Electrical equipment room environment humidity control method and system based on difference comparison

Abstract

The invention relates to an electrical equipment room environment humidity control method and system based on difference comparison. The method comprises the following steps: acquiring original humidity signal data through a humidity sensing array, and performing noise elimination processing to obtain humidity data; calculating a humidity change rate based on the humidity data, comparing the humidity change rate with a preset threshold value, and evaluating the environmental humidity complexity; establishing an equipment humidity demand evaluation model and generating an equipment humidity demand matrix in combination with the physical position, the humidity demand and the environmental humidity complexity of the electrical equipment; based on the matrix, fusing environment state parameters and equipment operation state parameters, and calculating a target comprehensive humidity value by adopting a dynamic weight distribution method; the deviation between the target comprehensive humidity value and the actual humidity data is calculated through a feedback control algorithm, and a humidity control instruction is generated to drive humidity adjusting equipment to execute operation. According to the method, through multi-dimensional parameter fusion and dynamic weight distribution, the accuracy and adaptability of humidity control in a complex environment are improved.

Description

Technical Field

[0001] This invention belongs to the field of industrial environmental control systems, and in particular relates to a method and system for controlling the humidity of electrical equipment rooms based on difference comparison. Background Technology

[0002] With the development of industrial automation and intelligent technologies, the requirements for the accuracy of humidity control in electrical equipment rooms (such as data center computer rooms, substation control rooms, and precision instrument rooms) are increasing. Humidity is one of the key environmental parameters affecting the safe and stable operation of electrical equipment. Excessively high or low humidity can lead to decreased equipment insulation performance, condensation short circuits, static electricity accumulation, and even equipment failure. Therefore, various environmental humidity monitoring and regulation technologies have emerged. Their core objective is to monitor humidity through sensors and drive humidifiers or dehumidifiers to maintain the environmental humidity near a set target value. In traditional technologies, humidity control in electrical equipment rooms is typically based on a simple setpoint control strategy. This mainly includes: deploying one or a small number of humidity sensors at key locations within the equipment room to collect environmental humidity data; comparing the collected humidity data with a preset fixed target humidity value to calculate the deviation; and using classic feedback control algorithms such as proportional-integral-derivative (PID) to generate control commands based on this deviation to drive the humidity regulation equipment. Current traditional humidity control methods suffer from the following main problems: Traditional methods typically use a single, fixed target humidity value for control, failing to fully consider the varying humidity requirements of different equipment within the electrical equipment room due to factors such as their physical location, heat dissipation characteristics, and insulation levels. Environmental humidity changes are often complex and time-varying, making it difficult for traditional methods to dynamically assess this complexity and incorporate it into control decisions. Control targets are rigid, lacking multi-factor fusion, relying solely on the deviation between a single humidity sensor reading and a fixed setpoint for control, without effectively combining real-time environmental parameters and the equipment's own operating parameters for dynamic weighting and comprehensive decision-making. This results in control commands failing to optimally adapt to actual environmental conditions and equipment operating needs, and in some cases, even exacerbating uneven or fluctuating humidity in localized areas. Parameter adjustment is lagy and lacks adaptability; traditional PID controller parameters are usually pre-set fixed values. When the environmental humidity change pattern deviates significantly from the preset control curve, the fixed PID parameters are difficult to adjust adaptively, potentially leading to slow system response, overshoot, or continuous oscillation, affecting control accuracy and equipment operational safety. Summary of the Invention

[0003] Therefore, it is necessary to provide a method and system for controlling the ambient humidity of electrical equipment rooms based on difference comparison that can solve the above problems.

[0004] In a first aspect, this application provides a method for controlling the ambient humidity of an electrical equipment room based on difference comparison, comprising:

[0005] Raw humidity signal data is collected by a humidity sensor array, and noise is removed from the raw humidity signal data to obtain humidity data.

[0006] Based on humidity data, the humidity change rate is calculated, and the difference between the humidity change rate and a preset threshold is compared to obtain the complexity of the environmental humidity.

[0007] Based on the physical location and humidity requirements of electrical equipment, and combined with the complexity of environmental humidity, an equipment humidity requirement assessment model is established to generate an equipment humidity requirement matrix.

[0008] Based on the equipment humidity demand matrix, and combined with environmental state parameters and equipment operating state parameters, a dynamic weight allocation method is used to calculate the target comprehensive humidity value.

[0009] Based on the target comprehensive humidity value and humidity data, a feedback control algorithm is used to calculate the deviation, and a humidity control command is generated based on the deviation to drive the humidity regulating equipment to perform operations.

[0010] In one embodiment, based on humidity data, the humidity change rate is calculated, and the difference between the humidity change rate and a preset threshold is compared to obtain the environmental humidity complexity, including:

[0011] Based on humidity data, extract its fluctuation characteristic statistics;

[0012] Based on fluctuation characteristic statistics, the time series differencing order is optimized using a genetic algorithm to determine the optimal differencing order N;

[0013] The humidity change rate is generated by calculating the humidity change over N adjacent sampling times using the difference order N.

[0014] Calculate the first threshold and the second threshold based on the statistical distribution of the humidity change rate within the preset sliding window;

[0015] Determine whether the rate of change in humidity falls within the range defined by the first and second thresholds:

[0016] When the rate of change of humidity is within the range, calculate the absolute difference between the rate of change of humidity and the median of the statistical distribution.

[0017] When the humidity change rate is outside the range, calculate the absolute difference between the humidity change rate and the nearest threshold.

[0018] The calculated absolute difference is input into a preset quantization function to map and generate the complexity of environmental humidity.

[0019] In one embodiment, based on the physical location and humidity requirements of the electrical equipment, and considering the complexity of the environmental humidity, an equipment humidity requirement assessment model is established to generate an equipment humidity requirement matrix, including:

[0020] Based on the operating parameters of the electrical equipment and the preset humidity tolerance range, generate the basic humidity requirement level for the equipment.

[0021] Based on the equipment's basic humidity requirement level, an initial weight vector is generated by converting it using a preset quantization function.

[0022] Based on the initial weight vector and the complexity of the environmental humidity, an adaptive adjustment is performed to generate a dynamic demand weight vector;

[0023] Establish a location mapping relationship between the physical location of electrical equipment and the monitoring area of ​​the humidity sensor array;

[0024] Based on the location mapping relationship and dynamic demand weight vector, a device humidity demand matrix is ​​constructed, where the row dimension of the matrix corresponds to the monitoring area and the column dimension corresponds to the priority of device humidity demand.

[0025] In one embodiment, based on the equipment humidity demand matrix and combining environmental state parameters and equipment operating state parameters, a dynamic weight allocation method is used to calculate the target comprehensive humidity value, including:

[0026] Feature extraction is performed on environmental state parameters to generate dynamic environmental influencing factors. Feature extraction includes the extraction of at least one physical parameter, such as environmental temperature gradient or airflow velocity variance.

[0027] The operating status parameters of the equipment are analyzed for operating characteristics to generate equipment operating weight coefficients. The operating characteristic analysis includes the analysis of at least one operating parameter among current harmonic distortion rate, radiator temperature rise rate, and power factor change.

