Constant temperature and constant pressure control method and high and low temperature low gas pressure test box

Abstract

This invention provides a constant temperature and pressure control method and a high and low temperature low pressure test chamber. Through high-precision sensor array multi-source data acquisition, signal filtering and feature extraction, combined with improved Granger causality test and coupling spectrum construction, the main propagation path and propagation characteristics of various disturbances are identified. By using sliding window online optimization and dynamic weight matrix, the disturbance suppression mechanism is embedded in the PID loop to achieve multi-channel weighted compensation of proportional, integral and derivative components. This invention can accurately identify and compensate for external disturbances of different paths, time sequences and frequency bands in real time, improve the control accuracy and anti-interference ability of the system, and ensure the stability and robustness of the test chamber operation.

Description

Technical Field

[0001] This invention relates to the field of environmental testing equipment control and multivariable coupling disturbance suppression technology, and particularly to a constant temperature and pressure control method and a high and low temperature low pressure test chamber. Background Technology

[0002] Currently, in the field of constant temperature and pressure control for high and low temperature and low pressure test chambers, the mainstream control strategy mostly adopts feedback control algorithms based on PID (proportional-integral-derivative) regulation, coupled with feedforward compensation or disturbance observation techniques to improve the dynamic response capability and steady-state accuracy of the system. To address the multi-source disturbances (such as ambient temperature, grid voltage, and power quality fluctuations) that occur during the operation of environmental test equipment, the controlled object typically incorporates a disturbance feedforward compensation mechanism. This aims to compensate for known external disturbances in advance, reduce the dynamic lag of the feedback loop, and improve robustness and anti-interference capability under complex operating conditions. In existing technologies, feedforward compensation model design mainly relies on single-variable disturbance path modeling. For example, a common method is to independently model external disturbances such as ambient temperature fluctuations and power supply voltage, and achieve simple advance compensation through empirical parameters or single-channel transfer functions. This type of method is relatively easy to implement and is suitable for some automated testing scenarios dominated by steady-state temperature or pressure. However, as test chamber control tasks develop towards higher precision, stronger dynamics, and multi-field coupling, the limitations of single-variable feedforward models are gradually becoming apparent. They mainly focus on the action chain of a single disturbance on a single controlled variable, failing to reveal the coupling effects of multi-dimensional disturbances such as temperature, air pressure, and power supply in complex thermo-fluid-electrodynamic systems, such as interaction, superposition, mutual amplification, or suppression. Most of the above models assume that each disturbance acts independently and is linearly additive, making it difficult to accurately compensate for path dependence, temporal distortion, and frequency domain component differences in disturbance propagation. In practical applications, high and low temperature low pressure test chambers often face the combined effects of multiple disturbances, such as sudden changes in ambient temperature, short-term ripple in the power grid, and power switching fluctuations. These disturbances propagate within the test chamber structure and control system through physical components such as wall panel heat capacity, heat exchangers, vacuum pump pipelines, and power converters, exhibiting strong nonlinearity, time-varying propagation delay, and frequency response characteristics. For example, sudden changes in ambient temperature can induce changes in chamber pressure through heat conduction, while power grid voltage fluctuations may simultaneously affect the heater's heat output and the throttling valve operation of the refrigeration system, leading to aliasing of same-frequency or different-frequency components in the feedback. Existing feedforward compensation mechanisms, due to their use of simple linear superposition models, cannot accurately characterize the distribution of different disturbance types across multiple control paths and the coupling strength over time and under varying operating conditions. This results in insufficient feedforward compensation accuracy, prolonged system settling time, and decreased robustness. Summary of the Invention

[0003] In order to solve the above-mentioned technical problems, the present invention provides a constant temperature and pressure control method and a high and low temperature low pressure test chamber.

[0004] The technical solution of this invention is implemented as follows: a constant temperature and pressure control method, comprising: S1: Multi-dimensional synchronous acquisition of external environmental disturbance sources of the high and low temperature low pressure test chamber, obtaining disturbance amplitude vectors composed of room temperature fluctuation, grid voltage fluctuation and power supply fluctuation, so as to form a real-time monitoring dataset containing three types of external disturbances. S2: Based on the real-time monitoring dataset obtained from S1, the improved Granger causality test combined with time-frequency coherence analysis method is used to identify the main propagation paths of three types of disturbances—temperature, air pressure, and power supply—in the thermodynamic-fluid dynamic-electrodynamic coupled system of the test chamber, so as to generate a multi-source disturbance coupling spectrum characterizing the intensity and phase delay of the disturbance propagation path. S3: Based on the multi-source disturbance coupling spectrum generated in S2 and the disturbance amplitude vector obtained in S1, calculate the dynamic correlation between disturbance type encoding, disturbance angular frequency and equivalent propagation delay through the spectrum mapping function, so as to output a dynamic weight matrix containing nine time-varying weight coefficients. S4: Embed the dynamic weight matrix output by S3 into the PID feedback loop to perform weighted processing on the temperature deviation signal and generate a weighted error term containing multivariate coupling effects, so as to achieve dynamic compensation of the proportional element of the temperature control channel. S5: Based on the dynamic weight matrix output by S3, the integral quantity of air pressure deviation is weighted to generate a weighted integral term that includes the grid voltage fluctuation compensation factor, so as to realize dynamic compensation of the integral link of the air pressure control channel. S6: The dynamic weight matrix output by S3 is used to perform path weighting on the power supply fluctuation micro-components to generate a weighted differential term containing the room temperature change prediction factor, so as to realize the feedforward differential prediction compensation for power supply disturbances. S7: Determine whether the weighted control terms generated by S4, S5, and S6 exceed the preset weight saturation limit range. If they do, perform weight limiting processing to ensure the system robustness of the multivariable coupled feedforward compensation process. S8: Based on the weighted control terms processed by S7, an online sliding window update mechanism is implemented. The dynamic weight matrix is ​​calculated every five seconds and a smooth transition is implemented to achieve precise suppression of multivariate coupled disturbances by frequency band, path, and time sequence.

[0005] The present invention also provides a high and low temperature and low pressure test chamber, which uses the above-mentioned constant temperature and pressure control method to control the temperature and pressure inside the test chamber.

[0006] The constant temperature and pressure control method and high and low temperature low pressure test chamber provided by this invention have the following beneficial effects: (1) This invention constructs a dynamic graph model that characterizes the real propagation path of three types of external disturbances—temperature, air pressure, and power supply—in a thermodynamic-fluid dynamic-electrodynamic composite system. This model accurately depicts the energy decay, temporal distortion, and nonlinear distortion behavior of disturbances. Based on this, a graph-driven dynamic weight generator is designed to output a time-varying weighted matrix in real time. This enables the PID controller to adaptively adjust the response sensitivity of each control channel to errors and their historical and prospective information according to the current disturbance type and intensity. This significantly improves the anti-interference capability and adjustment accuracy of the control system in complex multivariable strongly coupled environments, and effectively ensures the stability and reproducibility of the test environment.

[0007] (2) This invention introduces a coupled path identification method based on multi-source synchronous sampling and improved Granger causality test, which realizes the extraction and quantitative modeling of the main propagation path from the original disturbance input to the response of the key controlled variable, ensuring the lightweight nature and deployability of the algorithm. Combined with the sliding window online update mechanism and the weight saturation limiting strategy, the time-varying weighted matrix is ​​optimized every 5 seconds and a smooth transition is implemented. This not only ensures the control parameters' ability to quickly track the evolution of the operating conditions, but also prevents system oscillations caused by sudden weight changes, significantly enhancing the dynamic stability and robustness of the control process. (3) The graph mapping function proposed in this invention and the reconstructed weighted PID control law form a closed-loop enhancement mechanism of "disturbance perception - path analysis - dynamic weighting - precise compensation". It not only realizes the functional extension of the traditional PID structure rather than its replacement, but also retains its advantages of high real-time performance and industrial compatibility. It also gives it the ability to adapt to the coupling characteristics of complex systems. The whole scheme does not need to introduce complex architectures such as fuzzy rule sets, double-loop observers, Koopman operators or physical guided neural networks, nor does it rely on fixed decoupling matrices or preset disturbance intervals. It has good versatility and portability and is suitable for various environmental simulation equipment control scenarios with multi-field coupling effects. In particular, it achieves high-precision temperature and pressure joint regulation under extreme low enthalpy conditions. Attached Figure Description

[0008] Figure 1 This is a flowchart of a constant temperature and constant pressure control method according to the present invention; Figure 2 This is a sub-flowchart of a constant temperature and pressure control method according to the present invention; Figure 3 This is another sub-flowchart of a constant temperature and pressure control method according to the present invention. Detailed Implementation

