Intelligent control method and system for airplane ground air conditioner based on multi-source data fusion

Abstract

This invention relates to the field of air conditioning control technology, and particularly to an intelligent control method and system for aircraft ground air conditioning based on multi-source data fusion. The method involves: interacting with sensor data sources to acquire multi-source objective data; combining this multi-source objective data with a preset thermodynamic model to generate baseline control commands; interacting with a personnel information platform to collect anonymized information and obtain multi-source passenger information for the target flight; calculating preference coefficients based on the multi-source passenger information and a pre-trained preference mapping model, and performing preference compensation on the baseline control commands to obtain standard control commands; and performing ground air conditioning control operations according to the standard control commands. This achieves the technical effect of enabling aircraft ground air conditioning control to match the cabin's internal and external environment and the performance of the air conditioning equipment, while also catering to the thermal comfort preferences of different passenger groups, and ensuring data security and operational compliance.

Description

Technical Field

[0001] This invention relates to the field of air conditioning control technology, and in particular to an intelligent control method and system for aircraft ground air conditioning based on multi-source data fusion. Background Technology

[0002] In the field of aircraft ground air conditioning control, traditional technologies are mainly based on a single environmental parameter plus human experience in the control mode. They focus only on single-dimensional data such as the average temperature inside the cabin or the external ambient temperature, and mainly rely on preset fixed temperature thresholds or human adjustment of the air conditioning operation status based on experience. They do not consider the impact of individual differences such as the number of passengers, gender, and age on thermal comfort needs, resulting in technical problems such as low control accuracy and inability to meet the personalized thermal comfort needs of passengers. Summary of the Invention

[0003] This invention addresses the technical problems of low control precision and inability to meet the personalized thermal comfort needs of passengers in existing technologies by providing an intelligent control method and system for aircraft ground air conditioning based on multi-source data fusion.

[0004] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0005] In a first aspect, the present invention provides an intelligent control method for aircraft ground air conditioning based on multi-source data fusion, comprising: interacting with a sensor data source to acquire multi-source objective data, and combining the multi-source objective data with a preset thermodynamic model to generate a baseline control command; interacting with a personnel information platform to collect de-identified information and acquire multi-source passenger information for a target flight, wherein the multi-source passenger information includes at least passenger quantity information and passenger attribute information; calculating preference coefficients based on the multi-source passenger information and a pre-trained preference mapping model, and performing preference compensation on the baseline control command to obtain a standard control command; and performing control operations on the ground air conditioning according to the standard control command.

[0006] Optionally, the multi-source objective data includes at least cabin interior environmental data, external environmental data, and intrinsic performance parameters of the ground air conditioning system. The cabin interior environmental data includes the cabin interior temperature distribution, the external environmental data includes the external ambient temperature and external solar radiation intensity, and the intrinsic performance parameters include the adjustable performance range of the ground air conditioning system.

[0007] Specifically, the process of generating a baseline control command by combining the multi-source objective data with a preset thermodynamic model includes: calculating the real-time heat load of the cabin based on the cabin interior environmental data and the external environmental data using the thermodynamic model; and generating the baseline control command for controlling the operation of the ground air conditioning system based on the real-time heat load, the intrinsic performance parameters, and the preset standard cabin temperature.

[0008] The interactive personnel information platform collects anonymized information to obtain multi-source passenger information for the target shift. This multi-source passenger information includes at least passenger quantity information and passenger attribute information. The process involves: accessing the personnel information platform to collect anonymized information according to preset information collection standards, and obtaining the anonymized information collection results; statistically obtaining passenger quantity information based on the anonymized information collection results; parsing the anonymized information collection results, extracting passenger gender and age respectively, and constructing corresponding passenger gender and age distributions to obtain the passenger attribute information; and merging and outputting the passenger attribute information and passenger quantity information as the multi-source passenger information.