[0028] Based on the equipment humidity demand priority in the equipment humidity demand matrix, a multi-parameter fusion weight vector is generated by fusing environmental dynamic influencing factors and equipment operation weight coefficients through a dynamic weight allocation function.

[0029] The target comprehensive humidity value is obtained by correlation calculation based on the multi-parameter fusion weight vector and the equipment humidity requirement matrix.

[0030] In one embodiment, the correlation calculation to obtain the target comprehensive humidity value includes the following steps:

[0031] Equipment humidity requirement matrix Perform singular value decomposition to obtain its singular values ​​σ. k Left singular vector and right singular vectors Where k = 1, 2, ..., r, r = min(R, C) is the theoretical upper limit of the number of non-zero singular values ​​in the singular value decomposition;

[0032] Based on the singular value decomposition results, the nonlinear projection value is calculated using the following formula.

[0033]

[0034] Among them, u k (i) is the left singular vector u k The i-th component, v k (j) is the right singular vector v k The j-th component;

[0035] Extract the column vector d of the humidity requirement matrix D of the equipment. j =[D 1j D 2j D Rj ] T The column priority weight ω is calculated using the following formula. j :

[0036]

[0037] Based on multi-parameter fusion weight vector Equipment humidity requirement matrix D and column priority weights ω j The entropy weighting factor ξ is calculated using the following formula. ij :

[0038]

[0039] in, For the multi-parameter fusion weight vector w f The i-th component, D ij Let be the element in the i-th row and j-th column of the equipment humidity requirement matrix D;

[0040] Based on multi-parameter fusion weight vector w f Given the equipment humidity requirement matrix D, the adaptive adjustment factor ζ is calculated using the following formula:

[0041]

[0042] Among them, ||w f ||2 represents the multi-parameter fusion weight vector w f The L2 norm, ||D|| F Let Frobenius norm be the humidity requirement matrix D of the equipment. For the multi-parameter fusion weight vector w f The maximum value of each element;

[0043] Based on nonlinear projection values Entropy weighting factor ξ ijAnd the adaptive adjustment factor ζ, the target comprehensive humidity value is calculated using the following formula:

[0044]

[0045] Among them, H target The target comprehensive humidity value.

[0046] In one embodiment, after the humidity control device performs its operation, the method further includes:

[0047] Obtain the humidity data sequence H for the T consecutive sampling periods prior to the current time. t =[h t-T+1 , ..., h t ], where h t Here is the humidity data at time t;

[0048] Calculation of instantaneous humidity change rate based on humidity data sequence And based on the humidity change rate Δh t Generate humidity change curve C actual ;

[0049] The humidity change curve C was calculated using a dynamic time warping algorithm. actual Adjustment curve C with preset target target The similarity δDTW;

[0050] When the similarity δDTW is lower than the preset threshold δ0, the proportional-integral-derivative parameter weight matrix of the feedback control algorithm is updated using the following formula:

[0051] W PID ←W PID -η·(1-δDTW)·ΔW base

[0052] Where η is the learning rate factor, ΔW base This is a preset constant.

[0053] In one embodiment, noise removal processing is performed on the original humidity signal data to obtain humidity data, including:

[0054] To address the time-varying, non-steady-state noise characteristics of humidity signals in electrical equipment rooms, a state prediction model adapted to the noise characteristics is constructed.

[0055] Calculate the dynamic deviation between the original humidity signal and the predicted value output by the state prediction model;

[0056] The Kalman filter gain parameter is adjusted in real time based on dynamic bias, and the predicted value is fused with the adjusted Kalman filter gain parameter to generate a humidity estimate as humidity data.

[0057] Secondly, this application also provides an environmental humidity control system for electrical equipment rooms based on difference comparison, comprising:

[0058] The data acquisition and processing module is used to acquire raw humidity signal data through a humidity sensor array and to perform noise reduction processing on the raw humidity signal data to obtain humidity data.

[0059] The environmental complexity analysis module is used to calculate the humidity change rate based on humidity data, and compare the difference between the humidity change rate and a preset threshold to obtain the environmental humidity complexity.

[0060] The equipment requirements modeling module is used to establish an equipment humidity requirements assessment model based on the physical location and humidity requirements of electrical equipment, combined with the complexity of environmental humidity, and generate an equipment humidity requirements matrix.

[0061] The target humidity decision module is used to calculate the target comprehensive humidity value based on the equipment humidity demand matrix, combined with environmental status parameters and equipment operating status parameters, using a dynamic weight allocation method.

[0062] The feedback control execution module is used to calculate the deviation based on the target comprehensive humidity value and humidity data, and generate humidity control commands based on the deviation to drive the humidity regulating equipment to perform operations.

[0063] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described steps of the electrical equipment room environmental humidity control method based on difference comparison.

[0064] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method for controlling the ambient humidity of an electrical equipment room based on difference comparison.

[0065] The aforementioned method and system for controlling the ambient humidity of electrical equipment rooms based on difference comparison, along with the computer equipment and storage medium, acquires multi-dimensional raw data through a humidity sensor array and eliminates noise interference to obtain ambient humidity distribution information. Based on this information, the humidity change rate is calculated and compared with a preset threshold to quantify the complexity of the ambient humidity and assess humidity fluctuation characteristics. A humidity demand matrix is ​​established by combining the physical location of the electrical equipment with its humidity requirements, enabling differentiated humidity demand modeling. Environmental state parameters and equipment operating state parameters are integrated, and a dynamic weight allocation method is used to calculate the target comprehensive humidity value, allowing the control target to dynamically adapt to environmental complexity and equipment operating status. A feedback control algorithm transforms the deviation between the target humidity value and actual humidity data into control commands to drive the equipment, achieving multi-dimensional parameter fusion and dynamic weight allocation. This solves the problems of rigid control with a single fixed target value, failure to consider differences in equipment requirements, and lag in parameter adjustment, improving the accuracy, adaptability, and operational safety of humidity control in complex industrial environments. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0067] Figure 1 This is a flowchart illustrating the steps of the electrical equipment room environmental humidity control method based on difference comparison according to the present invention.

[0068] Figure 2 This is a structural diagram of the environmental humidity control system for electrical equipment rooms based on difference comparison according to the present invention. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0070] In one embodiment, such as Figure 1As shown, a method for controlling the environmental humidity of an electrical equipment room based on difference comparison is provided. This embodiment illustrates the application of this method to a terminal (such as an embedded industrial control computer). It is understood that this method can also be applied to a cloud server, or to a system where an edge computing terminal collaborates with a cloud server. In the implementation environment of the electrical equipment room, the hardware architecture includes a humidity sensor array deployed in each monitoring area, an environmental status sensor for collecting ambient temperature gradients and airflow speeds, an equipment operating status sensor for monitoring equipment current harmonic distortion rate and radiator temperature rise rate, and controlled humidity control equipment (such as a variable frequency dehumidifier or humidifier). In scenarios requiring high-precision humidity control (such as data center computer rooms or precision instrument rooms), the terminal receives the raw humidity signal from the sensor array in real time, generates reliable data through noise cancellation processing, and integrates the parameters of the environmental status sensor and the equipment operating status sensor, combined with the terminal's built-in equipment physical location mapping database, to dynamically construct an equipment humidity demand matrix. Based on this matrix, the terminal uses a dynamic weight allocation algorithm to calculate the target comprehensive humidity value and generates control commands to drive the humidity control equipment operation. In a terminal-server collaborative scenario, the terminal is responsible for real-time data acquisition and emergency control, while the server performs model training and sends updated parameters to the terminal, forming a closed-loop interaction through the network. In this embodiment, the method includes the following steps:

[0071] S01: The original humidity signal data is collected through the humidity sensor array, and the original humidity signal data is processed to remove noise to obtain humidity data.

