Aircraft environmental control system heat and humidity load prediction and regulation method, device and equipment and storage medium
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-15
- Publication Date
- 2026-03-31
AI Technical Summary
Traditional airborne environmental control systems cannot dynamically adapt to the drastic changes in environmental parameters and equipment computing load during flight, resulting in insufficient heat dissipation or energy waste. Existing AI-based prediction models have limited accuracy and generalization ability, and have failed to form an intelligent closed loop.
By employing high-quality data preprocessing and a deeply optimized BP neural network model, combined with the whale optimization algorithm, an accurate heat and humidity load prediction model is constructed. Through real-time data feedback, proactive decision-making and control are achieved, forming a closed-loop optimization.
It achieves accurate prediction and dynamic control of heat and humidity load, improves the accuracy of temperature and humidity control and equipment reliability, reduces the energy consumption of the environmental control system, and has strong adaptability and generalization ability.
Smart Images
Figure CN121764027A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of aircraft airborne system design, and in particular to a method, apparatus, equipment and storage medium for predicting and controlling the thermal and humidity load of an aircraft environmental control system. Background Technology
[0002] In recent years, eVTOL (electric vertical takeoff and landing) and drone technologies have been widely applied in numerous fields. With the rapid improvement in the functionality and integration of airborne electronic equipment, to achieve functions such as high-precision flight control, multi-device collaborative operation, and complex environment perception, a large number of electronic components are integrated within the aircraft, including flight control systems, navigation and positioning modules, multispectral sensors, high-definition image processing units, and wireless communication equipment. Some high-end models also incorporate AI computing chips to support real-time data processing and intelligent decision-making. These electronic devices continuously generate heat during operation. With increased component density and computational load, their heat generation per unit volume is 3-5 times higher than traditional models. If this heat cannot be dissipated in time, the equipment temperature will exceed the safe threshold, leading to decreased computational efficiency, data transmission interruptions, and even hardware burnout, placing extremely high demands on the environmental control system.
[0003] Traditional airborne environmental control systems often employ passive cooling or simple active cooling strategies based on fixed empirical values. These strategies cannot dynamically adapt to the complex changes in thermal and humidity loads caused by drastic changes in environmental parameters and fluctuations in equipment computing load during flight. For example, when an aircraft switches from low-altitude, low-speed flight to high-altitude, high-speed flight, changes in ambient temperature and air pressure affect cooling efficiency. At the same time, fluctuations in the computing load of electronic equipment can also lead to sudden changes in heat generation. Traditional systems, unable to adjust their cooling strategies in real time, are prone to insufficient cooling or energy waste.
[0004] Although there are some artificial intelligence-based prediction attempts in existing technologies, they often have the following shortcomings: 1) The data preprocessing process is simple and ineffective in handling noise and missing data in aircraft operation data; 2) The prediction model structure is fixed, and its generalization ability and accuracy are limited; 3) The prediction and control links are disconnected, and an intelligent closed loop of "prediction-decision-control-feedback" has not been formed.
[0005] Therefore, there is a need in this field for an integrated solution that can achieve accurate prediction, intelligent decision-making, and adaptive regulation. Summary of the Invention
[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method, device, equipment and storage medium for predicting and controlling the thermal and humidity load of an aircraft environmental control system. This invention can achieve accurate and rapid prediction of the thermal and humidity load of the environmental control system and perform forward-looking and energy-saving optimal control of the environmental control system based on the prediction results.
[0007] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for predicting and controlling the thermal and humidity load of an aircraft environmental control system, comprising the following steps: Collect historical operating data of the environmental control system, identify, repair, and standardize the data for missing and outlier values, and construct an offline database; Based on the offline database, a preset number of parameters are selected from the historical operation data through correlation analysis as input to construct a BP neural network prediction model; by traversing the number of intermediate layer nodes of the BP neural network, the network structure is determined with the minimization of prediction error as the evaluation index, and an optimization algorithm is used to optimize the weights and thresholds of the BP neural network to obtain the optimized prediction model. Real-time status data of the environmental control system is collected and input into the optimized prediction model to obtain the predicted value of thermal and humidity load. Based on the predicted value of thermal and humidity load, the operating parameters of the environmental control system are determined through active decision-making with the goal of optimizing system energy consumption, and the operating parameters are sent to the environmental control system actuator of the aircraft to regulate the environmental control system.