[0009] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0010] The following disclosure provides many different embodiments or examples for implementing different structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. Of course, these are merely examples and are not intended to limit the invention. Furthermore, reference numerals and / or letters may be repeated in different examples; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0011] like Figure 1 As shown, the present invention provides a constant temperature and constant pressure control method, specifically including: S1: Multi-dimensional synchronous acquisition of external environmental disturbance sources of the high and low temperature low pressure test chamber, obtaining disturbance amplitude vectors composed of room temperature fluctuation, grid voltage fluctuation and power supply fluctuation, so as to form a real-time monitoring dataset containing three types of external disturbances. S2: Based on the real-time monitoring dataset obtained from S1, the improved Granger causality test combined with time-frequency coherence analysis method is used to identify the main propagation paths of three types of disturbances—temperature, air pressure, and power supply—in the thermodynamic-fluid dynamic-electrodynamic coupled system of the test chamber, so as to generate a multi-source disturbance coupling spectrum characterizing the intensity and phase delay of the disturbance propagation path. S3: Based on the multi-source disturbance coupling spectrum generated in S2 and the disturbance amplitude vector obtained in S1, calculate the dynamic correlation between disturbance type encoding, disturbance angular frequency and equivalent propagation delay through the spectrum mapping function, so as to output a dynamic weight matrix containing nine time-varying weight coefficients. S4: Embed the dynamic weight matrix output by S3 into the PID feedback loop to perform weighted processing on the temperature deviation signal and generate a weighted error term containing multivariate coupling effects, so as to achieve dynamic compensation of the proportional element of the temperature control channel. S5: Based on the dynamic weight matrix output by S3, the integral quantity of air pressure deviation is weighted to generate a weighted integral term that includes the grid voltage fluctuation compensation factor, so as to realize dynamic compensation of the integral link of the air pressure control channel. S6: The dynamic weight matrix output by S3 is used to perform path weighting on the power supply fluctuation micro-components to generate a weighted differential term containing the room temperature change prediction factor, so as to realize the feedforward differential prediction compensation for power supply disturbances. S7: Determine whether the weighted control terms generated by S4, S5, and S6 exceed the preset weight saturation limit range. If they do, perform weight limiting processing to ensure the system robustness of the multivariable coupled feedforward compensation process. S8: Based on the weighted control terms processed by S7, an online sliding window update mechanism is implemented. The dynamic weight matrix is ​​calculated every five seconds and a smooth transition is implemented to achieve precise suppression of multivariate coupled disturbances by frequency band, path, and time sequence.

[0012] Step S1: Multi-dimensional synchronous acquisition of external environmental disturbance sources of the high-low temperature and low-pressure test chamber to obtain disturbance amplitude vectors composed of room temperature fluctuations, grid voltage fluctuations, and power supply fluctuations, thereby forming a real-time monitoring dataset containing three types of external disturbances. Specifically, this includes: S1.1: Deploy a high-precision sensor array for room temperature disturbance sources, grid voltage disturbance sources and power supply disturbance sources in the external environment of the high and low temperature low pressure test chamber. Obtain the original simulated room temperature signal based on thermocouple temperature sensor, obtain the original simulated grid voltage signal based on Hall voltage sensor, and obtain the original simulated power supply fluctuation signal based on current transformer, so as to form the basis for physical quantity detection of the three types of disturbance sources. Deploy a high-precision thermocouple temperature sensor network to detect room temperature disturbances in the external environment of the high and low temperature low pressure test chamber (parameter: measurement range). With a temperature range of 100℃ to 200℃, a resolution of 0.01℃, and a response time of <0.5s, the system accurately acquires the original analog signal at room temperature. A four-wire wiring method reduces conductor resistance errors, and a shielded twisted-pair structure suppresses electromagnetic coupling interference, ensuring the purity and stability of the thermocouple temperature signal. Furthermore, by deploying a Hall effect-based voltage sensor (parameters: measurement range 0V to 500V, sensitivity 2.5mV / V, bandwidth 100kHz) at the main power grid input line, high-frequency response acquisition of the raw analog signal of the power grid voltage is achieved. The detected signal is transmitted to the data acquisition module through a differential output interface to suppress common-mode interference and maintain signal amplitude integrity. Furthermore, by connecting a precision current transformer (parameters: rated primary current 0-50A, secondary output current 1A, accuracy class 0.2, bandwidth 20kHz) in series at the power input of the test chamber, real-time monitoring of the original analog signal of power fluctuations is achieved. An open-loop iron core structure is adopted to ensure high bandwidth response, and a load resistor is configured on the secondary side to convert the current signal into voltage, which facilitates subsequent data acquisition. Furthermore, by standardizing the interfaces of thermocouples, Hall voltage sensors and current transformer output terminals, selecting electromagnetic interference-resistant BNC interfaces and calibrating the zero-point drift of each channel, the interface consistency and reference consistency of the detection of physical quantities of the three types of disturbance sources are ensured. By selecting the above sensors and acquiring the signals, the results of the previous step are transformed into high-confidence original simulated signals of room temperature, grid voltage, and power fluctuation, which form the basis for detecting the physical quantities of external disturbance sources and provide accurate input data for subsequent synchronous sampling and digital processing of signals. For example, in an environmental reliability test of avionics equipment, the external temperature of the test chamber was maintained within the range of 20℃±3℃, and the thermocouple sensor range was configured as follows: The temperature range is 50℃ to 150℃, with a resolution set to 0.01℃. The power grid is three-phase 380V. The connected Hall voltage sensor has a range of 0V to 500V, a sensitivity of 2.5mV / V, and a bandwidth of 80kHz. The power supply is single-phase 220V with a maximum rated current of 15A. The current transformer used has a range of 0-20A and an accuracy class of 0.2. During data acquisition, the grounding resistance of the thermocouple output shield is 0.2Ω to ensure low-impedance grounding; the Hall sensor differential interface suppresses common-mode noise to within 0.05V; the current transformer load resistance is configured to 50Ω to convert the 1A secondary current into a 50V analog voltage signal for input to the acquisition module. Under the above configuration, the peak amplitude of the room temperature disturbance source signal is 0.8℃, the peak amplitude of the grid voltage fluctuation signal is 5.2V, and the peak amplitude of the power supply fluctuation signal is 3.8V. According to the preliminary signal quality assessment, the signal-to-noise ratio of the three types of disturbance signals is significantly improved, which meets the accuracy requirements of subsequent synchronous sampling and feature extraction. S1.2: Based on the original room temperature analog signal, the original grid voltage analog signal, and the original power fluctuation analog signal obtained in S1.1, perform synchronous sampling and analog-to-digital conversion processing, and use the isochronous triggering mechanism of the industrial-grade data acquisition card to perform time-aligned sampling of the three signals to generate a digital sampling sequence with consistent timestamps; Based on the original room temperature analog signal, the original grid voltage analog signal, and the original power fluctuation analog signal obtained by S1.1, a multi-channel isochronous trigger sampling method (parameters: three analog signal input channels, sampling rate 20kHz, trigger mode is hardware-level synchronous triggering) is adopted to achieve synchronous sampling of the three signals at the same sampling time. Furthermore, through high precision A novel analog-to-digital conversion algorithm (parameters: 24-bit quantization, 256 oversampling ratio) is used to achieve high-fidelity conversion of analog signals to digital signals, and to obtain digital sequences of room temperature, grid voltage, and power fluctuation. Furthermore, by using a timestamp alignment algorithm (parameters: reference clock is the local crystal oscillator frequency of the data acquisition card at 10MHz, synchronization deviation threshold ≤0.5μs), the time index of the three digital sampling data is unified, generating a sequence with consistent timestamps; Furthermore, a data frame structured encapsulation process (parameters: single frame length 50ms, containing three signal samples) is adopted to realize the batch encapsulation of sampled data and generate a data frame set with time alignment attributes; By using the above-mentioned multi-channel isochronous triggering sampling, analog-to-digital conversion, and time alignment processing methods, the physical quantity analog signal of the previous step is converted into a digital sampling sequence with consistent timestamps, thus realizing the digital signal foundation required for subsequent feature extraction and filtering processes. For example, in a certain environmental reliability test scenario, the raw analog signal at room temperature is input to the thermocouple channel, with a sampling rate set to 20kHz. The grid voltage and power supply fluctuation signals are respectively connected to the Hall sensor and current transformer channels. During the data acquisition phase, the industrial-grade data acquisition card enables hardware synchronization triggering to ensure that the three input signals are acquired at the same sampling time. During the analog-to-digital conversion process, [the following is used]. This ADC features a 24-bit quantization bit depth and an oversampling ratio of 256, ensuring complete quantization of low-frequency, weakly disturbed signals. During the time synchronization process, a 10MHz local crystal oscillator is used as the reference clock, and a timestamp metadata is appended to each sample. When the detection deviation exceeds... The correction algorithm is executed in real time, ultimately outputting a digital sampling sequence with three identical timestamps. In the data encapsulation stage, 50ms long samples are packaged into data frames, each containing three signal samples (room temperature, grid voltage, and power supply fluctuation) and unified timestamp information. After this step, the output digital sampling sequence is perfectly aligned in the time domain, and its fidelity meets the accuracy and stability requirements of subsequent Butterworth filtering and sliding window standard deviation feature extraction. S1.3: Apply Butterworth bandpass filtering algorithm to the digital sampling sequence with consistent timestamps generated in S1.2. Remove power frequency interference and DC drift components by setting the passband range from 0.01Hz to 10Hz to obtain an effective signal sequence that characterizes the disturbance. S1.4: Based on the effective signal sequence representing the disturbance characteristics obtained in S1.3, calculate the instantaneous fluctuation amplitude. Use the sliding window standard deviation algorithm to calculate the room temperature fluctuation for the effective signal sequence, the grid voltage fluctuation for the effective signal sequence, and the power supply fluctuation for the effective signal sequence, so as to generate scalar fluctuation values ​​for the three types of disturbances. S1.5: Combine the scalar values ​​of room temperature fluctuation, grid voltage fluctuation, and power supply fluctuation generated in S1.4 into a three-dimensional disturbance amplitude vector. Store the vector data in 50ms cycles based on the timestamp marking mechanism of the real-time operating system to form a real-time monitoring dataset containing three types of external disturbances.