[0072] This system utilizes a distributed humidity sensor array to collect raw humidity signal data from multiple monitoring areas in real time. This array, composed of multiple spatially discrete humidity sensors, comprehensively senses the environmental humidity distribution. Due to complex electromagnetic interference and equipment operating noise in electrical equipment rooms, the raw humidity signal exhibits time-varying and non-steady-state characteristics, requiring noise cancellation processing to improve data reliability. A state prediction model adapted to the noise characteristics of the electrical equipment room can be constructed to calculate the dynamic deviation between the raw signal and the model's predicted values ​​in real time. Based on this deviation, the Kalman filter gain parameters are adaptively optimized, and the predicted values ​​are fused with the dynamically adjusted filter parameters to reconstruct the signal, generating high-confidence humidity data. Through spatial redundancy sampling and adaptive noise suppression techniques, accurate humidity reference information is provided for subsequent control decisions, overcoming measurement distortion caused by signal interference in industrial environments.

[0073] S02, based on humidity data, calculate the humidity change rate and compare the difference between the humidity change rate and a preset threshold to obtain the environmental humidity complexity.

[0074] The steps for calculating the humidity change rate using humidity data are as follows: extract its fluctuation characteristic statistics, optimize the difference order of the time series using a genetic algorithm, determine the optimal difference order, calculate the humidity change at N adjacent sampling times based on this order N, and generate the humidity change rate (the absolute change in humidity per unit time, reflecting the dynamic fluctuation intensity of environmental humidity); the statistical distribution of the humidity change rate can be analyzed within a preset sliding window, the first threshold and the second threshold are calculated as preset boundary intervals, and it is determined whether the humidity change rate is within the threshold interval: when it is within the interval, the absolute difference between it and the median of the statistical distribution is calculated; when it is outside the interval, the absolute difference between it and the nearest threshold is calculated. The obtained absolute difference is input into a preset quantization function for nonlinear mapping to generate the environmental humidity complexity (a scalar value in the range of 0-1, representing the time-varying complexity and unpredictability of environmental humidity changes; high values ​​correspond to high volatility and control difficulty, and low values ​​represent a stable state).

[0075] S03. Based on the physical location and humidity requirements of electrical equipment, and combined with the complexity of environmental humidity, establish an equipment humidity requirement assessment model and generate an equipment humidity requirement matrix.

[0076] The equipment humidity demand assessment model is a dynamic calculation framework used to quantify the different humidity requirements of various devices. Based on the operating parameters of electrical equipment (such as rated power or insulation class) and a preset humidity tolerance range (such as ±5% RH), the model generates a basic humidity demand level for the equipment, representing the basic requirements for humidity stability. The basic demand level is converted into an initial weight vector through a preset quantization function (such as linear or nonlinear mapping). This vector initially assigns priority weights to each device in the control process. Adaptive adjustments are made to the initial weight vector based on the complexity of the environmental humidity, and the weights are dynamically corrected through a weighting function to generate a dynamic demand weight vector that reflects the real-time impact of environmental complexity on equipment requirements. A location mapping relationship is established between the physical location of the electrical equipment and the monitoring area of ​​the humidity sensor array, associating the equipment location with discrete monitoring points of the sensor array. Based on the location mapping relationship and the dynamic demand weight vector, an equipment humidity demand matrix is ​​constructed. Matrix operations are used to integrate spatial location and demand weights, providing structured input for subsequent control.

[0077] S04. Based on the equipment humidity demand matrix, combined with environmental state parameters and equipment operating state parameters, the target comprehensive humidity value is calculated using a dynamic weight allocation method.

[0078] In implementation, features are extracted from environmental state parameters (including real-time monitoring data of at least one physical parameter such as ambient temperature gradient or airflow velocity variance) to generate dynamic environmental influencing factors, quantifying the dynamic interference intensity of environmental factors on humidity control. Operational features are analyzed from equipment operating state parameters (including real-time collected data of at least one operating parameter such as current harmonic distortion rate, radiator temperature rise rate, or power factor change) to generate equipment operating weight coefficients, reflecting the priority impact of equipment operating status on humidity demand. The equipment humidity demand matrix is ​​a two-dimensional data structure, with rows corresponding to the monitoring area (representing spatial distribution) and columns corresponding to the equipment humidity demand priority (representing control importance). Based on the equipment humidity demand priority of this matrix, a dynamic weight allocation function (an adaptive algorithm) is used to fuse the dynamic environmental influencing factors and equipment operating weight coefficients to generate a multi-parameter fusion weight vector, integrating multi-dimensional parameters to dynamically allocate weights. Based on this weight vector and the equipment humidity demand matrix, a pre-set formula is used to perform correlation calculations to calculate the target comprehensive humidity value (the optimal humidity setpoint after multi-dimensional fusion), ensuring that the control target adapts to environmental changes and equipment operating status.

[0079] S05, based on the target comprehensive humidity value and humidity data, uses a feedback control algorithm to calculate the deviation, and generates a humidity control command based on the deviation to drive the humidity regulating equipment to perform the operation.

[0080] In the process of executing control based on the target comprehensive humidity value and humidity data, a feedback control algorithm is used to calculate the deviation (the difference between the target comprehensive humidity value and the actual humidity data). Based on this deviation, a humidity control command is generated to drive the humidity adjustment equipment (such as a variable frequency dehumidifier or humidifier) ​​to perform operations to adjust the ambient humidity. In implementation, the feedback control algorithm adopts a proportional-integral-derivative (PID) control mechanism, which dynamically calculates the control quantity to generate the command through a preset weight matrix. At the same time, after driving the humidity adjustment equipment to perform operations, the humidity data sequence of the previous T consecutive sampling periods can be obtained. Based on this sequence, the instantaneous humidity change rate (the ratio of the absolute difference in humidity between adjacent sampling points to the time interval) is calculated to generate the actual humidity change curve. The similarity between this curve and the preset target adjustment curve is compared by a dynamic time warping algorithm (defined as a quantitative index of the degree of curve shape matching). When the similarity is lower than a preset threshold, the PID parameter weight matrix is ​​updated using a learning rate factor and a preset constant to adaptively optimize the control performance.

[0081] The aforementioned method for controlling the humidity of electrical equipment rooms based on difference comparison collects multi-dimensional raw humidity signals through a humidity sensor array and performs noise cancellation processing to obtain reliable data. Based on this data, the humidity change rate is calculated and compared with a preset threshold to quantify the complexity of the environmental humidity. A two-dimensional demand matrix is ​​generated by establishing an equipment humidity demand assessment model by combining the physical location of electrical equipment and humidity requirements. The target comprehensive humidity value is calculated by integrating environmental state parameters (such as temperature gradient) and equipment operating state parameters (such as current harmonic distortion rate) using a dynamic weight allocation method. The humidity regulating equipment is driven to perform operations through a feedback control algorithm. By using multi-dimensional sensing and dynamic demand modeling, the rigidity of single fixed target value control is overcome. The difference comparison quantifies the environmental complexity and integrates multi-source parameters to achieve dynamic weight allocation, solving the shortcomings of traditional methods that do not consider the differences in equipment demand and the dynamic characteristics of the environment. At the same time, the real-time feedback and parameter fusion mechanism significantly improves the adaptive capability of the control system, achieving synergistic optimization of humidity control accuracy and equipment operation safety in complex industrial environments.