[0008] In a second aspect, the present invention provides a device for predicting and controlling the thermal and humidity load of an aircraft environmental control system, used to implement the method described in the first aspect, comprising: The data acquisition and preprocessing module is configured to acquire historical operating data of the environmental control system, and to identify and repair missing and outlier values and standardize the data to build an offline database. The model training and optimization module is configured to, based on the offline database, select a preset number of parameters from the historical running data through correlation analysis as input to construct a BP neural network prediction model, determine the network structure by traversing the number of intermediate layer nodes of the BP neural network and using the minimization of prediction error as the evaluation index, and optimize the weights and thresholds of the BP neural network using an optimization algorithm to obtain the optimized prediction model. The online prediction and decision control module is configured to collect real-time status data of the environmental control system and input it into the optimized prediction model to obtain the predicted value of heat and humidity load, and determine the operating parameters of the environmental control system based on the predicted value of heat and humidity load through active decision-making with the goal of optimizing system energy consumption. The environmental control system actuator is configured to receive the operating parameters and perform corresponding control actions.
[0009] Thirdly, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in the first aspect.
[0010] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.
[0011] The beneficial effects of this invention include: This invention achieves accurate prediction of thermal and humidity loads under complex operating conditions (10-30 seconds in advance) through high-quality data preprocessing and a deeply optimized BP neural network model, significantly reducing prediction errors. The prediction model is optimized through algorithms and continuously updated via a feedback mechanism, exhibiting excellent generalization ability and strong dynamic adaptability to different aircraft types and flight profiles. Through prediction-based multi-objective proactive decision-making, "on-demand cooling" is achieved, effectively reducing the energy consumption of the environmental control system while improving the accuracy of temperature and humidity control and equipment reliability. Attached Figure Description
[0012] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0013] Figure 1 This is a schematic diagram of the overall process of the aircraft environmental control system thermal and humidity load prediction and control method according to an embodiment of the present invention; Figure 2 This is a structural block diagram of the thermal and humidity load prediction and control device for the aircraft environmental control system according to an embodiment of the present invention. Detailed Implementation
[0014] The embodiments of this application will now be described in detail with reference to the accompanying drawings.
[0015] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0016] The following will combine Figure 1This invention provides a detailed description of a method for predicting and controlling the thermal and humidity load of an aircraft environmental control system. The method includes the following steps: S1: Data Acquisition and Preprocessing This step includes collecting historical operating data of the environmental control system, identifying, repairing, and standardizing missing and outlier values in the data, and constructing an offline database.
[0017] Specifically, such as Figure 1 As shown on the left, it includes: Data collection and recording: This step forms the data foundation for building the predictive model. The process begins with the environmental control system itself, continuously collecting and recording multi-dimensional data such as the power of electronic equipment, battery status, flight profile (e.g., aircraft altitude and speed), external environmental parameters (e.g., temperature, humidity, wind speed, and air pressure), and environmental control target parameters (e.g., set target temperature and humidity) through airborne sensors.
[0018] Outlier and Missing Value Identification: Based on historical operational data of the aircraft's environmental control system collected at multiple times, the data is first analyzed to identify outliers and missing values. In an exemplary embodiment, for missing values, a traversal method is used for identification; the program sequentially checks each data point in the dataset to determine if it is missing. For outliers, the Tukey fence method is used for identification; the program marks values outside the normal range as outliers based on the calculation of the quartiles and interquartile ranges of the data. Specifically, first, the 25th percentile of the data to be processed is calculated (…). ) and 75th percentile ( ); then, the interquartile range ( Its value is and The difference, that is = - Based on this, calculate the upper limit of the reasonable range of normal data ( ) and lower limit ( The calculation formula is as follows:
[0019]
[0020] Ultimately, the values in the dataset that are higher than or below The data points are marked as outliers.
[0021] Outlier and Missing Value Handling: After identifying missing and outlier values, the identified outliers and missing values are repaired. This embodiment of the invention uses a K-Means clustering algorithm based on Euclidean distance to repair outliers and missing values. The specific process is as follows: For each data point to be repaired, based on its other dimensions of data in the database, the Euclidean distance between it and all other data points is calculated to measure the similarity between the data. The formula for calculating the Euclidean distance is as follows:
[0022] in, For vectors and Euclidean distance, and They are vectors and The i-th component, k is the total dimension of the vector.