[0013] Step S2: Based on the real-time monitoring dataset obtained in S1, an improved Granger causality test combined with time-frequency coherence analysis is used to identify the main propagation paths of three types of disturbances—temperature, air pressure, and power supply—in the thermodynamic-fluid dynamic-electrodynamic coupled system of the test chamber, in order to generate a multi-source disturbance coupling spectrum characterizing the intensity and phase delay of the disturbance propagation path. Specifically, this includes: S2.1: Based on the standard Granger causality test framework, an integrated time-frequency coherence analysis module is used to perform cross-spectral density calculation and phase coherence analysis on multivariate disturbance time series to generate an improved Granger causality test model capable of quantifying nonlinear distortion. The input to this step is the standard Granger causality test algorithm and the multivariate disturbance time series. By performing time-frequency coherence analysis, the coherence function and phase difference between each disturbance source and the key controlled variable are calculated, and the output is an improved Granger causality test model that includes nonlinear distortion quantification capability. Based on the time-stamped, multivariate disturbance digital sampling sequence obtained by S1, the time series of key controlled variables such as the temperature at the center point of the test chamber and the average cavity pressure are selected as the target variable set, and a data matrix is ​​constructed for causal analysis. The standard Granger causality test method (parameter: the test lag order p is taken as the optimal value according to the Akaike Information Criterion (AIC)) is adopted to establish a multivariate linear prediction model between multivariate disturbances and key controlled variables, so as to realize the statistical judgment of whether each disturbance variable can significantly improve the prediction accuracy of the target variable; Furthermore, by using the cross-spectral density calculation method (parameter: frequency resolution 0.01Hz), frequency domain correlation analysis was performed on the disturbance time series and the controlled quantity time series to obtain amplitude spectrum and phase spectrum data at different frequency points, which were used to explore the performance of disturbance propagation characteristics in the frequency domain. Furthermore, a time-frequency coherence analysis method (parameters: window length 2 seconds, overlap rate 50%) is used to calculate the coherence function between each disturbance source and the key controlled variable. and phase difference Furthermore, a coherence threshold is introduced to screen significant time-frequency points, enabling time-segmented characterization of the instantaneous correlation between disturbances and responses; Furthermore, a nonlinear distortion calculation is performed on the coherence function and phase difference data, using the distortion formula:

[0014] in, For coherent function values, and These represent the lower and upper limits of the analysis frequency range, respectively, and the degree of nonlinear deviation of the disturbance in the frequency spectrum is quantified by this integral. By integrating the Granger causality test results, cross-spectral density analysis results, and time-frequency coherence analysis data, an improved Granger causality test model is constructed to simultaneously quantify causal strength and nonlinear distortion, and to provide a unified time-frequency domain criterion for subsequent main propagation path selection. This improved model transforms the above multi-source analysis results into a model structure that includes causality index, frequency domain gain, and distortion parameters, thereby achieving a full-domain characterization of the dependence of disturbances and controlled variables. For example, under the experimental conditions of -20℃ and cavity pressure of 5kPa, the sampling rates for room temperature fluctuations, grid voltage fluctuations, and power supply fluctuations were all 20Hz. The data window length was set to 60 seconds, and the lag order p was determined to be 5 using the AIC method. A Granger causality test was performed on the room temperature fluctuations and the center point temperature, yielding a causality index of 0.34 and a significance level less than 0.01. Cross-spectral density analysis of room temperature and center point temperature showed an amplitude spectral gain of 1.8 times at 0.3Hz, with a phase lag of 23°. In time-frequency coherence analysis, the coherence function remained stable above 0.85 in the 0.1~0.5Hz range, with the phase difference fluctuating within this range by no more than ±30°. The calculated nonlinear distortion integral was 0.12, indicating that the room temperature disturbance exhibited low-distortion propagation within the main propagation frequency band. Based on all indicators, the causal path from room temperature perturbation to the center point temperature in the improved Granger causal model has a high priority, and this path will be given priority in the selection of the main propagation path as the basis for the construction of the dynamic weight matrix. S2.2: Based on the disturbance amplitude vector time series obtained in S1 and the improved Granger causality test model generated in S2.1, perform causality strength calculation and path screening to identify the main propagation paths from three types of disturbances—temperature, air pressure, and power supply—to the key controlled variables. The inputs to this step are the disturbance amplitude vector time series and the improved Granger causality test model. By calculating the Granger causality index and combining it with the time-frequency coherence analysis results, the path priority is sorted, and the set of main propagation paths is output. S2.3: For the set of main propagation paths identified in S2.2, calculate the path gain, group delay, and nonlinear distortion of each path based on the time-frequency coherence analysis results to generate a set of disturbance propagation characteristic parameters. The input to this step is the set of main propagation paths and time-frequency coherence analysis data. By quantifying the energy attenuation and timing distortion characteristics during the disturbance propagation process, the output is a set of disturbance propagation characteristic parameters including path gain, group delay, and nonlinear distortion. S2.4: Based on the disturbance propagation characteristic parameter set generated in S2.3, construct a three-dimensional coupling map characterizing the disturbance angular frequency, equivalent propagation delay, and disturbance type encoding to generate a multi-source disturbance coupling map; the input of this step is the disturbance propagation characteristic parameter set, and by mapping the disturbance angular frequency, equivalent propagation delay, and disturbance type encoding to three-dimensional space, the output is a multi-source disturbance coupling map characterizing the disturbance propagation path strength and phase delay; S2.5: By injecting controlled step disturbance data and comparing the predicted response with the actual response, verify the accuracy of the multi-source disturbance coupling spectrum generated in S2.4 to ensure that it can truly reflect the disturbance propagation characteristics. The input of this step is the multi-source disturbance coupling spectrum and the controlled step disturbance data. By performing a comparative analysis of the disturbance propagation prediction and the measured data, the verified multi-source disturbance coupling spectrum is output.

[0015] like Figure 2 As shown, step S3 involves calculating the dynamic correlation between the perturbation type encoding, perturbation angular frequency, and equivalent propagation delay based on the multi-source perturbation coupling map generated in S2 and the perturbation amplitude vector obtained in S1, using a map mapping function, to output a dynamic weight matrix containing nine time-varying weight coefficients. Specifically, this includes: S3.1: Based on the path gain, group delay and nonlinear distortion data stored in the multi-source perturbation coupling spectrum, a spectrum mapping function is constructed. This function takes the perturbation type code, perturbation angular frequency and equivalent propagation delay as configuration parameters, takes the perturbation amplitude vector as input variable, and outputs the calculation basis of dynamic weight coefficients to realize the parameterized characterization of perturbation propagation characteristics. Based on the path gain, group delay, and nonlinear distortion data stored in the multi-source perturbation coupling spectrum output by S2, a parametric modeling method (input parameters: perturbation type encoding, perturbation angular frequency, and equivalent propagation delay) is used to realize the initialization structure of the spectrum mapping function. Furthermore, by using a weighted path fitting algorithm (parameters: path gain weight α, group delay weight β, nonlinear distortion weight γ), the perturbation characteristic vector is mapped in a multidimensional space, and a preliminary family of weight coefficient functions is obtained. Furthermore, a combination of multinomial regression and piecewise logical fitting is adopted (parameters: fitting times n=3, piecewise threshold based on perturbation angular frequency distribution) to achieve continuous and smooth processing of the family of weight coefficient functions and generate a closed analytical form suitable for real-time calculation. Furthermore, a time-frequency domain normalization processing method is introduced (parameter: the normalization benchmark is the path gain of each perturbation type code at the reference frequency) to achieve the dimension unification of the weight coefficient function corresponding to different perturbation type codes, and to generate a standardized weight function F that can be directly used for dynamic calculation; By constructing a graph mapping function, the combined characteristics of path gain, group delay, and nonlinear distortion are transformed into the basis for calculating dynamic weight coefficients, thereby realizing the parameterized characterization of multi-source disturbance propagation characteristics. For example, in a high-low temperature and low-pressure test chamber, a combined disturbance test scenario involving a room temperature change of ±6℃, a grid voltage fluctuation of ±4%, and a power supply fluctuation of 2% is performed. The disturbance type code T=[1,2,3] corresponds to temperature, air pressure, and power supply, respectively. The disturbance angular frequency range is [0.05Hz, 5Hz], and the equivalent propagation delay range is [0.1s, 1.2s]. The path gain vector G=[0.82, 0.65, 0.74], group delay vector τ=[0.35, 0.78, 0.66], and nonlinear distortion vector D=[0.12, 0.18, 0.14] output from S2 are input into the weighted path fitting formula:

[0016] in =0.4、 =0.35、 =0.25, and the initial weight coefficient vector W=[0.53,0.58,0.56] was calculated. This vector was fitted to the disturbance angular frequency range through cubic polynomial regression to obtain the frequency response weight function curve, and then normalized to form a standardized weight function F, so that the weight calculation of different disturbance types is carried out under a unified dimension. Example verification shows that the dynamic weight coefficients generated by this mapping function under the combined action of three types of disturbances can significantly improve the feedforward compensation accuracy, significantly reduce the overshoot of the temperature channel, and synchronously reduce the pressure coupling error, meeting the technical expectations of high-precision control. S3.2: Input the disturbance amplitude vector composed of the room temperature fluctuation, grid voltage fluctuation and power supply fluctuation obtained in S1, together with the multi-source disturbance coupling spectrum generated in S2, into the spectrum mapping function to form the input parameter set of the mapping function, providing a data basis for the calculation of dynamic weight coefficients; Based on the room temperature fluctuation, grid voltage fluctuation and power supply fluctuation obtained by S1, a three-dimensional disturbance amplitude vector is constructed. A real-time data reading interface (parameters: sampling period 50ms, data buffer length 20 sampling points) is selected to synchronously load the three types of disturbance amplitude vectors with the multi-source disturbance coupling spectrum generated by S2 to achieve consistent alignment between input data and spectrum parameters. An input parameter correction algorithm (parameters: normalization range [-1,1], missing value imputation strategy is linear prediction) is used to normalize the amplitude and align the timestamps of the three-dimensional perturbation amplitude vector, so as to achieve scale uniformity and temporal consistency of input variables; Furthermore, through the spectrum parameter analysis module (parameters: disturbance type encoding mapping table, angular frequency range from 0.01Hz to 10Hz, equivalent propagation delay unit ms), the path gain, group delay, and nonlinear distortion features corresponding to each disturbance type encoding, disturbance angular frequency, and equivalent propagation delay in the multi-source disturbance coupling spectrum are extracted to form a structured parameter set that matches the disturbance amplitude vector; Furthermore, a parameter set fusion algorithm (parameters: matrix dimension 3×N, initial fusion weight value 1) is used to merge the normalized perturbation amplitude vector with the structured graph parameter set column by column to generate a fusion data matrix containing input perturbation features and propagation characteristic parameters, so as to ensure the integrity of the input for subsequent graph mapping function calculation; By using the above parameter fusion processing method, the multi-source disturbance coupling spectrum and real-time disturbance amplitude vector are transformed into an input parameter set containing disturbance type encoding, disturbance angular frequency and equivalent propagation delay, thus achieving the expected technical effect of providing a highly consistent data foundation for dynamic weight coefficient calculation. For example, in a high-low temperature and low-pressure test chamber control example, the room temperature fluctuation is 0.85℃, the mains voltage fluctuation is 3.2V, the power supply fluctuation is 0.18A, the sampling period is set to 50ms, and the data buffer length is 20 sampling points. After normalizing the three-dimensional disturbance amplitude vector, the room temperature fluctuation is normalized to... The grid voltage fluctuation is normalized to Power fluctuations normalized to Analyze the multi-source perturbation coupling map to obtain the perturbation type encoding vector. Corresponding angular frequency Hz, equivalent propagation delays are respectively The model is analyzed to obtain the feature matrix consisting of path gain, group delay, and nonlinear distortion. The fusion process generates a 3×6 fused data matrix, with elements containing normalized perturbation amplitudes and corresponding propagation characteristic parameters. This parameter set is input to the spectral mapping function, which outputs the pre-calculation data for weight calculation. Experimental results verify that its consistency and stability meet the requirements for dynamic weight matrix calculation, maintaining the accuracy of weight coefficient calculation under different external perturbation conditions. S3.3: Based on the input parameter set, the dynamic correlation between the disturbance type code and the disturbance angular frequency is calculated using the spectral mapping function. The disturbance propagation intensity coefficient is generated by multiplying the path gain and the group delay to quantify the impact of disturbances of different frequencies on the control channel. Based on the input parameter set formed by S3.2, a graph mapping function (configuration parameters: perturbation type encoding, perturbation angular frequency, equivalent propagation delay) is used to realize the dynamic parameter matching function between different perturbation types and corresponding perturbation angular frequencies. Furthermore, by using the path gain retrieval algorithm (parameter source: path gain data in the multi-source disturbance coupling spectrum), the gain coefficient of each type of disturbance on the corresponding control channel is extracted, and the energy amplification or attenuation during the disturbance propagation process is obtained. Furthermore, by using a group delay retrieval algorithm (parameter source: group delay data in the multi-source disturbance coupling map), the hysteresis characteristics of the target disturbance in the control channel are quantified, and the time offset of the disturbance effect is obtained. Furthermore, the perturbation propagation strength coefficient is calculated using a product operation method (input: path gain and group delay), specifically according to the following formula:

[0017] in, The disturbance propagation intensity coefficient, This is the path gain coefficient. Group delay; Furthermore, by using the frequency mapping analytical method (parameters: disturbance angular frequency and disturbance propagation intensity coefficient), the influence intensity of disturbances of different frequency components on the three stages of the control channel, namely proportional, integral, and derivative, is quantified, and a frequency-intensity correspondence table is generated. Through the above calculation logic, the calculation results of the spectrum mapping function, path gain and group delay coefficient are transformed into disturbance propagation intensity data that can be used to generate weight coefficients, so as to achieve the expected technical effect of quantifying the impact of multi-band disturbances. For example, during the operation of the high and low temperature low pressure test chamber, the input parameter set is configured as follows: temperature disturbance type code is 1, and angular frequency is set to... Hz, equivalent propagation delay is s, path gain extraction value Group delay extraction value s. The product operation formula: The disturbance propagation intensity coefficient is obtained as The coefficient and the perturbation angular frequency are input together into a frequency mapping analytical algorithm to obtain the intensity in the proportional channel. The integral channel strength is The intensity of the differential channel is Actual operating results show that, under this mapping parameter, the PID control significantly suppresses temperature fluctuations during the temperature channel response process, reduces the steady-state deviation to less than half of the original level, and also significantly improves the coupling error of the air pressure control. S3.4: Based on the disturbance propagation intensity coefficient, combined with the equivalent propagation delay and nonlinear distortion parameters, the mapping relationship between the equivalent propagation delay and the weight coefficient is calculated through a time-frequency domain weighting algorithm to generate a time-varying weight coefficient vector to reflect the temporal characteristics of disturbance propagation. S3.5: Organize the time-varying weight coefficient vector into a 3×3 structure to form a dynamic weight matrix, where each element represents the instantaneous impact weight of the j-th type of disturbance on the i-th PID control channel. Output this matrix for subsequent weighted control.

[0018] like Figure 3 As shown, step S4 involves embedding the dynamic weight matrix output from S3 into the PID feedback loop to weight the temperature deviation signal, generating a weighted error term that includes multivariate coupling effects, thereby achieving dynamic compensation of the proportional element in the temperature control channel. Specifically, this includes: S4.1: Perform Butterworth bandpass filtering on the temperature deviation signal of the high and low temperature low pressure test chamber. Based on the temperature disturbance main path parameters in the dynamic weight matrix generated by S3, set the passband range from 0.05Hz to 5Hz to remove temperature sensor noise and DC drift components, and generate a standard error sequence characterizing the effective temperature disturbance. S4.2: Based on the standard error sequence generated in S4.1 and the dynamic weight matrix output in S3, the first row and first column of the matrix are extracted as temperature disturbance weight coefficients. The instantaneous product of these coefficients and the standard error sequence is calculated using the time-frequency domain weighted mapping algorithm to generate the dynamic compensation vector of the main path of temperature disturbance. Based on the standard error sequence generated by S4.1 and the dynamic weight matrix output by S3, the matrix element retrieval method (parameter: matrix row and column index [1,1]) is used to extract the main path weight coefficients corresponding to temperature perturbation from the dynamic weight matrix. Furthermore, by using a time-frequency domain weighted mapping algorithm (parameters: frequency range 0.05Hz to 5Hz, time delay range 0.1s to 1.0s), the frequency band and time series correlation model of the weight coefficients and the standard error sequence is realized, and the instantaneous weighted factor sequence is obtained; Furthermore, an instantaneous product operation method is adopted (parameters: input is the standard error sequence E(t) and the instantaneous weighted factor sequence). The calculation of the compensation amount for the main path of temperature disturbance is performed time-by-time, and the formula is as follows:

[0019] in, For dynamic compensation amount, This is the temperature disturbance weighting coefficient. The standard error sequence value; Furthermore, a sliding window smoothing filtering algorithm is used (parameters: window length is 5 sampling points, smoothing coefficient...). =0.65), to realize the transition vibration suppression processing of the dynamic compensation vector in the time domain, and generate a smooth temperature main path dynamic compensation vector; Furthermore, an amplitude normalization processing method (parameter: target interval [-1,1]) is adopted to unify the dimensions of the smoothing compensation amount, ensuring that the compensation amount of other control channels is numerically comparable in subsequent weighted calculations; Through the above time-frequency domain mapping and product smoothing, the standard error sequence of the previous step is transformed into a dynamic compensation vector that reflects the impact of instantaneous disturbances in the main temperature path, thereby achieving the expected technical effect of precise dynamic compensation for the proportional element. For example, in a high and low temperature low pressure test chamber operating at -60℃ / 1.2kPa, with a sampling period of 50ms, the standard error sequence... The amplitude range within this period is [-0.95, 0.88], and the value of the element in the first row and first column of the dynamic weight matrix ranges from 0.82 to 1.16. The matrix elements are used as temperature perturbation weight coefficients and input into the time-frequency domain weighted mapping algorithm. The frequency range is set to 0.05Hz-5Hz, and the time delay range is set to 0.15s-0.95s, resulting in an instantaneous weighted factor sequence. The range is [0.84, 1.12]. Perform the instantaneous product operation formula:

[0020] At some point , =0.72, =0.94, then =0.6768. (The rest of the text appears to be a mix of characters and symbols, possibly representing numbers or symbols.) The sequence was smoothed using a sliding smoothing process with a window length of 5 (smoothing coefficient 0.65), and the output smoothing compensation value stabilized within the range of [-0.65, 0.74]. After normalization and mapping to the [-1, 1] interval, it was weighted and fused with other compensation values. Actual operating results show that the overshoot of the temperature channel was significantly reduced, and the steady-state temperature fluctuation amplitude tended to be significantly stabilized, meeting the dynamic compensation performance requirements of the proportional element under strong disturbance conditions. S4.3: Perform sliding window differentiation on the scalar values ​​of room temperature fluctuation and grid voltage fluctuation obtained in S1. Based on the dynamic compensation vector and the elements of the first row and second column and the first row and third column of the dynamic weight matrix generated in S4.2, perform path weighted summation to generate a disturbance compensation superposition term that includes multivariate coupling effects. Based on the scalar values ​​of room temperature fluctuation and grid voltage fluctuation obtained by S1, a sliding window differential algorithm (parameters: window length L=5, sampling period 50ms, difference order 1) is used to calculate the rate of change of the fluctuation signal in the time domain in order to extract the short-term disturbance change trend characteristics. Furthermore, by using a differential smoothing filtering method (parameters: the weighted coefficient distribution conforms to a triangular window [0.2, 0.4, 0.8, 0.4, 0.2]), the differential sequence is smoothed, the instantaneous spike noise introduced by the original differential process is suppressed, and the smoothed room temperature change rate vector and grid voltage change rate vector are obtained. Furthermore, based on the dynamic compensation vector generated by S4.2, the elements in the first row and second column of the dynamic weight matrix are extracted. with the element in the first row and third column As path weighting coefficients, they correspond to the room temperature disturbance weighting coefficient and the grid voltage disturbance weighting coefficient, respectively; Furthermore, a path-weighted summation method is adopted (the formula is as follows:)

[0021] in, Let the vector be the rate of change of room temperature. (For the grid voltage change rate vector), this method enables weighted superposition calculation of the two types of disturbance change rates, and generates a disturbance compensation superposition term that includes multivariate coupling effects. ; Furthermore, through vector normalization processing (parameter: normalization interval [...]), 1,1]), to achieve dimensional uniformity of the disturbance compensation superposition term, ensuring that the data can be processed by linear combination of proportional elements with the standard error sequence; Through the above chain derivation process, the sliding window differential result and the auxiliary path weight in the dynamic weight matrix are weighted and summed to transform into a disturbance compensation superposition term that characterizes the multivariate coupling compensation effect of room temperature and grid voltage, thereby realizing high-precision disturbance information input for proportional compensation. For example, in a high-low temperature and low-pressure test, the scalar value of room temperature fluctuation is 0.84℃, the scalar value of grid voltage fluctuation is 2.8V, the sampling period is set to 50ms, and the window length is 5. After executing the sliding window differential algorithm, the mean of the room temperature change rate vector is 0.0168℃ / sampling point, and the mean of the grid voltage change rate vector is 0.056V / sampling point. After triangular window weighted smoothing, the noise peak amplitude is reduced to one-quarter of the original. The element in the first row and second column of the dynamic weight matrix is ​​extracted as 1.12, and the element in the first row and third column is extracted as 0.87. Substituting these values ​​into the weighted summation formula:

[0022] The disturbance compensation superposition term was calculated. It is 0.064. Normalized to [ After the interval [1,1], the compensation superposition term increases the instantaneous deviation correction of the temperature control channel by about 0.064 units in the subsequent proportional link mapping conversion. The measured results show that the fluctuation of the temperature response curve is effectively suppressed and the steady-state temperature difference is maintained within 0.3℃. S4.4: Perform vector linear combination processing on the standard error sequence generated in S4.1 and the disturbance compensation superposition term generated in S4.3. Construct a weighted coefficient set based on the elements of the first row of the dynamic weight matrix. Perform proportional link mapping transformation on the combination result to output the weighted error term containing multivariate coupling effects to the temperature control channel. The standard error sequence generated by S4.1 and the disturbance compensation superposition term generated by S4.3 are combined using a vector linear combination algorithm (parameters: input vector length N=1024, operation precision 64-bit floating point) to achieve element-by-element combination processing of the two temperature control related signals. Furthermore, a method for constructing a weighted coefficient set is used (parameter source: the set of elements in the first row of the dynamic weight matrix). This allows for the extraction and one-dimensional arraying of the three-channel weights in the proportional circuit, so that subsequent combination operations can be directly indexed and called. Furthermore, a time-frequency domain mapping transformation algorithm (parameters: mapping window length 50ms, frequency resolution 0.1Hz) is adopted to fit the frequency domain correspondence between the weighted coefficient set and the standard error sequence and the perturbation compensation superposition term, and to obtain the coefficient sequence corrected by the frequency response. Furthermore, this can be achieved through a combination of matrix multiplication and vector summation (the formula is as follows:)

[0023] in, The output vector is the weighted error term. For the weighted coefficient set, The standard error sequence, (For disturbance compensation superposition), to achieve rapid calculation of dynamic compensation signal of proportional element; Furthermore, using the proportional element mapping function (parameter: proportional gain) Derived from the initial design value of the controller (with a limit range of 0.1 to 2.5), the combined calculation result is mapped to the proportional input signal that adapts to the PID control law, and a final weighted error term dataset containing multivariate coupling effects is formed. By using the above linear combination and proportional mapping processing method, the standard error sequence and disturbance compensation superposition term are transformed into a weighted error term that meets the input requirements of the proportional element of the controller, thereby achieving the expected technical effect of dynamic compensation for the proportional element of the temperature control channel. For example, in a certain model of high and low temperature low pressure test chamber operating at... Under conditions of 60℃ and 1.5kPa pressure, the standard error sequence length is set to 1024, the sampling rate is 200Hz, and the disturbance compensation superposition term is of uniform length after sliding window processing. The first row of the dynamic weight matrix has the following values: =1.25、 =0.85、 =1.05, and after constructing a weighted coefficient set, the array [1.25, 0.85, 1.05] is obtained. Under the conditions of a mapping window length of 50ms and a frequency resolution of 0.1Hz, a time-frequency domain mapping transformation is performed, which increases the gain of the standard error by 3% at the center frequency of 0.5Hz and attenuates the compensation superposition term by 2% at 1.2Hz. Matrix multiplication and vector summation operations yield the weighted error term sequence. And through the proportional element mapping function =1.15 after processing. The actual test verified that under the interference of external room temperature fluctuation of ±7℃, the peak amplitude of the proportional loop response curve driven by the weighted error term was significantly reduced, the temperature overshoot was converged to within 0.35℃, and the linkage fluctuation amplitude of the air pressure control was kept within 0.1kPa, which effectively improved the dynamic stability of the system under multiple disturbance coupling conditions; S4.5: Perform real-time over-limit detection processing on the weighted error term output by S4.4. Based on the preset weighted saturation limit range [0.1, 2.5], compare the upper and lower thresholds. When an over-limit state is detected, trigger the amplitude clipping algorithm to generate the final proportional element dynamic compensation signal that meets the system robustness requirements. For the weighted error term output by S4.4, a real-time over-limit detection algorithm (parameter: preset weight saturation limit range [0.1, 2.5]) is used to realize the function of comparing the upper and lower thresholds of the instantaneous amplitude of the weighted signal in the proportional circuit. Furthermore, a dual threshold comparison method (parameter: upper threshold) is used. =2.5, lower threshold =0.1), realize the sampling comparison between the signal amplitude and the limiting range, and generate an over-limit status flag bit matrix to clarify the over-limit direction and over-limit degree of each sampling point; Furthermore, an amplitude clipping algorithm (parameters: clipping mode is saturation clamping, processing step size is 1 sampling period) is adopted to perform amplitude forced constraint processing on the sampling points marked in the over-limit state flag matrix, setting the sampling values ​​exceeding the upper limit to zero. Set sampled values ​​below the lower limit to ; Furthermore, through an amplitude smoothing transition algorithm (parameter: smoothing coefficient) =0.7, window length of 3 sampling points), to achieve smoothing of the signal in the time domain after clipping, avoiding oscillation of the proportional circuit output caused by instantaneous decrease or increase; Through the above real-time over-limit detection and trimming smoothing process, the weighted error term output by S4.4 is transformed into a final proportional element dynamic compensation signal that meets the weight saturation limit requirements, thereby enhancing the robustness of the proportional element of the temperature control channel and providing over-bias protection. For example, in the temperature control scenario of a high and low temperature low pressure test chamber, the weighted error term output by S4.4 is a sampling sequence of length 100 with a sampling period of 50ms. The maximum value is 3.1 and the minimum value is 0.05. The double threshold comparison formula is then executed:

[0024] in, For the input sequence element values, =2.5, =0.1. After comparison, there are 4 points exceeding the upper limit and 6 points exceeding the lower limit. Apply the saturation clamping formula to the out-of-limit values:

[0025] Elements exceeding the upper limit are directly set to 2.5, and elements exceeding the lower limit are set to 0.1. The pruned sequence is then subjected to exponential weighted smoothing with a window of length 3. =0.7), resulting in a smooth final compensation signal. This signal avoids oscillations in the proportional output of the temperature control channel caused by instantaneous jumps in the error term during actual control, ensuring significantly improved stability of the temperature response curve and over-bias protection performance under room temperature fluctuations of ±7℃.