[0082] In one embodiment, based on humidity data, the humidity change rate is calculated, and the difference between the humidity change rate and a preset threshold is compared to obtain the environmental humidity complexity, including:

[0083] S11, based on humidity data, extract its fluctuation characteristic statistics;

[0084] S12, Based on fluctuation characteristic statistics, the time series differencing order is optimized using a genetic algorithm to determine the optimal differencing order N;

[0085] S13, use the difference order N to calculate the humidity change over N adjacent sampling times and generate the humidity change rate;

[0086] S14, Calculate the first threshold and the second threshold based on the statistical distribution of the humidity change rate within the preset sliding window;

[0087] S15, determine whether the humidity change rate is within the interval formed by the first threshold and the second threshold:

[0088] S15.1 When the rate of change of humidity is within the range, calculate the absolute difference between the rate of change of humidity and the median of the statistical distribution;

[0089] S15.2, When the humidity change rate is outside the range, calculate the absolute difference between the humidity change rate and the nearest threshold;

[0090] S16, input the calculated absolute difference into the preset quantization function to map and generate the environmental humidity complexity.

[0091] Specifically, the fluctuation characteristic statistics of humidity data are extracted (statistical indicators such as variance and standard deviation calculated from the humidity signal sequence to quantify the dynamic fluctuation characteristics of environmental humidity). Based on these fluctuation characteristic statistics, a genetic algorithm is used to optimize the time series differencing order. By simulating the natural selection process (including selection, crossover, and mutation operations), the optimal differencing order N is iteratively searched to capture the non-stationary characteristics of the humidity sequence to the greatest extent. This optimal differencing order N is then used to calculate the humidity change (the absolute difference in humidity between adjacent sampling points) over N adjacent sampling times, generating the humidity change rate, which reflects the humidity change per unit time. Dynamic change intensity; within a preset sliding window (N = 5-15 sampling periods, where a single sampling period is 1-5 minutes, which can be adaptively adjusted according to the sampling frequency of the humidity sensor array based on control accuracy requirements), analyze the statistical distribution (normal or skewed distribution) of the humidity change rate, and calculate the first threshold (as shown in the lower quartile) and the second threshold (as shown in the upper quartile) based on the distribution characteristics as boundary intervals to distinguish humidity change patterns; determine whether the current humidity change rate is within the interval formed by the first threshold and the second threshold: if it is within the interval, calculate the absolute difference between the humidity change rate and the median of the statistical distribution, and evaluate... Estimate the degree of deviation from the central trend; if it is outside the interval, calculate the absolute difference between the humidity change rate and the nearest threshold (the boundary value close to the change rate) to quantify its abnormal deviation amplitude; input the calculated absolute difference into a preset quantization function (Sigmoid or linear mapping function, in this embodiment a piecewise linear quantization function is used, let the absolute difference be Δ, and the environmental humidity complexity be C (value range [0,1], 0 represents the simplest, 1 represents the most complex), when Δ∈[0,a], C=k1·Δ where a is the 1 / 4 quantile of the statistical distribution of the humidity change rate, k1=1 / a; when Δ∈(a,b) When b is the 3 / 4 quantile of the statistical distribution of the humidity change rate, k2 = 0.5 / (ba); when Δ∈(b, +∞), C = 1; where a and b are calculated based on the statistical distribution of the humidity change rate within a preset sliding window (related to the first threshold and the second threshold, a≤ the first threshold, b≥ the second threshold). By nonlinearly mapping the absolute difference to a scalar value of environmental humidity complexity in the range of 0-1, where a high value (close to 1) represents high volatility and control complexity, and a low value (close to 0) represents a stable state, a quantitative assessment of the dynamic characteristics of environmental humidity is achieved.

[0092] In one embodiment, based on the physical location and humidity requirements of the electrical equipment, and considering the complexity of the environmental humidity, an equipment humidity requirement assessment model is established to generate an equipment humidity requirement matrix, including:

[0093] S21, Based on the operating parameters of the electrical equipment and the preset humidity tolerance range, generate the basic humidity requirement level of the equipment;

[0094] S22, based on the equipment's basic humidity requirement level, generates an initial weight vector through a preset quantization function;

[0095] S23, based on the initial weight vector and the complexity of the environmental humidity, perform adaptive adjustment to generate a dynamic demand weight vector;

[0096] S24, Establish the location mapping relationship between the physical location of electrical equipment and the monitoring area of ​​the humidity sensor array;

[0097] S25. Based on the location mapping relationship and dynamic demand weight vector, construct the equipment humidity demand matrix, where the row dimension of the matrix corresponds to the monitoring area and the column dimension corresponds to the equipment humidity demand priority.

[0098] For example, based on the operating parameters of the electrical equipment (such as rated power and insulation class) and the preset humidity tolerance range (such as ±5%RH), a basic humidity requirement level for the equipment is generated (representing the basic requirement level of the equipment for humidity stability, defined by quantifying the difference between the equipment's own characteristics and humidity tolerance). Based on the basic humidity requirement level, the basic humidity requirement level L (value range [1,5], where 1 is the minimum requirement and 5 is the maximum requirement) is converted into an initial weight vector w using a preset quantization function (linear mapping or Sigmoid function; in this embodiment, the Sigmoid function is used). i (Value range [0,1]), in the following form: Where λ = 1.2 is the adjustment parameter, and an initial weight vector is generated by conversion (calibrated using historical equipment operation data). The relative weights of each device in the control priority are initially assigned and are positively correlated with the demand level. Based on the initial weight vector and the environmental humidity complexity (0-1 scalar value), an adaptive adjustment is performed to generate a dynamic demand weight vector. The initial weights are corrected in real time through a weighting function (multiplication factor or exponential decay), so that the weight vector dynamically responds to changes in environmental complexity (increasing the weight of high-demand devices in high complexity, and maintaining or reducing the weight in low complexity), thus optimizing the adaptability of control decisions. A monitoring area for the physical location of electrical equipment and humidity sensor array is established. The location mapping relationship of the domain can be achieved through spatial coordinate transformation or regional correlation matrix, mapping the device location to discrete monitoring points of the sensor array; based on the location mapping relationship and dynamic demand weight vector, a device humidity demand matrix is ​​constructed, where the matrix row dimension corresponds to the monitoring area (representing spatial distribution) and the column dimension corresponds to the device humidity demand priority (representing control importance). By integrating spatial location and dynamic weight through matrix multiplication or tensor operation, a two-dimensional demand structure is generated, providing input data for subsequent calculation of the target comprehensive humidity value. Through differentiated demand modeling and dynamic adjustment, the shortcomings of traditional control in ignoring the differences in device location and the dynamic influence of the environment are solved.