[0009] The acquisition of the preference mapping model includes: interacting with the personnel information platform to obtain historical crew information, historical baseline control commands, historical average cabin temperature, and corresponding historical somatosensory feedback information, wherein the historical somatosensory feedback information includes a thermal sensation scale; extracting and constructing historical crew gender distribution and historical crew age distribution based on the historical crew information; extracting the thermal sensation scale based on the historical somatosensory feedback information, calculating the corresponding average thermal sensation scale, and calculating and obtaining a first target preference coefficient in combination with the historical average cabin temperature; obtaining the prior comfort temperature corresponding to the historical crew information by combining with a thermal comfort knowledge graph, and calculating and obtaining a second target preference coefficient based on the historical average cabin temperature and the prior comfort temperature; constructing and training the preference mapping model using the historical crew gender distribution, the historical crew age distribution, and the historical baseline control commands as inputs, and using the mean of the first target preference coefficient and the second target preference coefficient as supervision.

[0010] The process of performing preference compensation on the baseline control command to obtain the standard control command further includes: inputting the gender distribution of the passengers, the age distribution of the passengers, and the baseline control command into the preference mapping model to obtain the preference coefficient; calculating based on the multi-source passenger information and the personnel heat load calculation model to generate the personnel heat load coefficient; and adjusting the baseline control command with weights according to the personnel heat load coefficient and the preference coefficient to obtain the standard control command.

[0011] This also includes: obtaining real-time updated boarding status information; calculating and obtaining the boarding time for each boarding passenger based on the boarding status information, and obtaining the real-time boarding time distribution; obtaining the median boarding time based on the real-time boarding time distribution, and calculating the real-time activity correction coefficient based on the ratio of the median boarding time to the preset typical seating time; and compensating the standard control command based on the real-time activity correction coefficient.

[0012] Secondly, the present invention provides an intelligent control system for aircraft ground air conditioning based on multi-source data fusion, comprising:

[0013] The benchmark control command generation module is used to interact with the sensor data source, acquire multi-source objective data, and combine the multi-source objective data with a preset thermodynamic model to generate benchmark control commands.

[0014] The multi-source passenger information acquisition module is used to collect de-identified information from the interactive personnel information platform and acquire multi-source passenger information for the target shift. The multi-source passenger information includes at least passenger quantity information and passenger attribute information.

[0015] The standard control instruction generation module is used to calculate the preference coefficient based on the multi-source passenger information and the pre-trained preference mapping model, and to perform preference compensation on the benchmark control instruction to obtain the standard control instruction;

[0016] The standard control command execution module is used to perform control operations on the ground air conditioner according to the standard control commands.

[0017] By implementing this invention, interactive sensor data sources can be used to acquire multi-source objective data. Combined with the multi-source objective data and a preset thermodynamic model, a baseline control command can be generated. The multi-dimensional objective data covers key internal, external, and equipment-specific factors affecting cabin temperature, avoiding control deviations caused by single data points. This ensures that the baseline control command is generated based on a real and complete environment and equipment status. Furthermore, by calculating real-time heat load through a professional thermodynamic model, compared to traditional experience-based control, it can more accurately match the actual heat dissipation and heat generation needs of the cabin, providing a precise "baseline" for subsequent control and reducing the blindness of air conditioning operation.

[0018] By implementing this invention, the interactive personnel information platform can collect de-identified information and obtain multi-source passenger information for the target shift. The multi-source passenger information includes at least passenger quantity information and passenger attribute information. The de-identification process ensures that passenger personal information is not leaked, which meets the requirements of data security and privacy protection. At the same time, it effectively obtains key passenger information, resolves the contradiction between data acquisition and privacy protection, and provides core data support for subsequent preference compensation, avoiding a one-size-fits-all control model.

[0019] By implementing this invention, it is possible to calculate preference coefficients based on the multi-source occupant information and a pre-trained preference mapping model, and to perform preference compensation on the baseline control commands to obtain standard control commands. The preference mapping model is trained based on historical occupant information, haptic feedback, and thermal comfort knowledge graphs, which can accurately match the thermal comfort preferences of different genders and age groups, avoiding discomfort for some occupants caused by the standardization of baseline control commands. Furthermore, by correcting the personnel thermal load coefficient, the influence of occupant number and attributes on cabin thermal load is integrated into the control commands, so that the standard control commands simultaneously meet environmental and personnel needs, further improving the accuracy of control.