[0026] Step S5: Based on the dynamic weight matrix output in S3, the integral quantity of the air pressure deviation is weighted to generate a weighted integral term that includes the grid voltage fluctuation compensation factor, so as to achieve dynamic compensation of the integral link of the air pressure control channel. Specifically, it includes: S5.1: Perform Butterworth bandpass filtering on the grid voltage fluctuation obtained in S1 to extract the fluctuation component of the 50Hz power frequency interference band and generate a grid voltage fluctuation characteristic sequence as the compensation input source for the air pressure control channel. S5.2: Based on the grid voltage fluctuation feature sequence generated by S5.1, the multi-source disturbance coupling spectrum constructed by S2 is used for path matching processing to identify the main propagation path gain and phase delay parameters of grid voltage fluctuation to the pressure control channel, and output the grid voltage fluctuation coupling strength index. Based on the grid voltage fluctuation feature sequence generated by S5.1, a path matching algorithm (parameters: multi-source disturbance coupling spectrum G(ω,τ,θ), matching criteria are minimum phase difference and maximum path gain) is used to realize the correlation and identification of the propagation path of the feature sequence and the voltage disturbance node in the coupling spectrum. Furthermore, by using the cross-spectral analysis method (parameters: window length of 4 seconds, overlap rate of 50%, frequency resolution of 0.25Hz), the frequency domain phase synchronization of the power grid voltage fluctuation characteristic sequence and the air pressure control channel response signal is calculated, and the phase delay data of each frequency band is obtained. Furthermore, the path gain estimation algorithm (parameters: based on time-frequency coherence coefficients, the gain calculation formula is as follows):

[0027] in This is a characteristic sequence of grid voltage fluctuations. This is a pressure response signal. For covariance, Let V be the variance of the grid voltage fluctuation signal. (The variance of the pressure response signal) is used to quantify the impact of voltage disturbances on pressure control and obtain the path gain index. ; Furthermore, by using a phase delay correction method (parameter: based on the delay compensation model, the delay parameter τ is weighted and smoothed according to the corresponding frequency), the main path delay parameter is accurately fitted, generating a coupling characteristic vector containing path gain and delay. The comprehensive calculation formula based on coupling strength is as follows:

[0028] in As a coupling strength index, For path gain, The phase delay (in seconds) transforms the result of the previous step into an index characterizing the coupling strength from grid voltage fluctuations to the gas pressure control channel, thus providing accurate input for subsequent weight matrix element calculations. For example, in a high-low temperature and low-pressure test chamber operating at -65℃ and 2kPa ambient pressure, the sampling frequency of the power grid voltage fluctuation characteristic sequence was set to 200Hz. Using a path matching algorithm, the main propagation path numbered P_VP3 was identified in the multi-source disturbance coupling spectrum, with a phase delay of 0.42 seconds in the 1.25Hz band. The time-frequency coherence coefficient output by cross-spectral analysis reached 0.87 in this frequency band, and the gain was calculated using the path gain formula. After delay correction seconds, substituting into the coupling strength formula, yields the result. This coupling strength index is used as input into the matrix elements of S5.3. Dynamic mapping processing significantly improved the stability of the pressure integral control quantity under voltage fluctuations in actual operation, reducing the peak-to-peak pressure fluctuation to within 0.12 kPa. S5.3: Based on the grid voltage fluctuation coupling strength index output by S5.2 and the dynamic weight matrix output by S3, perform time-frequency mapping processing of matrix elements, calculate the instantaneous influence weight coefficient of grid voltage fluctuation on the integral of air pressure deviation, and generate a time-varying sequence of grid voltage fluctuation compensation factor. Based on the grid voltage fluctuation coupling strength index output by S5.2 and the dynamic weight matrix output by S3, a time-frequency mapping analysis method (parameters: path gain, phase delay, disturbance spectral density) is used to realize the matrix elements. Time-frequency domain correlation modeling with grid voltage fluctuation characteristics ensures that the weighting coefficients are dynamically adjusted with the disturbance frequency and phase; Furthermore, through a frequency domain weighted calculation method (parameter: initial weight value) Target frequency band 50Hz±5Hz, phase delay ), to implement matrix elements The instantaneous impact of the weights is corrected, and a time-varying weight coefficient sequence is obtained; Furthermore, a normalized amplitude correction algorithm (parameter: coupling strength index) is used. Disturbance amplitude This process normalizes the magnitude of the weight coefficient sequence and generates a dimensionless compensation factor benchmark sequence. ; Furthermore, by combining linear interpolation and smoothing filtering (parameters: sliding window length L=10, smoothing coefficient α=0.6), dynamic smoothing of the compensation factor benchmark sequence is achieved, avoiding output oscillation of the integral stage caused by sudden weight changes; Through the above time-frequency mapping and smoothing data processing, the coupling strength index of the previous step is transformed into a time-varying sequence of the power grid voltage fluctuation compensation factor, realizing instantaneous dynamic weighting of the integral of the air pressure deviation, and providing high-precision compensation data for the subsequent component weighting fusion of S5.4. For example, in a certain avionics environmental test, the path gain was set to 0.85 and the phase delay was... The parameters are: ms, coupling strength exponent γ = 0.92, sampling frequency = 200Hz, and target frequency band = 50Hz±5Hz. The grid voltage fluctuation characteristic sequence obtained in S5.2 is input into the time-frequency mapping calculation module. A Hanning window with a length of 40 points is used to perform a Fast Fourier Transform to obtain the spectral density. Matrix elements are then calculated. Frequency domain correction value :

[0029] in For the center angular frequency, The phase delay is used. The corrected weighting coefficients are normalized to generate a reference sequence, and smoothed using an exponentially weighted moving average method with a sliding window length of 10 and a smoothing coefficient of 0.6. The final output time-varying sequence of the compensation factor C(t) maintains a stable trend throughout the entire test period. After being fused with the original gas pressure deviation integral, it effectively reduces the accumulation of integral error caused by grid voltage fluctuations and significantly improves the dynamic response stability of the gas pressure control channel. S5.4: Perform component weighted fusion processing on the time-varying sequence of the grid voltage fluctuation compensation factor generated in S5.3 and the original pressure deviation integral quantity, and perform dynamic scaling operation on the pressure deviation integral quantity to output a weighted pressure deviation integral term containing the grid voltage fluctuation compensation factor. Based on the time-varying sequence of the grid voltage fluctuation compensation factor generated by S5.3 and the original gas pressure deviation integral, a component weighted fusion algorithm is adopted (parameter: weighting coefficients are derived from the elements of the dynamic weight matrix). This enables the component-by-component linear combination processing of the power grid voltage fluctuation compensation factor and the integral of the air pressure deviation. Furthermore, through the normalization mapping method (parameter: normalization interval [...]), [1,1]), to unify the dimensions of different physical quantities, ensure the comparability of the values ​​of each component in the weighted operation, and obtain the normalized compensation factor sequence and integral quantity sequence; Furthermore, a vector scaling operation with time-by-time multiplication (parameter: element-by-element correspondence) is used to achieve dynamic scaling of the compensation factor sequence on the integral sequence and generate the fused weighted integral vector. Furthermore, a weighted smoothing filtering algorithm (parameters: window length of 5 sampling points, weight distribution conforming to the triangular window function) is used to achieve a smooth transition of the scaled weighted integral vector in the time domain, eliminate the peak fluctuations in the integral caused by instantaneous changes, and generate a smooth weighted integral term. By using dynamic scaling and smoothing, the time-varying sequence of compensation factors and the original integral amount of air pressure deviation from the previous step are transformed into stable weighted air pressure control integral data that includes grid voltage fluctuation compensation factors, thereby realizing dynamic compensation of the integral link of the air pressure control channel under grid fluctuation interference. For example, in a high-low temperature and low-pressure test, the initial sequence of the pressure deviation integral is set as a sampling vector of length 100, the sampling period is 50ms, and the dynamic weight matrix elements are... The value ranges from 0.85 to 1.15 within this period, corresponding to the time-varying sequence range of the normalized compensation factor. 0.3 to 0.4. The original pressure deviation integral sequence is normalized, and the values ​​are mapped to [...]. The power grid compensation factor sequence is also normalized within the interval [1,1]. The scaling result is calculated using element-wise multiplication, as shown in the following formula:

[0030] in, This is the normalized integral of the pressure deviation. The normalized power grid fluctuation compensation factor. The value is the scaled weighted integral. The calculated sequence is then smoothed by a sliding weighted smoothing process of length 5, with the smoothing weights distributed in a triangular window [0.2, 0.4, 0.8, 0.4, 0.2]. The output is a smoothed weighted integral term. It is verified that under the condition of ±4% grid voltage fluctuation, the weighted integral term significantly reduces the peak fluctuation compared to the uncompensated condition, and the response curve of the air pressure control system tends to be stable. S5.5: Perform integral saturation suppression processing on the weighted pressure deviation integral term output by S5.4, and perform amplitude clipping operation based on the preset integral limit threshold to eliminate control overshoot caused by cumulative error and generate a stable weighted integral control quantity suitable for the pressure control channel.