[0099] In one embodiment, based on the equipment humidity demand matrix and combining environmental state parameters and equipment operating state parameters, a dynamic weight allocation method is used to calculate the target comprehensive humidity value, including:

[0100] S31, extract features from environmental state parameters to generate environmental dynamic influencing factors. Feature extraction includes the extraction of at least one physical parameter from environmental temperature gradient or airflow velocity variance.

[0101] S32, perform operational feature analysis on the equipment operating status parameters and generate equipment operating weight coefficients. The operational feature analysis includes the analysis of at least one of the following operating parameters: current harmonic distortion rate, radiator temperature rise rate, and power factor change.

[0102] S33, based on the equipment humidity demand priority in the equipment humidity demand matrix, uses a dynamic weight allocation function to fuse environmental dynamic influence factors and equipment operation weight coefficients to generate a multi-parameter fusion weight vector;

[0103] S34, based on the multi-parameter fusion weight vector and the equipment humidity requirement matrix, obtains the target comprehensive humidity value through correlation calculation.

[0104] Specifically, features are extracted from environmental state parameters (including at least one physical parameter such as ambient temperature gradient or airflow velocity variance, which directly reflects the intensity of environmental interference with humidity distribution; parameter selection depends on sensor availability and environmental complexity) to generate dynamic environmental influencing factors. The ambient temperature gradient is the rate of temperature change per unit distance (obtained by differential calculation of adjacent sensor data). The airflow velocity variance is the degree of dispersion of the airflow velocity sequence (calculated based on the variance formula), used to quantify the dynamic interference intensity of environmental factors on humidity control; for equipment operating status parameters (including at least one operating parameter such as current harmonic distortion rate, radiator temperature rise rate, or power factor change, for high-precision equipment (such as relays): current harmonic distortion rate is preferred because current quality directly affects insulation humidity; for heat dissipation-critical equipment (such as power cabinets): radiator temperature rise rate is preferred; for energy-sensitive equipment: power factor change is preferred. This is achieved through a preset rule base, which is generated by training from historical equipment data), performing operational feature analysis to generate equipment operating weight coefficients. Current harmonic distortion rate is the degree of current waveform distortion (calculated through FFT analysis), radiator temperature rise rate... The rate of change of temperature over time (calculated by slope) and the change in power factor (the amplitude of power factor fluctuations, calculated by difference statistics) reflect the priority impact of equipment operating status on humidity demand. Based on the equipment humidity demand priority (control importance corresponding to column dimensions) in the equipment humidity demand matrix, a multi-parameter fusion weight vector is generated by fusing environmental dynamic influence factors and equipment operating weight coefficients through a dynamic weight allocation function (such as weighted average or neural network fusion). This vector dynamically allocates the weights of each parameter to adapt to real-time complexity. Based on this multi-parameter fusion weight vector and the equipment humidity demand matrix, the target comprehensive humidity value is obtained through correlation operations, including four steps of correlation operations: singular value decomposition, nonlinear mapping, entropy weight fusion, and adaptive adjustment, to achieve precise optimization of the dynamic humidity setpoint.

[0105] In one embodiment, the correlation calculation to obtain the target comprehensive humidity value includes the following steps:

[0106] S41, Equipment Humidity Demand Matrix Perform singular value decomposition to obtain its singular values ​​σ. k Left singular vector and right singular vectors Where k = 1, 2, ..., r, r = min(R, C) is the theoretical upper limit of the number of non-zero singular values ​​in the singular value decomposition;

[0107] S42. Based on the singular value decomposition results, calculate the nonlinear projection value using the following formula.

[0108]

[0109] Among them, u k (i) is the left singular vector u k The i-th component, v k (j) is the right singular vector v k The j-th component;

[0110] S43, Extract the column vector d of the equipment humidity requirement matrix D. j =[D 1j D 2j D Rj ] T The column priority weight ω is calculated using the following formula. j :

[0111]

[0112] S44, based on multi-parameter fusion weight vector Equipment humidity requirement matrix D and column priority weights ω j The entropy weighting factor ξ is calculated using the following formula. ij :

[0113]

[0114] in, For the multi-parameter fusion weight vector w f The i-th component, D ij Let be the element in the i-th row and j-th column of the equipment humidity requirement matrix D;

[0115] S45, based on multi-parameter fusion weight vector w f Given the equipment humidity requirement matrix D, the adaptive adjustment factor ζ is calculated using the following formula:

[0116]

[0117] Among them, ||w f ||2 represents the multi-parameter fusion weight vector w f The L2 norm, ||D|| F Let Frobenius norm be the humidity requirement matrix D of the equipment. For the multi-parameter fusion weight vector w f The maximum value of each element;

[0118] S46, based on nonlinear projection values Entropy weighting factor ξ ij And the adaptive adjustment factor ζ, the target comprehensive humidity value is calculated using the following formula:

[0119]

[0120] Among them, H target The target comprehensive humidity value.

[0121] For example, in the process of calculating the target comprehensive humidity value in the correlation operation, the equipment humidity requirement matrix is ​​first processed. Perform singular value decomposition to obtain singular values ​​σ. kLeft singular vector and right singular vectors The index k ranges from k = 1, 2, ..., r, and r = min(R, C) represents the theoretical upper limit of the number of non-zero singular values. The spatial distribution characteristics of humidity demand are revealed by deconstructing the intrinsic structure of the matrix. Based on the singular value decomposition results, the formula is applied... Calculate nonlinear projection values Where u k (i) is the left singular vector u k The i-th component, v k (j) is the right singular vector v k The j-th component is mapped from the linear combination to the nonlinear domain using the hyperbolic tangent function, enhancing adaptability to complex humidity interactions; the column vector d of the equipment humidity demand matrix is ​​extracted. j =[D 1j D 2j D Rj ] T And use the formula Calculate column priority weight ω j Quantify the priority of humidity requirements for each column dimension; combine multiple parameters to form a weight vector. Equipment humidity requirement matrix D and column priority weights ω j Using the formula Calculate the entropy weighting factor ξ ij ,in, For the multi-parameter fusion weight vector w f The i-th component, D ij For the element in the i-th row and j-th column of the equipment humidity demand matrix D, a weight is dynamically assigned based on the principle of information entropy to reflect the fusion uncertainty of multi-source parameters. A higher weight indicates a more significant contribution of that parameter to the target humidity decision. Based on the multi-parameter fusion weight vector w... f Application formula of equipment humidity requirement matrix D Calculate the adaptive adjustment factor ζ, where ||w f ||2 represents the multi-parameter fusion weight vector w f The L2 norm, ||D|| F Let Frobenius norm be the humidity requirement matrix D of the equipment. For the multi-parameter fusion weight vector w f The maximum value of each element is determined by automatically adjusting the calculation parameters to adapt to dynamic environmental changes; nonlinear projection values ​​are integrated. Entropy weighting factor ξ ij and the adaptive adjustment factor ζ, through the formula Calculate the target comprehensive humidity value H targetBy integrating multi-dimensional data in a weighted average form, the humidity setpoint is accurately adapted to complex environments and equipment requirements, thereby improving the robustness and accuracy of the control system.

