison result.

[0198] The data acquisition module 1 is connected to the model establishment module 2 and the trend analysis module 3, and the trend analysis module 3 is connected to the adjustment control module 4.

[0199] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A condensation forecast control method of an electrical cabinet, characterized by, The condensation prediction control method of the electrical cabinet comprises the following steps: S1, obtaining real-time environmental data in the explosion-proof electrical cabinet, and preprocessing the real-time environmental data to obtain environmental characteristic data; S2, establishing a condensation prediction model using the obtained environmental characteristic data, and performing time series analysis on the environmental characteristic data in the current explosion-proof electrical cabinet, and obtaining a condensation trend based on the time series analysis result; S3, based on the condensation trend, randomly generating an initial key factor combination and its weight, calculating the condensation trend fitting degree of each key factor combination, determining the current optimal key factor combination, presetting a conversion probability, if the conversion probability is greater than a random number, updating the weight for global search; otherwise, fine-tuning the weight in the neighborhood of the current optimal key factor combination for local optimization, evaluating the fitting degree of the new key factor combination, if it is better, updating the key factor combination as the key factor affecting the condensation in the explosion-proof electrical cabinet; S4, comparing the identified key factor affecting the condensation in the explosion-proof electrical cabinet with the known factors affecting the condensation in the explosion-proof electrical cabinet, and adjusting the condensation prevention measures according to the comparison result; The step of establishing a condensation prediction model using the obtained environmental characteristic data, and performing time series analysis on the environmental characteristic data in the current explosion-proof electrical cabinet, and obtaining a condensation trend based on the time series analysis result comprises the following steps: S21, obtaining the environmental characteristic data in the current explosion-proof electrical cabinet, and recording the historical change of the environmental characteristic data; S22, performing time series analysis on the environmental characteristic data in the current explosion-proof electrical cabinet, and extracting the periodic characteristics of the environmental characteristic data; S23, based on the periodic characteristics of the environmental characteristic data, using the condensation prediction model to predict the condensation state in the electrical cabinet at a future time, and obtaining a condensation trend; The step of using the condensation prediction model to predict the condensation state in the electrical cabinet at a future time based on the periodic characteristics of the environmental characteristic data comprises the following steps: S231, using a test method to test the time series stationarity of the environmental characteristic data; S232, performing autocorrelation graph and partial autocorrelation graph analysis on the stationary processed environmental characteristic data time series, and determining the best lag order of the environmental characteristic data; S233, using a statistical method to test the residual sequence in the environmental characteristic data; S234, combining the time series stationarity test result, the determined best lag order and the residual sequence, selecting the autoregressive order and the moving average order, and constructing a condensation prediction model; S235, using the constructed condensation prediction model to predict the condensation state in the electrical cabinet at a future time, and obtaining a condensation trend; The step of, based on the condensation trend, randomly generating an initial key factor combination and its weight, calculating the condensation trend fitting degree of each key factor combination, determining the current optimal key factor combination, presetting a conversion probability, if the conversion probability is greater than a random number, updating the weight for global search; otherwise, fine-tuning the weight in the neighborhood of the current optimal key factor combination for local optimization, evaluating the fitting degree of the new key factor combination, if it is better, updating the key factor combination as the key factor affecting the condensation in the explosion-proof electrical cabinet comprises the following steps: S31, initialize parameters of the trend analysis algorithm, set the maximum number of iterations and the preset conversion probability, randomly generate an initial set of condensation key factors, and set the initial influence weight of each factor; S32, calculate the fitting degree of each factor combination to the condensation trend, identify the current optimal key factor combination and its corresponding fitting degree value, and establish a preliminary condensation influence model; S33, if the conversion probability is greater than the random number, adjust the weight of the key factor combination using the strategy adjustment algorithm to expand the global search range of the condensation key factors; S34, if the conversion probability is less than or equal to the random number, perform local optimization within the neighborhood of the current optimal key factor combination, and adjust the weight of the key factor combination by uniformly distributed random; S35, evaluate the fitting degree of the newly generated key factor, if the new key factor combination is better than the current key factor combination, update the key factor combination; otherwise, keep the original analysis result to ensure continuous optimization of the condensation trend analysis; S36, check whether the maximum number of iterations is reached, if yes, output the optimal key factor combination as the key factor affecting the condensation in the explosion-proof electrical cabinet, otherwise, continue iteration analysis; The identified key factors affecting the condensation in the explosion-proof electrical cabinet are compared with known factors affecting the condensation in the explosion-proof electrical cabinet, and the condensation prevention measures are adjusted according to the comparison result, including the following steps: S41, set the data set of the key factors affecting the condensation in the explosion-proof electrical cabinet, and preset the number of mutation and crossover operations; S42, perform mutation operation on the data set of the key factors affecting the condensation in the explosion-proof electrical cabinet, determine the key factor data in the data set that has the greatest impact on the condensation by calculating the difference vector between each key factor data; S43, perform crossover operation on the new key factor data affecting the condensation in the explosion-proof electrical cabinet, and determine the key factor data combination that most significantly affects the condensation; S44, use clustering analysis method to perform new grouping on the mutated key factor data, determine the cluster where each key factor data is located, and find the best correlation distance between different clusters; S45, compare the key factor data obtained by mutation and crossover with the preset target function, and select the key factor data with the lowest target function as the key factor data affecting the condensation in the explosion-proof electrical cabinet in the new key factor data set; S46, judge whether the iteration termination condition is met, if yes, terminate evolution, take the current best key factor data set as the key factor affecting the condensation in the explosion-proof electrical cabinet, and adjust the condensation prevention measures.