[0020] By implementing this invention, the ground air conditioner can be controlled according to the standard control instructions. This replaces the traditional manual control based on experience, reduces human error, and ensures stable and controllable air conditioner operation.

[0021] In summary, by implementing this invention, it is possible to achieve the technical effect of enabling aircraft ground air conditioning control to match the internal and external environment of the cabin with the performance of the air conditioning equipment, to cater to the thermal comfort preferences of different passenger groups, and to ensure data security and operational standards. Attached Figure Description

[0022] Figure 1 A flowchart illustrating an intelligent control method for aircraft ground air conditioning based on multi-source data fusion provided by this invention;

[0023] Figure 2 This invention provides a schematic diagram of the structure of an intelligent control system for aircraft ground air conditioning based on multi-source data fusion.

[0024] In the attached diagram, the components represented by each number are as follows:

[0025] The module includes a baseline control command generation module 11, a multi-source occupant information acquisition module 12, a standard control command generation module 13, and a standard control command execution module 14. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0028] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0029] Example 1, as Figure 1 As shown, this embodiment of the invention provides an intelligent control method for aircraft ground air conditioning based on multi-source data fusion, including:

[0030] S100: Interactive sensing data source, acquires multi-source objective data, and combines the multi-source objective data with a preset thermodynamic model to generate a benchmark control command;

[0031] S200: The interactive personnel information platform collects de-identified information to obtain multi-source passenger information for the target shift, wherein the multi-source passenger information includes at least passenger quantity information and passenger attribute information;

[0032] S300: Calculate the preference coefficients based on the multi-source passenger information and the pre-trained preference mapping model, and perform preference compensation on the benchmark control command to obtain the standard control command;

[0033] S400: Perform the control operation of the ground air conditioner according to the standard control command.

[0034] In step S100 of this application embodiment, the multi-source objective data includes at least cabin interior environmental data, external environmental data, and intrinsic performance parameters of the ground air conditioning. The cabin interior environmental data includes the cabin interior temperature distribution, the external environmental data includes the external ambient temperature and the external solar radiation intensity, and the intrinsic performance parameters include the adjustable performance range of the ground air conditioning.

[0035] In this application embodiment, obtaining the multi-source objective data provides a comprehensive and accurate quantitative basis for the benchmark control command, avoids the deviation caused by traditional single-data control, and lays a scientific control foundation.

[0036] The cabin environment data mainly refers to the temperature distribution inside the cabin. This can be achieved by strategically placing multiple temperature sensors in different areas such as under the seats, near the windows, and in the middle of the cabin to collect temperature data at each point in real time. The data is then integrated to form a temperature distribution map inside the cabin, rather than just collecting the temperature at a single center point. This ensures that the temperature distribution inside the cabin reflects the local thermal differences within the cabin.

[0037] The external environmental data includes external ambient temperature and external solar radiation intensity. The external ambient temperature can be directly collected from ambient temperature sensors on the aircraft exterior or airport ground to obtain real-time external ambient temperature. The external solar radiation intensity can be detected by solar radiation sensors deployed on the aircraft fuselage or in designated areas of the airport, measuring real-time solar radiation flux (unit: W / m²). 2 This allows for the quantification of the sun's impact on the cabin's heat input.

[0038] The intrinsic performance parameters refer to the adjustable performance range of the terrestrial air conditioner. The adjustable performance range depends on the hardware parameters of the terrestrial air conditioner system, including performance parameters such as maximum cooling capacity, minimum cooling capacity, maximum heating power, and fan speed adjustment range. The adjustable range can be directly obtained from the terrestrial air conditioner manual and used as the intrinsic performance parameters.

[0039] In step S100 of this application embodiment, a baseline control command is generated by combining the multi-source objective data with a preset thermodynamic model, including:

[0040] Based on the cabin interior environmental data and the external environmental data, the real-time heat load of the cabin is calculated using the thermodynamic model.

[0041] Based on the real-time heat load, the intrinsic performance parameters, and the preset standard cabin temperature, a reference control command is generated to control the operating status of the ground air conditioning system.