[0122] In one embodiment, after the humidity control device performs its operation, the method further includes:

[0123] S51, Obtain the humidity data sequence H for the T consecutive sampling periods prior to the current time. t =[h t-T+1 , ..., h t ], where h t Here is the humidity data at time t;

[0124] S52, Calculate the instantaneous humidity change rate based on humidity data sequence And based on the humidity change rate Δh t Generate humidity change curve C actual ;

[0125] S53, calculates the humidity change curve C using a dynamic time warping algorithm. actual Adjustment curve C with preset target target The similarity δDTW;

[0126] S54, when the similarity δDTW is lower than the preset threshold δ0, the proportional-integral-derivative parameter weight matrix of the feedback control algorithm is updated using the following formula:

[0127] W PID ←W PID -η·(1-δDTW)·ΔW base

[0128] Where η is the learning rate factor, ΔW base This is a preset constant.

[0129] Specifically, during the adaptive optimization process after the humidity control device performs its operation, the humidity data sequence H for the T consecutive sampling periods prior to the current moment is acquired. t =[h t-T+1 , ..., h t ], where h t The humidity data at time t constitute the original dataset reflecting recent dynamic changes in humidity; the instantaneous humidity change rate is calculated based on this sequence. Where Δt is the sampling time interval, and the actual humidity change curve C is generated by connecting the rate of change values ​​at each time point. actual The curve, with time on the horizontal axis and instantaneous rate of change on the vertical axis, quantifies the real-time fluctuation characteristics of ambient humidity; C is calculated using the Dynamic Time Warping (DTW) algorithm. actual Adjustment curve C with preset target targetThe similarity δDTW is calculated by nonlinearly aligning the time axes of the two curves, minimizing the cumulative distance to eliminate time offset interference, and accurately evaluating the degree of matching between the actual control effect and the expected trajectory. When the similarity δDTW is lower than the preset threshold δ0 (determined based on the humidity tolerance range of the electrical equipment: for humidity-sensitive equipment, such as precision relays, tolerance ±3%RH), δ0≥0.85; for ordinary equipment (such as low-voltage switchgear, tolerance ±5%RH), δ0≥0.75. By statistically analyzing the historical adjustment curve similarity of the equipment under qualified humidity conditions, the lower limit of the 95% confidence interval is taken as δ0 to ensure that more than 95% of normal adjustment processes meet the similarity requirements (default value is 0.8). This indicates that the actual adjustment process deviates from the target shape, and the formula W is used. PID ←W PID -η·(1-δDTW)·ΔW base Update the proportional-integral-derivative (PID) parameter weight matrix, where η is the learning rate factor (the adjustment range of the control parameters), and ΔW base A preset constant (basic adjustment step size, value range [0.05, 0.2], default value is 0.1, which can be dynamically adapted according to the sensitivity of the equipment and the convergence requirements of the system) is used to dynamically reduce the risk of control overshoot and oscillation through the principle of negative feedback, so that the PID controller can adapt to the dynamic characteristics of the environment in real time, and finally achieve adaptive stability optimization of the humidity control system.

[0130] In one embodiment, noise removal processing is performed on the original humidity signal data to obtain humidity data, including:

[0131] S61, for the time-varying non-steady-state noise characteristics of the humidity signal in the electrical equipment room, construct a state prediction model adapted to the noise characteristics;

[0132] S62, calculate the dynamic deviation between the original humidity signal and the predicted value output by the state prediction model;

[0133] S63 adjusts the Kalman filter gain parameters in real time based on dynamic bias, and merges the predicted value with the adjusted Kalman filter gain parameters to generate a humidity estimate as humidity data.

[0134] For example, during the noise cancellation process, a state prediction model adapted to the time-varying, non-stationary noise characteristics of the humidity signal in the electrical equipment room (the non-stationary, time-varying interference pattern of the humidity signal caused by electromagnetic interference and operating noise of electrical equipment) can be constructed. This model integrates autoregressive (AR) components and machine learning modules within a state-space framework, with the state equation modeling the humidity evolution law, the observation equation describing the noise interference, and an embedded LSTM network capturing long-term dependent patterns of noise (such as periodic equipment interference). The model is pre-trained offline using maximum likelihood estimation to initialize parameters based on historical noisy datasets, and the model weights are adjusted in real-time according to the dynamic deviation (the difference between the original signal and the predicted value). The state space is updated using a recursive least squares algorithm. The parameters, model output, and Kalman filter gain are linked. When sudden noise is detected, the weights of the LSTM noise classifier are automatically strengthened to ensure that the humidity estimate adapts to environmental changes in real time. The dynamic deviation between the original humidity signal and the predicted value output by the state prediction model is calculated (quantifying the real-time noise intensity and obtaining it in real time through differential operations). Based on the dynamic deviation, the Kalman filter gain parameters are adjusted in real time (by optimizing the Kalman gain matrix to minimize the estimation error covariance), and the predicted value and the adjusted Kalman filter gain parameters are fused. The Kalman filter recursive algorithm is applied to generate the humidity estimate as the humidity data (signal reconstruction is achieved through weighted fusion, eliminating non-steady-state noise and improving data reliability), providing a high-precision humidity benchmark for subsequent control.

[0135] The aforementioned method for controlling the humidity of electrical equipment rooms based on difference comparison acquires multi-dimensional raw signals through a humidity sensor array and uses Kalman filtering and dynamic gain adjustment to eliminate noise, obtaining high-confidence humidity data. It combines a genetic algorithm to optimize the difference order to calculate the humidity change rate and quantifies the complexity of environmental humidity through statistical distribution threshold difference comparison, accurately assessing environmental fluctuation characteristics. It generates basic demand levels based on the physical location of equipment and humidity tolerance range, integrates a dynamically corrected weight vector based on environmental complexity, and constructs a spatially and priority-related equipment humidity demand matrix. It extracts dynamic influencing factors such as environmental temperature gradient and airflow velocity variance, analyzes operating parameters such as equipment current harmonic distortion rate and radiator temperature rise rate, and uses singular value decomposition and nonlinear projection to fuse multi-source parameters to generate a target comprehensive humidity value, enabling the setpoint to dynamically adapt to environmental complexity and equipment operating status. Finally, it uses a dynamic time warping (DTW) algorithm to compare the similarity between the measured humidity curve and the target adjustment curve in real time, adaptively updating the PID weight matrix to drive the humidity control equipment to respond precisely. By employing four core technologies—difference comparison to quantify environmental complexity, matrix modeling of differentiated equipment requirements, dynamic weighting to fuse multi-source parameters, and closed-loop feedback adaptive optimization—this approach overcomes three major shortcomings of traditional methods: rigidity of single target values, neglect of differences in equipment requirements, and lag in parameter adjustment. It significantly improves the accuracy of humidity control, equipment safety, and system adaptability in complex industrial environments.

[0136] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0137] Based on the same inventive concept, this application also provides a humidity control system for electrical equipment room based on difference comparison, used to implement the above-mentioned humidity control method for electrical equipment room based on difference comparison. The solution provided by this system is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the humidity control system for electrical equipment room based on difference comparison provided below can be found in the limitations of the humidity control method for electrical equipment room based on difference comparison described above, and will not be repeated here.