2. The condensed water prediction control method of an electrical cabinet according to claim 1, characterized by, The formula for testing the residual sequence in the environmental characteristic data using statistical method is: ； In the formula, T is expressed as a test statistic; M is expressed as sample volume; representing the first S autocorrelation coefficient value of the first residual sequence; L denotes the number of autocorrelation coefficients.

3. The condensed water prediction control method of an electrical cabinet according to claim 1, characterized by, If the conversion probability is greater than the random number, the weight of the key factor combination is adjusted using the strategy adjustment algorithm to expand the global search range of the condensation key factors, including the following steps: S331, if the conversion probability is greater than the random number, initialize the parameters of the strategy adjustment algorithm, set the maximum number of iterations and the constraint condition, randomly generate a key factor combination and its initial weight in the defined parameter space as the starting point of the current condensation trend analysis; S332, randomly assign initial values to each key factor affecting the condensation state to form several sets of preliminary condensation trend analysis data, evaluate and determine the current best key factor combination according to the pros and cons of trend fitting degree; S333, verify whether the newly generated key factor combination meets all the preset constraints, if not, cancel the newly generated key factor combination, and try to evaluate other key factor combinations until a key factor combination that can meet all the preset constraints is obtained; S334, repeatedly execute factor searching and combination verification, continuously expand and optimize the key factor combination, and when the trend fitting degree reaches the local optimum, temporarily determine the current key factor combination as the best key factor combination; S335, introduce random factor disturbance, adjust the weight configuration of the temporarily determined current best key factor combination, and search again to obtain a new key factor combination; S336, compare the trend fitting degrees of the new and old key factor combinations, if the new key factor combination is better, update it as the global optimal key factor combination, otherwise, output the current global optimal key factor combination.

4. The condensed water prediction control method of an electrical cabinet according to claim 3, characterized by, The formula for introducing random factor disturbance and adjusting the weight configuration of the temporarily determined current best key factor combination is: ； In the formula, R new represents the new key factor combination vector after disturbance by random factors and weight configuration adjustment; R best represents the tentative current best key factor combination vector; Δ a represents the random disturbance coefficient of the a th key factor; M a represents the random disturbance vector of the a th key factor; α a represents the weight adjustment coefficient of the a th key factor; Q a represents the weight adjustment vector of the a th key factor; ω a represents the weight vector of the a th key factor in the current best key factor combination.

5. The condensed water prediction control method of an electrical cabinet according to claim 1, characterized by, The new grouping of the mutated key factor data using the clustering analysis method, determining the cluster of each key factor data, and finding the best correlation distance between different clusters includes the following steps: S441, randomly select several key factor data from the mutated key factor data set as the center points of the initial clusters; S442, assign each unassigned key factor data to the cluster center point with the highest similarity; S443, randomly select a key factor data in the key factor data set that does not belong to the cluster center as a potential new cluster center candidate; S444, calculate the total cost change caused by replacing the current cluster center point with the randomly selected key factor data, and update the cluster center set; S445, when the cluster center point no longer changes, determine the final cluster attribution of each key factor data, and calculate the best correlation distance between different clusters.

6. A condensation forecast control system of an electrical cabinet for implementing the condensation forecast control method of the electrical cabinet according to any one of claims 1 to 5, characterized by The condensation prediction control system of the electrical cabinet comprises: A data acquisition module is configured to acquire real-time environmental data in the explosion-proof electrical cabinet and pre-process the real-time environmental data to obtain environmental feature data. A model establishment module is configured to establish a condensation prediction model using the obtained environmental feature data and perform time series analysis on the environmental feature data in the current explosion-proof electrical cabinet, and obtain a condensation trend based on the time series analysis result. A trend analysis module is configured to randomly generate an initial key factor combination and its weight based on the condensation trend, calculate a condensation trend fitting degree of each key factor combination, determine a current optimal key factor combination, preset a conversion probability, and if the conversion probability is greater than a random number, update the weight for global search; otherwise, fine-tune the weight in the neighborhood of the current optimal key factor combination for local optimization, evaluate a new key factor combination fitting degree, and if the new key factor combination fitting degree is better, update the key factor combination as a key factor affecting the condensation in the explosion-proof electrical cabinet. The adjustment control module is configured to compare the identified key factor affecting the condensation in the explosion-proof electrical cabinet with known factors affecting the condensation in the explosion-proof electrical cabinet, and adjust the condensation prevention measures according to the comparison result. The data acquisition module is connected with the model establishment module and the trend analysis module, and the trend analysis module is connected with the adjustment control module.

Images