[0042] In this embodiment, the purpose of generating a baseline control command by combining the multi-source objective data with a preset thermodynamic model is to transform the multi-source objective data into a baseline control command that can be directly used for control through the thermodynamic model, thus solving the problem of the lack of theoretical basis in traditional empirical control. Simultaneously, combining the multi-source objective data can avoid control imbalances caused by relying solely on environmental data or equipment parameters. This achieves a balance between considering the influence of the cabin's internal and external environment on the thermal state, being limited by the actual operating capacity of the ground air conditioning, and anchoring to a preset standard cabin temperature target. These three factors work together to ensure the feasibility and effectiveness of the baseline control command.

[0043] In this embodiment, the preset thermodynamic model is a mathematical model based on heat transfer and thermodynamic principles. It is used to quantitatively calculate the heat exchange relationship between the cabin environment and the external environment, ultimately determining the energy regulation required to maintain the preset temperature. The model is constructed as follows: first, fixed parameters such as aircraft surface area, wall thickness, and thermal conductivity are obtained from aircraft cabin design drawings. Then, variable parameters such as convective heat transfer coefficient and transmittance are preset using industry standards or experimental data. Next, multiple sets of multi-source objective data—correlated sample data of actual heat load—are collected in a real-world scenario. These correlated samples are input into the thermodynamic model to calculate the theoretical heat load value. The deviation between the theoretical and actual heat load values ​​is compared, and the thermodynamic model parameters are optimized using algorithms such as the least squares method until the calculation error of the thermodynamic model is controlled within a preset range, such as ±5%. This yields the preset thermodynamic model.

[0044] Next, based on the cabin interior environmental data and the external environment data, the real-time heat load of the cabin needs to be calculated using the thermodynamic model. Specifically, data such as the cabin interior temperature distribution, external ambient temperature, and external solar radiation intensity need to be imported into a preset thermodynamic model. Based on the principles of heat transfer, the thermodynamic model calculates the heat exchange between the cabin and the external environment, such as heat transferred by solar radiation and heat dissipation from the cabin walls. Combining this with the difference between the initial temperature and the standard temperature inside the cabin, the real-time heat load (in kW) required to maintain the standard temperature is obtained, which is the cooling / heating power that the ground air conditioning needs to provide. This is the final output value, such as 5kW for cooling, 3kW for heating, etc.

[0045] Furthermore, based on the real-time heat load, the intrinsic performance parameters, and the preset standard cabin temperature, a baseline control command is generated to control the operation of the ground air conditioning system. First, the intrinsic performance parameters of the ground air conditioning system need to be retrieved, such as the adjustable cooling capacity range of 3-10kW and fan speed levels of 1-5, to determine if the real-time heat load is within the equipment's adjustment range. Then, based on the difference between the real-time heat load and the standard cabin temperature, the heat load requirement is converted into specific ground air conditioning operating parameters. For example, if the real-time heat load is 6kW, and the ground air conditioning system outputs 50% of its power when cooling at 6kW, a baseline control command of "cooling mode, power 50%" or "cooling mode, set temperature 25℃" corresponding to 50% power is generated.

[0046] If the real-time heat load exceeds the maximum / minimum performance range of the terrestrial air conditioner, such as the required cooling capacity being 8kW but the maximum cooling capacity of the terrestrial air conditioner being 6kW, the limit value of the terrestrial air conditioner will be automatically taken as the reference control command, such as "cooling mode, power 100%", or "cooling mode, set temperature 16℃" corresponding to 100% power, to avoid invalid commands.

[0047] Finally, following the above logic, baseline control commands are generated for the operating mode, power, fan speed, and other parameters of the terrestrial air conditioner. These baseline control commands for multiple parameters are then integrated into a complete baseline control command, which serves as the original basis for subsequent preference compensation. The operating modes of the terrestrial air conditioner include cooling, heating, and natural ventilation.

[0048] In step S200 of this application embodiment, the interactive personnel information platform performs de-identified information collection to obtain multi-source passenger information for the target shift. The multi-source passenger information includes at least passenger quantity information and passenger attribute information, including:

[0049] Based on the preset information collection specifications, access the personnel information platform to collect de-identified information and obtain the de-identified information collection results;

[0050] Based on the de-identified information collection results, the number of passengers was statistically obtained.