[0138] In one exemplary embodiment, such as Figure 2 As shown, an environmental humidity control system for an electrical equipment room based on difference comparison is provided, including:

[0139] The data acquisition and processing module 101 is used to acquire raw humidity signal data through a humidity sensor array and perform noise cancellation processing on the raw humidity signal data to obtain humidity data.

[0140] The environmental complexity analysis module 102 is used to calculate the humidity change rate based on humidity data and compare the difference between the humidity change rate and a preset threshold to obtain the environmental humidity complexity.

[0141] The equipment requirement modeling module 103 is used to establish an equipment humidity requirement assessment model based on the physical location and humidity requirements of electrical equipment, combined with the complexity of environmental humidity, and generate an equipment humidity requirement matrix.

[0142] The target humidity decision module 104 is used to calculate the target comprehensive humidity value based on the equipment humidity demand matrix, combined with environmental status parameters and equipment operating status parameters, using a dynamic weight allocation method.

[0143] The feedback control execution module 105 is used to calculate the deviation based on the target comprehensive humidity value and humidity data using a feedback control algorithm, and generate a humidity control command based on the deviation to drive the humidity regulating device to perform operations.

[0144] In one embodiment, the environment complexity analysis module 102 is further configured to:

[0145] Based on humidity data, extract its fluctuation characteristic statistics;

[0146] Based on fluctuation characteristic statistics, the time series differencing order is optimized using a genetic algorithm to determine the optimal differencing order N;

[0147] The humidity change rate is generated by calculating the humidity change over N adjacent sampling times using the difference order N.

[0148] Calculate the first threshold and the second threshold based on the statistical distribution of the humidity change rate within the preset sliding window;

[0149] Determine whether the rate of change in humidity falls within the range defined by the first and second thresholds:

[0150] When the rate of change of humidity is within the range, calculate the absolute difference between the rate of change of humidity and the median of the statistical distribution.

[0151] When the humidity change rate is outside the range, calculate the absolute difference between the humidity change rate and the nearest threshold.

[0152] The calculated absolute difference is input into a preset quantization function to map and generate the complexity of environmental humidity.

[0153] In one embodiment, the equipment requirement modeling module 103 is further configured to:

[0154] Based on the operating parameters of the electrical equipment and the preset humidity tolerance range, generate the basic humidity requirement level for the equipment.

[0155] Based on the equipment's basic humidity requirement level, an initial weight vector is generated by converting it using a preset quantization function.

[0156] Based on the initial weight vector and the complexity of the environmental humidity, an adaptive adjustment is performed to generate a dynamic demand weight vector;

[0157] Establish a location mapping relationship between the physical location of electrical equipment and the monitoring area of ​​the humidity sensor array;

[0158] Based on the location mapping relationship and dynamic demand weight vector, a device humidity demand matrix is ​​constructed, where the row dimension of the matrix corresponds to the monitoring area and the column dimension corresponds to the priority of device humidity demand.

[0159] In one embodiment, the target humidity decision module 104 is further configured to:

[0160] Feature extraction is performed on environmental state parameters to generate dynamic environmental influencing factors. Feature extraction includes the extraction of at least one physical parameter, such as environmental temperature gradient or airflow velocity variance.

[0161] The operating status parameters of the equipment are analyzed for operating characteristics to generate equipment operating weight coefficients. The operating characteristic analysis includes the analysis of at least one operating parameter among current harmonic distortion rate, radiator temperature rise rate, and power factor change.

[0162] Based on the equipment humidity demand priority in the equipment humidity demand matrix, a multi-parameter fusion weight vector is generated by fusing environmental dynamic influencing factors and equipment operation weight coefficients through a dynamic weight allocation function.

[0163] The target comprehensive humidity value is obtained by correlation calculation based on the multi-parameter fusion weight vector and the equipment humidity requirement matrix.

[0164] In one embodiment, the target humidity decision module 104 is further configured to:

[0165] Equipment humidity requirement matrix Perform singular value decomposition to obtain its singular values ​​σ. k Left singular vector and right singular vectors Where k = 1, 2, ..., r, r = min(R, C) is the theoretical upper limit of the number of non-zero singular values ​​in the singular value decomposition;

[0166] Based on the singular value decomposition results, the nonlinear projection value is calculated using the following formula.

[0167]

[0168] Among them, u k (i) is the left singular vector u k The i-th component, v k (j) is the right singular vector v k The j-th component;

[0169] Extract the column vector d of the humidity requirement matrix D of the equipment. j =[D 1j D 2j D Rj ] T The column priority weight ω is calculated using the following formula. j :

[0170]

[0171] Based on multi-parameter fusion weight vector Equipment humidity requirement matrix D and column priority weights ω j The entropy weighting factor ξ is calculated using the following formula. ij :

[0172]

[0173] in, For the multi-parameter fusion weight vector w f The i-th component, D ij Let be the element in the i-th row and j-th column of the equipment humidity requirement matrix D;

[0174] Based on multi-parameter fusion weight vector w f Given the equipment humidity requirement matrix D, the adaptive adjustment factor ζ is calculated using the following formula:

[0175]

[0176] Among them, ||w f ||2 represents the multi-parameter fusion weight vector w f The L2 norm, ||D|| F Let Frobenius norm be the humidity requirement matrix D of the equipment. For the multi-parameter fusion weight vector w f The maximum value of each element;

[0177] Based on nonlinear projection values Entropy weighting factor ξ ij And the adaptive adjustment factor ζ, the target comprehensive humidity value is calculated using the following formula:

[0178]

[0179] Among them, H target The target comprehensive humidity value.

[0180] In one embodiment, the feedback control execution module 105 is further configured to:

[0181] Obtain the humidity data sequence H for the T consecutive sampling periods prior to the current time. t =[h t-T+1 , ..., h t ], where h t Here is the humidity data at time t;

[0182] Calculation of instantaneous humidity change rate based on humidity data sequence And based on the humidity change rate Δh t Generate humidity change curve C actual ;

[0183] The humidity change curve C was calculated using a dynamic time warping algorithm.actual Adjustment curve C with preset target target The similarity δDTW;

[0184] When the similarity δDTW is lower than the preset threshold δ0, the proportional-integral-derivative parameter weight matrix of the feedback control algorithm is updated using the following formula:

[0185] W PID ←W PID -η·(1-δDTW)·ΔW base

[0186] Where η is the learning rate factor, ΔW base This is a preset constant.

[0187] In one embodiment, the data acquisition and processing module 101 is further configured to:

[0188] To address the time-varying, non-steady-state noise characteristics of humidity signals in electrical equipment rooms, a state prediction model adapted to the noise characteristics is constructed.

[0189] Calculate the dynamic deviation between the original humidity signal and the predicted value output by the state prediction model;

[0190] The Kalman filter gain parameter is adjusted in real time based on dynamic bias, and the predicted value is fused with the adjusted Kalman filter gain parameter to generate a humidity estimate as humidity data.

[0191] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the above-described method for controlling the ambient humidity of an electrical equipment room based on difference comparison.