[0023] Subsequently, the five valid data points that are closest to the Euclidean distance of the data point to be repaired are selected, and the weighted average of the values of these five data points in the missing or abnormal dimension is taken to complete the repair.
[0024] Discrete-level averaging: To reduce data fluctuations and extract steady-state features, the repaired data is discretized and aggregated according to time windows. Specifically, based on the total duration of the flight mission, continuous time series data is divided into equal-length discrete time intervals. The length of these intervals can be set to 5 minutes, 10 minutes, or 30 minutes depending on data processing requirements. Within each time interval, the arithmetic mean of all valid data points falling within that interval is calculated, and this mean is used as the representative data output for that time interval, ultimately stored in the database. This process effectively smooths out short-term noise, preserves the macroscopic trend of load changes, and provides a high-quality data foundation for model training.
[0025] Database: Finally, the standardized data, after missing and outlier repair and discretized averaging according to time windows, is stored in a structured storage system, thus completing the construction of the offline database. This offline database serves as the data foundation for the method described in this embodiment of the invention, providing rich, clean, and time-ordered historical data samples for the subsequent training and optimization of the prediction model.
[0026] S2: Model Training and Optimization This step includes: based on the offline database, selecting a preset number of parameters from the historical operating data as input through correlation analysis to construct a BP neural network prediction model; determining the network structure by traversing the number of intermediate layer nodes of the BP neural network (backpropagation neural network) with the minimization of prediction error as the evaluation index; and using an optimization algorithm to optimize the weights and thresholds of the BP neural network to obtain the optimized prediction model.
[0027] For details, see attached. Figure 1 As shown in the middle section, the prediction model establishment process includes three main stages: input selection and determination, prediction model construction, and optimization of structural parameters and hyperparameters. It should be noted that the model construction and optimization work at this stage is based on historical operational data stored in the offline database.
[0028] Input Filtering and Determination: The system extracts historical operational data from an offline database. Based on this data, it uses correlation analysis to filter a predetermined number of parameters from multiple dimensions, including external meteorological parameters, historical heat and humidity loads, and environmental control target parameters, as model inputs. For example, the predetermined number of parameters are those significantly correlated with heat and humidity load prediction (such as flight altitude, flight speed, ambient temperature, and ambient pressure). In practice, the Pearson correlation coefficient method is used to calculate the correlation between each parameter and the heat and humidity load. The formula for calculating the Pearson correlation coefficient is:
[0029] in, The Pearson correlation coefficient between variables x and y. and Let x and y be the i-th sample points of variables x and y, respectively. and ...
[0030] Then, parameters with an absolute correlation coefficient greater than 0.4 were selected as input features. The selected parameters included at least one of environmental parameters, equipment power parameters, historical heat and humidity load parameters, and environmental control target parameters. This selection process ensured that there was a clear physical correlation and statistical significance between the model input parameters and the prediction target.
[0031] Prediction Model Construction: Using the previously selected parameters as the input layer and the heat and humidity load data as the output layer, a basic framework for a BP neural network prediction model is constructed and trained. The network adopts a forward propagation structure, and performs nonlinear transformations on the input features through hidden layers to learn the complex mapping relationship from input parameters to output heat and humidity load.
[0032] Structural and Hyperparameter Optimization: After determining the basic structure of the BP neural network, the performance of the prediction model is further improved through a two-stage optimization process. This optimization process is also based on training and validation sets divided from historical operating data.
[0033] Structural parameter traversal: By systematically traversing different numbers of intermediate layer nodes in the BP neural network, the optimal number of nodes is found to evaluate the model performance under each structure. The network structure is determined using the minimization of prediction error as the evaluation metric. Specifically, the mean absolute deviation (MAD) is used as the evaluation metric. This metric assesses prediction accuracy by calculating the mean absolute error between the predicted and actual values. The formula for MAD is:
[0034] in, Let j be the actual heat and humidity load value of the j-th sample. Let m be the predicted value of the model for the j-th sample, and m be the number of sample data in the validation set.