[0051] The anonymized information collection results are analyzed, and the gender and age of the pilots are extracted respectively. The gender distribution and age distribution of the pilots are constructed accordingly to obtain the pilot attribute information.

[0052] The combined output of the crew attribute information and the crew quantity information constitutes the multi-source crew information.

[0053] In this embodiment, the purpose of step S200 is to obtain key information about the passengers on the target shift while ensuring privacy and security, providing personnel-level data support for subsequent personalized control of ground air conditioning. Furthermore, by de-identifying and removing personally identifiable information, only statistical data is obtained, thus balancing data utilization with privacy protection.

[0054] First, it is necessary to access the personnel information platform to collect anonymized information in accordance with the preset information collection specifications, and obtain the results of the anonymized information collection. Specifically, the core rules for anonymization collection can be defined in advance in the system. For example, only the gender, age, and target shift identifier of the trainee can be retained, while personal identification information such as name and ID number can be removed. Among them, the age of the trainee can be identified by intervals, such as under 12 years old, 12-60 years old, and over 60 years old.

[0055] The ground air conditioning control system accesses the personnel information platform through pre-defined interfaces such as the airport passenger service system and the airline passenger management system. Based on the target flight identifier, it calls the anonymized data interface that conforms to the collection specifications to obtain the anonymized information collection results for that flight. Furthermore, access to the personnel information platform can be restricted to the ground air conditioning control system only through API keys, IP whitelist interface authentication, and other means, with access logs kept throughout the process to ensure personnel privacy and security.

[0056] Next, it is necessary to analyze the desensitized information collection results, extract the gender and age of the crew members respectively, and construct the gender distribution and age distribution of the crew members accordingly to obtain the crew member attribute information.

[0057] Specifically, the length of the "passengers" array in the anonymized information collection results can be counted to directly obtain the total number of passengers for the target shift. For example, if the array length is 80, the number of passengers is 80. Then, the "gender" field in the anonymized information is iterated to count the number of passengers for each gender, forming gender distribution data. For example, if the count is 45 males and 35 females, the gender distribution is 56.25% males and 43.75% females. Next, the "age range" field in the anonymized information is iterated to count the number of passengers in each age range, forming age distribution data. For example, if the count is "5 under 12 years old, 68 between 12 and 60 years old, and 7 over 60 years old," the age distribution is: 6.25% under 12 years old, 85% between 12 and 60 years old, and 8.75% over 60 years old.

[0058] Finally, the gender distribution and age distribution are integrated to form complete crew attribute information, and the crew attribute information and the crew quantity information are merged and output as the multi-source crew information.

[0059] In step S300 of this application embodiment, obtaining the preference mapping model includes:

[0060] Interact with the personnel information platform to obtain historical crew information, historical baseline control commands, historical average cabin temperature and corresponding historical somatosensory feedback information, wherein the historical somatosensory feedback information includes thermal sensation scale;

[0061] Based on the historical passenger information, the gender distribution and age distribution of historical passengers were extracted and constructed respectively.

[0062] Based on the historical somatosensory feedback information, the thermal sensation scale is extracted, the average thermal sensation scale is calculated accordingly, and the first target preference coefficient is calculated and obtained by combining the historical average cabin temperature.

[0063] By combining the thermal comfort knowledge graph, the prior comfort temperature corresponding to the historical occupant information is obtained, and the second target preference coefficient is calculated based on the historical average cabin temperature and the prior comfort temperature.

[0064] Using the historical crew gender distribution, the historical crew age distribution, and the historical baseline control instructions as inputs, and with the mean of the first target preference coefficient and the second target preference coefficient as supervision, the preference mapping model is constructed and trained.

[0065] In step S300 of this application embodiment, obtaining the preference mapping model is to establish a mapping relationship between occupant characteristics and control command preference compensation by integrating historical data and thermal comfort knowledge. This enables the model to accurately output preference coefficients that meet the thermal comfort needs of the group based on the current occupant information, solving the problem of uniform baseline control commands and providing a quantitative basis for personalized control.