[0031] Step S6: The dynamic weight matrix output from S3 is used to perform path-weighted processing on the power supply fluctuation micro-components, generating a weighted differential term containing a room temperature change prediction factor, to achieve feedforward differential prediction compensation for power supply disturbances. Specifically, this includes: S6.1: Based on the disturbance amplitude vector obtained in S1, extract the power fluctuation signal and room temperature fluctuation signal to obtain real-time external disturbance data as the basic input for differential processing; S6.2: Perform a first-order differential algorithm on the power fluctuation signal and room temperature fluctuation signal obtained in S6.1 to generate a power fluctuation differential signal and a room temperature change prediction factor, which are used to quantify the characteristics of the disturbance change rate. S6.3: Based on the dynamic weight matrix output by S3, extract the room temperature weight coefficient and power supply weight coefficient to determine the weight distribution of the instantaneous influence of temperature-pressure-power supply disturbance on the PID derivative element. Based on the dynamic weight matrix data structure output by S3, the weight coefficient elements corresponding to room temperature disturbance and the weight coefficient elements corresponding to power supply disturbance in the matrix are taken as extraction targets to establish the calculation basis for the instantaneous weighted distribution of the PID derivative element. A matrix analysis method (parameters: matrix dimension 3×3, element indexing rule (i,j)) is used to realize the correlation elements of room temperature perturbation in the dynamic weight matrix. Elements associated with power supply disturbances The location and separation of the two types of perturbations are determined, and the initial weight coefficient set of the two types of perturbations is obtained. Furthermore, by using the element normalization algorithm (parameter: normalization interval [0,1]), the initial set of weight coefficients is subjected to amplitude scaling processing to realize the comparison and operation of different perturbation weights in a unified numerical domain, and to generate a set of normalized weight coefficients; Furthermore, by using a time-frequency matching method (parameters: perturbation type encoding, perturbation angular frequency), the normalized weights are bound to the corresponding perturbation frequency components, resulting in the room temperature perturbation weight spectral distribution and the power supply perturbation weight spectral distribution, providing a frequency domain reference index for the subsequent weighted combination of differential prediction factors; Furthermore, by using a weight distribution mapping process, the room temperature disturbance weight spectrum distribution and the power supply disturbance weight spectrum distribution are transformed into an instantaneous influence weight vector. ,in Let be the instantaneous weight of the room temperature perturbation on the differential element. This is used to determine the instantaneous weight of power supply disturbances on the derivative element, enabling accurate determination of the weight distribution of the instantaneous effects of three types of disturbances—temperature, air pressure, and power supply—on the PID derivative element. By using matrix analysis and weight mapping, the dynamic weight matrix result from the previous step is transformed into paired data of room temperature weight coefficient and power supply weight coefficient. The output can be directly used for weighted summation processing in S6.4, realizing the coupled weighting of power supply disturbance differential term and room temperature change prediction factor. For example, in the dynamic weight matrix of a high and low temperature low pressure test chamber, at a period of t=12.5s, the matrix elements... =1.85、 =0.94. Both are scaled to the interval [0,1] using a normalization algorithm to obtain the normalized room temperature weight. =0.74, Normalized power source weight =0.38. Combined with perturbation type encoding. =1、 =3 and the corresponding perturbation angular frequency =0.25Hz =0.4Hz, and the weight values ​​are bound to their respective frequency domains using a time-frequency matching method, forming a weighted spectral distribution {room temperature: 0.74@0.25Hz, power supply: 0.38@0.4Hz}. The instantaneous influence weight vector generated by the mapping is... The weighted combination circuit input to S6.4 achieves synchronous weighting of the room temperature mutation factor and the power supply differential signal in actual operation, ensuring that the differential circuit can maintain the stability of the control signal and the anti-interference capability when facing the simultaneous occurrence of rapid room temperature rise and power supply fluctuation. S6.4: Based on the power fluctuation differential signal and room temperature change prediction factor generated in S6.2, and the weight coefficients extracted in S6.3, perform weighted summation to generate a weighted differential combination term containing the room temperature change prediction factor; Based on the power fluctuation differential signal and room temperature change prediction factor generated by S6.2, and the weight coefficients extracted by S6.3, the component weighted superposition algorithm (parameters: weight coefficient vector, signal component) is adopted to realize the quantification of the joint effect of multiple disturbance factors in the differential stage. Furthermore, by performing vector multiplication, the differential power fluctuation signal is multiplied by the power weighting coefficient to obtain the weighted differential power fluctuation components, and a sequence of component values ​​is obtained. Furthermore, by using vector multiplication, the room temperature mutation prediction factor and the room temperature weight coefficient are multiplied to obtain the room temperature prediction component after weight adjustment, and a sequence of component values ​​is obtained. Furthermore, an arithmetic summation method (parameters: weighted power fluctuation differential component and room temperature prediction component) is adopted to achieve periodic summation and fusion of the two components, and generate a time series of weighted differential combination terms to ensure that the term simultaneously reflects the power disturbance change rate and the room temperature change early warning effect. Furthermore, through dynamic normalization processing (parameters: time series of weighted differential combination terms, normalization interval [-1,1]), the numerical range of the combination terms is limited to the preset interval to generate normalized weighted differential combination terms, so as to avoid the distortion of the overall control signal caused by the abnormal amplitude of a single disturbance term. By weighted superposition and normalization of components, the results of the previous step are transformed into a weighted differential combination term containing the room temperature mutation prediction factor, so as to realize the fine response control of the feedforward differential link to multivariable coupled perturbations. For example, during the operation of the high and low temperature low pressure test chamber, the external room temperature fluctuation is differentiated to obtain the room temperature change prediction factor, with a value of [value missing]. The differential signal value of the power supply fluctuation is The room temperature weighting coefficient extracted from the dynamic weighting matrix is: The power supply weighting coefficient is Component-weighted calculation is used: Room temperature component = = Power component = = Summing two components: = This yields the original value of the weighted differential combination term. After normalization (interval [-1, 1]), the value of this combination term is... As the input to the PID derivative element, it enables the derivative control to respond sensitively to sudden changes in room temperature while maintaining stable compensation against power supply disturbances. In this embodiment, the controller achieves significant convergence of error fluctuations under complex disturbances, and the smoothness of the output signal is significantly improved, verifying the effectiveness and robustness of the weighted combination term construction. S6.5: The weighted derivative combination term generated in S6.4 is output as a weighted derivative term to the derivative link of the PID controller to realize feedforward derivative prediction compensation for power supply disturbance, thereby reducing control error and improving the system's anti-interference capability.

[0032] Step S7: Determine whether the weighted control terms generated in S4, S5, and S6 exceed the preset weight saturation limit. If they do, perform weight limiting processing to ensure the system robustness of the multivariable coupled feedforward compensation process. Specifically, this includes: S7.1: Perform synchronous reading operations on the weighted error term of the temperature control channel generated in S4, the weighted integral term of the air pressure control channel generated in S5, and the weighted differential term of the power disturbance generated in S6 to obtain the set of weighted control terms, which will serve as the input data source for subsequent over-limit judgment. For the weighted error term of the temperature control channel generated by S4, the weighted integral term of the air pressure control channel generated by S5, and the weighted differential term of the power disturbance generated by S6, a multi-channel data synchronous reading method is adopted (parameters: data buffer pointer position, sampling period 50ms, trigger source is real-time operating system clock) to realize the function of simultaneous acquisition of weighted signals of three different control channels. Furthermore, a high-precision timestamp marking algorithm (parameter: clock source is IEEE1588 Precision Time Protocol PTP hardware clock) is used to perform nanosecond-level timestamp appending processing on the acquired signal, and weighted signal recording data with strict time alignment characteristics is obtained. Furthermore, through a multi-channel data buffer queue management method (parameters: the length of the circular buffer queue is 1024 frames, and the write triggering strategy is double buffer switching), the collected data is stored in memory in an orderly manner, and a temporary storage structure of weighted control items classified by channel is generated. Furthermore, a signal type and source identification coding algorithm (parameters: temperature channel identifier Code_T=01, air pressure channel identifier Code_P=02, power channel identifier Code_V=03) is adopted to realize the signal category identification of each weighted control item and generate a set of weighted control items containing channel category, sampling timestamp and signal value; By using the above-mentioned synchronous reading and time alignment processing method, the multi-channel weighted control signal of the previous step is transformed into a set of weighted control items with a unified data structure and accurate time reference, thereby achieving the expected technical effect of providing a consistent input data source for subsequent over-limit judgment. For example, during this synchronous reading operation, the sampled value of the weighted error term for the temperature control channel is 0.85, the sampled value of the weighted integral term for the air pressure control channel is 1.92, and the sampled value of the weighted differential term for the power supply disturbance is 0.15. The sampling period is set to 50ms, and an IEEE 1588 hardware clock is used to add a timestamp accurate to nanoseconds at each sampling instant. The identification code for the temperature signal is 01, the identification code for the air pressure signal is 02, and the identification code for the power supply signal is 03. The data is stored sequentially in a circular buffer queue, forming a circular buffer capacity of 1024 frames. Under the above parameter conditions, the three weighted signals read can be compared under the same time base, ensuring that there is no time misalignment or channel confusion problem in the subsequent over-limit judgment calculation within the limiting range [0.1, 2.5], thus improving the accuracy of the limiting processing and the robustness of the system. S7.2: Based on the preset weighted saturation limit range [0.1, 2.5], perform upper and lower limit threshold comparison processing on each element in the weighted control item set to generate an over-limit status indication signal, which clearly identifies the position and over-limit direction of the element that exceeds the limit range; S7.3: Based on the over-limit status indication signal, perform an over-limit element filtering operation on the weighted control item set to identify the over-limit weighted control item elements that need to be limited, ensuring that subsequent processing is only for valid over-limit objects; S7.4: For the identified over-limit weighted control item elements, a numerical clamping algorithm combining the minimum and maximum value functions is applied to perform amplitude limiting processing, so as to force the element values ​​to be constrained within the weight saturation amplitude limit range, and generate the amplitude-limited element values. S7.5: Integrate the element values ​​after amplitude limiting with the element values ​​that have not exceeded the limit in the weighted control term set to output the amplitude-limited weighted control term set, ensuring the system robustness of the multivariable coupled feedforward compensation process.