[0035] The optimal network structure is the number of intermediate layer nodes corresponding to the minimum MAD value.
[0036] Algorithm optimization: Based on the determined optimal network structure, the Whale Optimization Algorithm (WOA) is used to globally optimize the weights and thresholds of the BP neural network. This algorithm simulates the intelligent foraging behavior of a whale pod, efficiently finding the optimal combination of weights and thresholds in the solution space through the synergistic effect of three stages: surrounding prey, bubble-web attack, and random search. The aim is to achieve the best prediction accuracy and convergence stability for the predictive model. The resulting optimized predictive model exhibits the highest prediction accuracy.
[0037] S3: Online Predictive and Decision Control This step includes collecting real-time status data of the environmental control system, inputting it into the optimized prediction model to obtain the predicted value of thermal and humidity load; based on the predicted value of thermal and humidity load, determining the operating parameters of the environmental control system through active decision-making with the goal of optimizing system energy consumption, and sending the operating parameters to the environmental control system actuator of the aircraft to regulate the environmental control system.
[0038] Specifically, such as Figure 1 As shown on the right, this stage is crucial for the optimized predictive model to be put into practical operation and realize its value. Its core is to leverage the results of previous steps to achieve a transition from "passive response" to "active regulation." This includes: Heat and humidity load prediction: In this stage, the system continuously collects real-time status data of the environmental control system through a real-time data collection module. After undergoing the same preprocessing process as the training data, this real-time data is input into the BP neural network prediction model trained and optimized in step S2. Based on the input real-time parameters, the model quickly calculates the predicted heat and humidity load values (heat load prediction value, humidity load prediction value) for the next control cycle (e.g., 10-30 seconds). This step achieves forward-looking perception of the system's future state.
[0039] Environmental control strategy setting and active decision-making algorithm: Based on the predicted heat and humidity load, the system combines a preset environmental control strategy and performs calculations through active decision-making with the goal of optimizing system energy consumption to determine the optimal operating parameters of the environmental control system. The active decision-making process simultaneously considers control accuracy and equipment lifespan, weighing multiple objectives, and ultimately determines at least one control parameter, such as fan speed, pipeline flow rate, and equipment start / stop status, with system energy consumption optimization as the ultimate goal.
[0040] Determine the optimal parameters of the system: The active decision-making algorithm outputs the set of operating parameters of the environmental control system.
[0041] Actuator Actions: Ultimately, these operating parameters are converted into specific control commands and sent to the aircraft's environmental control system actuators (such as brushless fans, electric pumps, and thermoelectric coolers) to perform controls including equipment start-stop control, fan speed adjustment, and system coordination parameters, thereby achieving precise and forward-looking control of the environmental control system.
[0042] S4. Closed-loop optimization The steps described in this embodiment of the invention may further include a closed-loop optimization step to achieve self-evolution. After completing online prediction and active control, the actual operating data of the environmental control system is used as new historical data and fed back into the offline database. Specifically, a sliding window method (e.g., retaining the most recent 5000 sets of data) is used to update the database. When the prediction error continuously exceeds a threshold or reaches a set time period (e.g., 24 hours), it serves as a trigger condition to re-execute the model training and optimization steps, thereby iteratively optimizing the prediction model to continuously adapt to changes in the environment and equipment status.
[0043] This invention achieves high-precision, feedforward prediction and control of the thermal and humidity load of an aircraft's environmental control system through high-quality data preprocessing, input parameter screening based on correlation analysis, and a BP neural network deeply optimized by the whale optimization algorithm. This not only significantly improves the stability and accuracy of temperature and humidity control and effectively ensures the reliable operation of airborne electronic equipment, but also achieves "on-demand heat dissipation" through prediction-based proactive decision-making, greatly reducing system energy consumption. At the same time, its closed-loop optimization mechanism ensures the method's strong adaptability and generalization ability to different flight conditions.
[0044] Those skilled in the art will understand that the above method can be implemented in various forms. For example, the present invention can be embodied as a virtual device or system, in which each module corresponds one-to-one with each step of the above method, and is configured by computer program code, performing corresponding functions when running on a processor.