[0066] First, it is necessary to interact with the personnel information platform to obtain historical passenger information, historical baseline control commands, historical average cabin temperature, and corresponding historical thermal feedback information. The historical passenger information refers to the passenger information of historical flights, i.e., the anonymized information mentioned in the previous steps. The historical baseline control commands are the baseline control commands corresponding to historical flights, such as cooling modes. The historical average cabin temperature is the actual average cabin temperature after executing historical baseline control commands on historical flights, such as 24.5℃. The historical thermal feedback information is a scale of passenger thermal sensation after executing historical baseline control commands on historical flights, such as a seven-level scale from -3 to +3, where -3 = extremely cold, -2 = cold, -1 = slightly cool, 0 = comfortable, +1 = slightly warm, +2 = warm, and +3 = extremely hot.

[0067] Then, based on the historical crew information, it is necessary to extract and construct the gender distribution and age distribution of the historical crew members. For example, there were 48 males, accounting for 60%; and 32 females, accounting for 40%. The age distribution was as follows: under 12 years old 6.25%, 12-60 years old 85%, and over 60 years old 8.75%.

[0068] Furthermore, it is necessary to extract the thermal sensation scale based on the historical somatosensory feedback information, calculate the corresponding average thermal sensation scale, and combine it with the historical average cabin temperature to calculate and obtain the first target preference coefficient.

[0069] Suppose that in the example above, a total of 60 valid thermal sensation scales were collected, specifically distributed as follows: -2 (2), -1 (8), 0 (35), +1 (13), +2 (2), with no feedback at -3 or +3. Then, we can first calculate the total score based on the thermal sensation scale assignments, and then divide it by the number of feedback items:

[0070] Specifically, the calculation method is as follows: Total score = (-2×2) + (-1×8) + (0×35) + (+1×13) + (+2×2) = -4 - 8 + 0 + 13 + 4 = 5. Then, the average thermal scale = total score ÷ number of feedback items = 5 ÷ 60 ≈ 0.083.

[0071] If the average thermal sensitivity scale is greater than 0, it indicates that the actual cabin temperature is slightly lower than the group's comfort requirements, meaning the feedback is slightly warm, and appropriate warming compensation is needed; if the average thermal sensitivity scale is less than 0, cooling compensation is needed. The absolute value of the first objective preference coefficient needs to reflect the intensity of the compensation demand.

[0072] Therefore, the first target preference coefficient can be calculated as follows: First target preference coefficient = 0.1 × (average thermal scale) + (preset standard temperature - historical average cabin temperature) × 0.02. Here, 0.1 and 0.02 are empirical weighting coefficients used to balance the influence of sensory feedback and cabin temperature deviation, and can be adjusted according to actual conditions.

[0073] Assuming the preset standard temperature is 25℃, substituting the data into the above formula yields: First target preference coefficient = 0.1 × 0.083 + (25 - 24.5) × 0.02 = 0.0083 + 0.01 = 0.0183. A positive first target preference coefficient indicates that the temperature needs to be slightly increased or the cooling intensity needs to be reduced based on the baseline control command, with the compensation direction being temperature increase; a negative first target preference coefficient indicates temperature decrease; and if the first target preference coefficient is 0, no compensation is performed.

[0074] Then, it is necessary to combine the thermal comfort knowledge graph to obtain the prior comfort temperature corresponding to the historical occupant information, and calculate the second target preference coefficient based on the historical average cabin temperature and the prior comfort temperature. The thermal comfort knowledge graph contains comfort temperature range data for different genders and age groups.

[0075] First, the difference between the historical average cabin temperature and the prior comfort temperature needs to be calculated as the cabin temperature deviation. Assuming that the prior comfort temperature corresponding to the historical occupant information obtained from the thermal comfort knowledge graph in the above example is 25.75℃, then the cabin temperature deviation = historical average cabin temperature - prior comfort temperature = 24.5 - 25.75 = -1.25℃. The cabin temperature deviation is negative, indicating that the actual cabin temperature is lower than the comfort temperature recommended by the knowledge graph, and temperature compensation is required.

[0076] Furthermore, the second target preference coefficient can be calculated as follows: Second target preference coefficient = -cabin temperature deviation × 0.015. Substituting the data, we get: Second target preference coefficient = -(-1.25) × 0.015 = 0.01875, with the compensation direction being temperature increase. Here, 0.015 is a weighting coefficient used to convert the cabin temperature deviation into a second target preference coefficient of the same order of magnitude as the first target preference coefficient. Its specific value can be adjusted according to the actual situation.