[0033] Step S8: Based on the weighted control terms processed in S7, a sliding window online update mechanism is implemented, calculating the dynamic weight matrix every five seconds and implementing a smooth transition to achieve precise suppression of multivariate coupled disturbances by frequency band, path, and time sequence. Specifically, this includes: S8.1: Based on the multivariable coupling system characteristics and real-time control requirements of the high and low temperature low pressure test chamber, a sliding window online update mechanism is constructed. This mechanism sets the window length to 30 seconds and the rolling step to 5 seconds. It performs time-aligned storage processing on the historical disturbance amplitude vector and weighted control response sequence to generate a structured sliding window dataset, providing a reliable data source for the rolling calculation of the dynamic weight matrix. S8.2: Based on the structured sliding window dataset generated by S8.1, extract the room temperature fluctuation, grid voltage fluctuation and power supply fluctuation within the latest 5-second period from the weighted control terms processed by S7, and perform timestamp alignment and normalization preprocessing on the disturbance amplitude vector to form the disturbance input dataset of the current rolling window, which serves as the real-time input basis for the dynamic weight matrix update. S8.3: Apply the least squares optimization algorithm to the current rolling window disturbance input dataset formed in S8.2 to calculate the gain coefficient and phase delay parameter of the disturbance propagation path. Perform matrix solving based on the multi-source disturbance coupling spectrum to generate an updated dynamic weight matrix to compensate for the frequency band coupling effect of the three types of disturbances: temperature, air pressure, and power supply. S8.4: Based on the updated dynamic weight matrix generated in S8.3 and the weight matrix stored in the previous period, perform exponential weighted moving average smoothing, calculate the transition increment of the weight coefficients and perform amplitude limiting correction to generate a smooth transition dynamic weight matrix and avoid control oscillations caused by sudden weight changes. S8.5: The smooth transition dynamic weight matrix output from S8.4 is embedded into the control law of the PID feedback loop. The error term of the proportional element, the deviation of the integral element, and the rate of change of the derivative element are replaced by component weighting to achieve precise suppression of multivariable coupled disturbances by path and time sequence, thereby improving the control accuracy of the system.

[0034] The present invention also provides a high and low temperature and low pressure test chamber, which uses the above-mentioned constant temperature and pressure control method to control the temperature and pressure inside the test chamber.

[0035] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0036] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and rules of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for controlling constant temperature and pressure, characterized in that, Includes the following steps: S1: Multi-dimensional synchronous acquisition of external environmental disturbance sources of the high and low temperature low pressure test chamber, obtaining disturbance amplitude vectors composed of room temperature fluctuation, power grid voltage fluctuation and power supply fluctuation, forming a real-time monitoring dataset. S2: Based on the real-time monitoring dataset, using the improved Granger causality test combined with time-frequency coherence analysis, the main propagation paths of three types of disturbances—temperature, air pressure, and power supply—in the thermodynamic-fluid dynamic-electrodynamic coupled system of the test chamber are identified, and a multi-source disturbance coupling spectrum is generated. S3: Based on the multi-source disturbance coupling spectrum and the disturbance amplitude vector, calculate the dynamic correlation between the disturbance type code, the disturbance angular frequency and the equivalent propagation delay, and output the dynamic weight matrix; S4: Embed the dynamic weight matrix into the PID feedback loop to perform weighted processing on the temperature deviation signal and generate a weighted error term that includes multivariate coupling effects; S5: Based on the dynamic weight matrix, perform component weighting on the integral of the air pressure deviation to generate a weighted integral term that includes the power grid voltage fluctuation compensation factor; S6: Use the dynamic weight matrix to perform path weighting on the power supply fluctuation micro-components to generate a weighted differential term that includes a room temperature change prediction factor; S7: Determine whether the weighted error term, the weighted integral term, and the weighted derivative term exceed the preset weight saturation limit range. If they do, perform weight limiting processing.

2. The constant temperature and pressure control method according to claim 1, characterized in that, The process following step S7 also includes: S8: Based on the weighted control terms processed in step S7, a sliding window online update mechanism is implemented, and the dynamic weight matrix is ​​calculated every 5 seconds and a smooth transition is implemented to achieve precise suppression of multivariate coupled disturbances by frequency band, path, and time sequence.

3. The constant temperature and constant pressure control method according to claim 1, characterized in that, The multi-dimensional synchronous acquisition includes deploying a high-precision sensor array for room temperature disturbance sources, power grid voltage disturbance sources, and power supply disturbance sources in the external environment of the high-low temperature and low-pressure test chamber. The sensor array includes thermocouple temperature sensors, Hall voltage sensors, and current transformers. The original simulated room temperature signal is obtained based on the thermocouple temperature sensors, the original simulated power grid voltage signal is obtained based on the Hall voltage sensors, and the original simulated power supply fluctuation signal is obtained based on the current transformers.

4. The constant temperature and pressure control method according to claim 1, characterized in that, Step S3 specifically includes: Based on the path gain, group delay, and nonlinear distortion data stored in the multi-source perturbation coupling map generated in step S2, a map mapping function is constructed. The disturbance amplitude vector obtained in step S1 and the multi-source disturbance coupling spectrum are input into the spectrum mapping function to form the input parameter set of the mapping function; Based on the input parameter set, the dynamic correlation between the disturbance type code and the disturbance angular frequency is calculated using the spectral mapping function. The disturbance propagation strength coefficient is generated by multiplying the path gain and the group delay. Based on the disturbance propagation intensity coefficient, combined with the equivalent propagation delay and nonlinear distortion parameters, the mapping relationship between the equivalent propagation delay and the weight coefficient is calculated, and a time-varying weight coefficient vector is generated. The time-varying weight coefficient vector is organized into a 3×3 structure to form a dynamic weight matrix.

5. The constant temperature and constant pressure control method according to claim 4, characterized in that, The spectral mapping function uses perturbation type encoding, perturbation angular frequency, and equivalent propagation delay as configuration parameters, and perturbation amplitude vector as input variables.

6. The constant temperature and pressure control method according to claim 1, characterized in that, Step S4 specifically includes: The temperature deviation signal of the high and low temperature low pressure test chamber is processed by Butterworth bandpass filtering. The passband range is set based on the temperature disturbance main path parameters in the dynamic weight matrix generated in step S3, and a standard error sequence is generated. Based on the standard error sequence and the dynamic weight matrix, the instantaneous product of the first row and first column element of the dynamic weight matrix and the standard error sequence is calculated to generate the dynamic compensation vector of the main path of temperature disturbance. The scalar values ​​of room temperature fluctuation and grid voltage fluctuation obtained in step S1 are subjected to sliding window differentiation processing. Based on the dynamic compensation vector and the elements of the first row and second column of the dynamic weight matrix and the elements of the first row and third column of the dynamic weight matrix, path weighted summation is performed to generate a disturbance compensation superposition term. The standard error sequence and the disturbance compensation superposition term are subjected to vector linear combination processing. A weighted coefficient set is constructed based on the first row of the dynamic weight matrix. The combination result is subjected to proportional element mapping transformation, and the weighted error term is output to the temperature control channel. Real-time over-limit detection processing is performed on the weighted error term. The upper and lower thresholds are compared based on the preset weighted saturation limit range. When an over-limit state is detected, the amplitude clipping algorithm is triggered to generate the final proportional dynamic compensation signal.

7. The constant temperature and pressure control method according to claim 6, characterized in that, The elements in the first row and first column of the dynamic weight matrix correspond to the temperature disturbance weight coefficient, the elements in the first row and second column of the dynamic weight matrix correspond to the room temperature disturbance weight coefficient, and the elements in the first row and third column of the dynamic weight matrix correspond to the grid voltage disturbance weight coefficient.

8. The constant temperature and constant pressure control method according to claim 1, characterized in that, Step S5 specifically includes: Perform Butterworth bandpass filtering on the grid voltage fluctuation obtained in step S1 to extract the fluctuation component of the power frequency interference band and generate a grid voltage fluctuation feature sequence. Based on the power grid voltage fluctuation characteristic sequence, the multi-source disturbance coupling spectrum constructed in step S2 is used for path matching processing to identify the main propagation path gain and phase delay parameters of the power grid voltage fluctuation to the pressure control channel, and the power grid voltage fluctuation coupling strength index is output. Based on the grid voltage fluctuation coupling strength index and the dynamic weight matrix output in step S3, perform time-frequency mapping processing of matrix elements, calculate the instantaneous influence weight coefficient of grid voltage fluctuation on the integral of air pressure deviation, and generate a time-varying sequence of grid voltage fluctuation compensation factor. The time-varying sequence of the power grid voltage fluctuation compensation factor is combined with the original gas pressure deviation integral quantity by performing component weighted fusion processing, and the gas pressure deviation integral quantity is dynamically scaled to output the weighted gas pressure deviation integral term. The integral saturation suppression process is performed on the weighted pressure deviation integral term, and the amplitude is clipped based on the preset integral limit threshold to generate a stable weighted integral control quantity.

9. The constant temperature and pressure control method according to claim 2, characterized in that, The sliding window online update mechanism is set with a window length of 30 seconds and a scrolling step of 5 seconds. The smooth transition is to recalculate the weight parameters using least squares every 5 seconds based on the monitoring data and perform exponential weighted moving average smoothing with the weight matrix of the previous period.

10. A high and low temperature low pressure test chamber, characterized in that: The constant temperature and pressure control method described in any one of claims 1-9 is used to control the internal temperature and pressure of the test chamber.

Images