[0045] Specifically, embodiments of the present invention also provide a device for predicting and controlling the thermal and humidity load of an aircraft environmental control system. For example... Figure 2 As shown, the device includes: Data Acquisition and Preprocessing Module 201: This module includes various sensors on the aircraft (such as temperature, humidity, pressure, and power sensors) and their signal conditioning circuits, as well as one or more processors. The processors are configured to execute instructions stored in memory to read sensor data, acquire historical operating data from the environmental control system 207, and perform preprocessing procedures such as missing value identification, outlier identification (Tukey fence method), data repair (K-Means and Euclidean distance weighting), and standardization (Z-score). Finally, an offline database 306 is constructed and maintained. This offline database can be stored in onboard non-volatile memory (such as a solid-state drive).
[0046] Model Training and Optimization Module 202: This module can be a software functional unit in an airborne computer or run on a ground server during the design phase. It includes a processor and a memory, the memory storing program instructions. When executed by the processor, these instructions perform a series of operations, including selecting input parameters from the offline database 206, constructing a BP neural network, traversing the number of intermediate layer nodes, and running the Whale Optimization Algorithm (WOA) to optimize the network weights and thresholds, ultimately generating an optimized prediction model file. Specifically, this module is configured to, based on the offline database, select a preset number of parameters from the historical operating data as input through correlation analysis to construct a BP neural network prediction model; determine the network structure by traversing the number of intermediate layer nodes of the BP neural network, using the minimization of prediction error as an evaluation metric; and optimize the weights and thresholds of the BP neural network using an optimization algorithm to obtain the optimized prediction model.
[0047] Online Prediction and Decision Control Module 203: This module is the core control unit of the system, typically implemented by an onboard real-time processor. It loads the optimized prediction model file generated by the model training and optimization module. During flight, this module continuously receives real-time data streams from the data acquisition and preprocessing module, inputs them into the prediction model for forward calculation, and obtains the predicted thermal and humidity load values. Subsequently, the decision logic embedded in this module (such as querying the control strategy table or running a lightweight optimization algorithm) calculates the optimal operating parameters (such as fan speed and equipment start / stop commands) based on the predicted thermal and humidity values and with proactive decision-making aimed at optimizing system energy consumption.
[0048] Environmental control system actuator 204: This refers to the environmental control system actuator of the aircraft, configured to receive operating parameters and execute corresponding control actions. The actuator is the physical terminal of the system, including but not limited to a brushless DC fan and its drive circuit, an electronic expansion valve, a pump, and relays (used to control equipment start-up and shutdown). They receive electrical signals or digital commands from the online predictive and decision control module and execute corresponding physical actions to change the operating state of the environmental control system.
[0049] Feedforward compensation module 205: Configured to adjust the environmental control system actuator 204 in advance based on the flight profile signal sent by the aircraft's flight management system and / or the environmental sensor signal of the environmental control system. Specifically, this module can be integrated into the online predictive and decision control module as an independent control algorithm thread. It monitors the flight profile data (such as predetermined climb and descent curves) and environmental sensor signals sent by the flight management system in real time. When a known, impending disturbance is detected (such as an approaching high-temperature zone), this module generates an additional control command to adjust the operating point of the actuator in advance to compensate for dynamic disturbances that the predictive model may not have fully captured.
[0050] In this embodiment of the invention, the online prediction and decision control module 203 and the environmental control system actuator 204 constitute a closed-loop control system. The actual operating data of the environmental control system 207 is fed back to the data acquisition and preprocessing module 201 for updating the offline database 206 and re-optimizing the prediction model. Specifically, after the environmental control system actuator 204 is activated, the actual state of the environmental control system (such as the final cabin temperature and humidity, and actual energy consumption) is collected again by sensors in the data acquisition and preprocessing module 201, forming new data. This new data is fed back to the system via the data bus. This new data is used to update the offline database and can trigger the model training and optimization module 202 to periodically or event-triggeredly retrain and optimize the prediction model, thereby enabling the system to have adaptive and learning evolution capabilities.
[0051] This invention also provides an electronic device, which includes at least one processor (CPU), a memory, an input / output interface (I / O), and a communication interface. The memory includes a non-volatile storage medium (such as a hard disk) and RAM. The non-volatile storage medium stores an operating system, a computer program, and the offline database. When the computer program is loaded into memory by the processor and executed, the aforementioned method for predicting and controlling the thermal and humidity loads of an aircraft environmental control system is implemented. This electronic device can be an onboard mission computer on an aircraft or a server at a ground station used for model training and data analysis.