[0192] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0193] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0194] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for controlling the ambient humidity of an electrical equipment room based on difference comparison, characterized in that, The method comprises: Collecting original humidity signal data through a humidity sensing array, and performing noise elimination processing on the original humidity signal data to obtain humidity data; Based on the humidity data, calculate the humidity change rate, and difference compare the humidity change rate with the preset threshold to obtain the environmental humidity complexity; Based on the physical location and humidity demand of the electrical equipment, combined with the environmental humidity complexity, an equipment humidity demand evaluation model is established to generate an equipment humidity demand matrix; Based on the equipment humidity demand matrix, combined with the environmental state parameters and the equipment operation state parameters, a dynamic weight distribution method is used to calculate the target comprehensive humidity value; Based on the target comprehensive humidity value and the humidity data, a feedback control algorithm is used to calculate the deviation, and a humidity control instruction is generated based on the deviation to drive the humidity adjusting equipment to perform operation.

2. The method of claim 1, wherein, Based on the humidity data, the humidity change rate is calculated, and the humidity change rate is compared with the preset threshold by difference to obtain the environmental humidity complexity, which comprises: Based on the humidity data, extract its fluctuation characteristic statistics; Based on the fluctuation characteristic statistics, the time series difference order is optimized by genetic algorithm to determine the optimal difference order N; The humidity change amount of the adjacent N sampling time is calculated by using the difference order N to generate the humidity change rate; According to the statistical distribution of the humidity change rate in the preset sliding window, the first threshold and the second threshold are calculated; Determine whether the humidity change rate is in the interval formed by the first threshold and the second threshold: When the humidity change rate is in the interval, the absolute difference between the humidity change rate and the median of the statistical distribution is calculated; When the humidity change rate is outside the interval, the absolute difference between the humidity change rate and the nearest threshold is calculated; The calculated absolute difference is input into a preset quantization function to map and generate the environmental humidity complexity.

3. The method of claim 1, wherein, Based on the physical location and humidity demand of the electrical equipment, combined with the environmental humidity complexity, an equipment humidity demand evaluation model is established to generate an equipment humidity demand matrix, which comprises: According to the operation parameters of the electrical equipment and the preset humidity tolerance range, the basic humidity demand level of the equipment is generated; Based on the basic humidity demand level of the equipment, an initial weight vector is generated by a preset quantization function; Based on the initial weight vector and the environmental humidity complexity, a dynamic demand weight vector is generated by adaptive adjustment; Establish the position mapping relationship between the physical location of the electrical equipment and the monitoring area of the humidity sensing array; Based on the position mapping relationship and the dynamic demand weight vector, the equipment humidity demand matrix is constructed, wherein the row dimension of the matrix corresponds to the monitoring area, and the column dimension corresponds to the equipment humidity demand priority.

4. The method of claim 1, wherein, Based on the equipment humidity demand matrix, combined with the environmental state parameters and the equipment operation state parameters, a dynamic weight distribution method is used to calculate the target comprehensive humidity value, which comprises: Perform feature extraction on the environmental state parameters to generate environmental dynamic influence factors, and the feature extraction comprises extraction of at least one physical parameter such as environmental temperature gradient or air flow speed variance; The device operation state parameters are analyzed to generate device operation weight coefficients, and the operation characteristic analysis includes analysis of at least one of current harmonic distortion rate, radiator temperature rise rate, and power factor change amount; Based on the device humidity requirement priority in the device humidity requirement matrix, the environment dynamic influence factor and the device operation weight coefficient are fused through a dynamic weight distribution function to generate a multi-parameter fusion weight vector; Based on the multi-parameter fusion weight vector and the device humidity requirement matrix, the target comprehensive humidity value is obtained through correlation operation.

5. The method of claim 4, wherein, The correlation operation obtains the target comprehensive humidity value, including the following steps: a matrix of humidity requirements of the equipment singular value decomposition to obtain singular values σ k left singular vectors and right singular vectors where k = 1, 2,..., r, r = min(R, C) is the theoretical upper limit of the number of non-zero singular values in the singular value decomposition; Based on the singular value decomposition result, a non-linear projection value is calculated using the following formula where u k (i) is the i-th component of the left singular vector u k k (j) is the j-th component of the right singular vector v k .​ extracting column vectors d of the device humidity demand matrix D j = [D 1j , D 2j ,..., D Rj ] T , using the following formula to calculate column priority weight ω j : based on the multi-parameter fusion weight vector the device humidity demand matrix D and the column priority weight ω j The entropy weight factor ξ is calculated using the following formula ij : wherein, is the i-th component of the multi-parameter fusion weight vector w f is the i-th component of the device humidity demand matrix D ij is the element of the device humidity demand matrix D in the i-th row and j-th column; based on the multi-parameter fusion weight vector w f and the device humidity demand matrix D, using the following formula to calculate the adaptive adjustment factor ζ: where ||w f ||2 is the L2 norm of the multi-parameter fusion weight vector w f F ||D||F is the Frobenius norm of the device humidity demand matrix D, where ||w f is the maximum value of each element.​ based on the non-linear projection value entropy weight factor ξ ij and an adaptive adjustment factor ζ, using the following formula to calculate the target integrated humidity value: where H target is the target integrated humidity value.

6. The method of claim 1, wherein, After the driving humidity adjusting device performs the operation, the method further includes: obtain a humidity data sequence H of T continuous sampling periods before the current time t = [h t-T+1 , …, h t ], wherein h t is humidity data at time t; calculating an instantaneous humidity change rate based on the humidity data sequence and based on the humidity change rate Δh t generating a humidity change curve C actual ; calculating the humidity change curve C by a dynamic time warping algorithm actual similarity δDTW with a preset target adjustment curve C target ​ When the similarity δDTW is lower than a preset threshold δ0, the proportional integral derivative parameter weight matrix of the feedback control algorithm is updated using the following formula: W PID ←W PID -η·(1-δDTW)·ΔW base where η is a learning rate factor, ΔW base is a predetermined constant.

7. The method of claim 1, wherein, The noise elimination processing is performed on the original humidity signal data to obtain humidity data, including: A state prediction model is constructed to adapt to the noise characteristics of the time-varying non-steady-state noise of the electrical equipment room humidity signal; A dynamic deviation between the original humidity signal and a predicted value output by the state prediction model is calculated; Based on the dynamic deviation, the Kalman filter gain parameter is adjusted in real time, and the predicted value and the adjusted Kalman filter gain parameter are fused to generate a humidity estimation value as the humidity data.

8. An electrical equipment room environment humidity control system based on difference comparison, characterized by, The system includes: A data acquisition and processing module is configured to acquire original humidity signal data through a humidity sensing array and to perform noise elimination processing on the original humidity signal data to obtain humidity data; An environment complexity analysis module is configured to calculate a humidity change rate based on the humidity data, and to perform difference comparison between the humidity change rate and a preset threshold to obtain an environment humidity complexity degree; A device requirement modeling module is configured to establish a device humidity requirement evaluation model based on a physical location of electrical equipment and humidity requirements, and to combine the environment humidity complexity degree to generate a device humidity requirement matrix; A target humidity decision module is configured to calculate a target comprehensive humidity value based on the device humidity requirement matrix, in combination with environment state parameters and device operation state parameters, and by using a dynamic weight distribution method; A feedback control execution module is configured to calculate a deviation by using a feedback control algorithm based on the target comprehensive humidity value and the humidity data, and to generate a humidity control instruction based on the deviation to drive a humidity adjusting device to perform an operation. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the method of any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 7.

Images