[0052] Furthermore, embodiments of the present invention also provide a computer-readable storage medium, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), flash memory (FLASH), a solid-state drive (SSD), or any other non-transitory storage medium. One or more computer programs (or software units, instructions) are stored on this storage medium. When these programs are executed by the processor of a device (such as the aforementioned electronic device), the device performs the aforementioned method for predicting and controlling the thermal and humidity loads of an aircraft environmental control system. The programs may be stored in the form of source code, interpreted language bytecode, or executable files compiled by a compiler.
[0053] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the embodiments of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for predicting and controlling the thermal and humidity load of an aircraft environmental control system, characterized in that, Includes the following steps: Collect historical operating data of the environmental control system, identify, repair, and standardize the data for missing and outlier values, and construct an offline database; Based on the offline database, a preset number of parameters are selected from the historical operation data through correlation analysis as input to construct a BP neural network prediction model; by traversing the number of intermediate layer nodes of the BP neural network, the network structure is determined with the minimization of prediction error as the evaluation index, and an optimization algorithm is used to optimize the weights and thresholds of the BP neural network to obtain the optimized prediction model. Real-time status data of the environmental control system is collected and input into the optimized prediction model to obtain the predicted value of thermal and humidity load. Based on the predicted value of thermal and humidity load, the operating parameters of the environmental control system are determined through active decision-making with the goal of optimizing system energy consumption, and the operating parameters are sent to the environmental control system actuator of the aircraft to regulate the environmental control system.
2. The method according to claim 1, characterized in that, The outlier identification method uses the Tukey fence method; the missing and outlier repair is performed using the K-Means clustering algorithm based on Euclidean distance.
3. The method according to claim 1, characterized in that, The selection of the preset number of parameters is accomplished by calculating the Pearson correlation coefficient between the parameters and the heat and humidity load, and selecting parameters with an absolute value of correlation coefficient greater than 0.4; the selected parameters include at least one of environmental parameters, equipment power parameters, historical heat and humidity load parameters, and environmental control target parameters.
4. The method according to claim 1, characterized in that, The number of intermediate layer nodes traversed in the BP neural network is evaluated using the mean absolute deviation as the evaluation metric; the optimization algorithm is the whale optimization algorithm.
5. The method according to claim 1, characterized in that, The active decision-making is based on the predicted heat and humidity load values, with the goal of optimizing system energy consumption, to determine at least one control parameter among fan speed, pipeline flow rate, and equipment start-up / shutdown status.
6. A device for predicting and controlling the thermal and humidity load of an aircraft environmental control system, used to implement the method according to any one of claims 1 to 5, characterized in that, include: The data acquisition and preprocessing module is configured to acquire historical operating data of the environmental control system, and to identify and repair missing and outlier values and standardize the data to build an offline database. The model training and optimization module is configured to, based on the offline database, select a preset number of parameters from the historical running data through correlation analysis as input to construct a BP neural network prediction model, determine the network structure by traversing the number of intermediate layer nodes of the BP neural network and using the minimization of prediction error as the evaluation index, and optimize the weights and thresholds of the BP neural network using an optimization algorithm to obtain the optimized prediction model. The online prediction and decision control module is configured to collect real-time status data of the environmental control system and input it into the optimized prediction model to obtain the predicted value of heat and humidity load, and determine the operating parameters of the environmental control system based on the predicted value of heat and humidity load through active decision-making with the goal of optimizing system energy consumption. The environmental control system actuator is configured to receive the operating parameters and perform corresponding control actions.
7. The apparatus according to claim 6, characterized in that, The online prediction and decision control module and the environmental control system actuator constitute a closed-loop control system, and feed back the actual operating data of the environmental control system to the data acquisition and preprocessing module for updating the offline database and re-optimizing the prediction model.
8. The apparatus according to claim 6, characterized in that, The device also includes a feedforward compensation module, configured to adjust the actuators of the environmental control system in advance based on the flight profile signal sent by the flight management system of the aircraft and / or the environmental sensor signal of the environmental control system.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 5.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 5.