[0077] Furthermore, it is necessary to calculate the average of the first target preference coefficient and the second target preference coefficient to obtain the comprehensive target preference coefficient, which serves as the supervision label for the preference mapping model. In the example above, the comprehensive target preference coefficient = (0.0183 + 0.01875) ÷ 2 ≈ 0.0185.

[0078] Using this comprehensive target preference coefficient as the supervision label for training the preference mapping model, it means that when the input is "60% male, 40% female, 6.25% under 12 years old, 85% 12-60 years old, and 8.75% over 60 years old" and the baseline control command is cooling, the preference mapping model should output a preference coefficient of about 0.0185, which is used to make a small temperature increase compensation to the baseline control command, so that the control is more in line with the comfort needs of this group.

[0079] For the specific task type of preference mapping model, a random forest regression model can be used to build it.

[0080] In the preference mapping model parameter settings, the number of decision trees is 100. Ensemble learning of multiple trees reduces the risk of overfitting and balances model performance and computational efficiency. The maximum depth of each tree is 8. Limiting the tree growth depth avoids overfitting the training data while ensuring sufficient learning of feature relationships. The maximum number of features considered when splitting a node is set to 3. Three features are randomly selected for evaluation at each split, enhancing the model's randomness and generalization ability. The minimum number of samples for leaf nodes is set to 5. This ensures that leaf nodes have sufficient samples to support their predictions and improves prediction stability. The minimum number of samples for splitting internal nodes is set to 10. This controls the conditions for node splitting and avoids the generation of too many fragmented nodes.

[0081] In the preference mapping model settings, the learning rate is set to 0.05. This is used to adjust the update step size of the preference mapping model during the iteration process to ensure convergence stability. The regularization coefficient is set to 0.01. This slight regularization constraint reduces the risk of model overfitting. The random seed is set to 2024. The training samples consist of the historical passenger information, historical baseline control commands, historical average cabin temperature, and corresponding historical sensory feedback information. 3000 sets of historical sample data are selected and divided into training and validation sets in a 7:3 ratio. This covers scenarios with different flight types, seasons, time periods, and passenger characteristics to ensure the diversity and representativeness of the samples, meeting the needs of the preference mapping model to learn various situations.

[0082] The maximum number of training epochs is set to 200. Through multiple iterative training epochs, the preference mapping model gradually optimizes its parameters and fully learns the patterns in the data. When the change in the validation set loss value is less than 0.0001 over 15 consecutive training epochs, the model is considered converged, and training is stopped. This criterion effectively avoids overtraining and ensures the model's generalization performance on the validation set.

[0083] In step S300 of this application embodiment, the process of performing preference compensation on the benchmark control command to obtain the standard control command further includes:

[0084] Input the gender distribution of the crew members, the age distribution of the crew members, and the baseline control command into the preference mapping model to obtain the preference coefficients;

[0085] Based on the multi-source occupancy information and the personnel heat load calculation model, the personnel heat load coefficient is generated.

[0086] The baseline control command is weighted and adjusted based on the personnel heat load coefficient and the preference coefficient to obtain the standard control command.

[0087] In this embodiment of the application, the purpose of the above steps is to transform multi-source occupancy information into a quantitative basis for regulation and correction through a preference mapping model and a personnel heat load calculation model, so that the preference compensation process is interpretable and quantifiable, avoids the arbitrariness of subjective experience correction, and ensures that the standard control instructions are both scientific and in line with actual needs, providing the final execution basis for the precise operation of ground air conditioning.

[0088] First, the preference coefficient needs to be obtained. The gender distribution of the crew, the age distribution of the crew, and the baseline control command generated in step S200 are input into the preference mapping model, and the preference coefficient reflecting the thermal comfort preference of the crew group is output, such as 0.02. A positive value means that the temperature compensation needs to be slightly increased.

[0089] Then, based on the multi-source occupancy information and the occupancy heat load calculation model, it is necessary to calculate and generate the occupancy heat load coefficient. Assuming the occupancy is 80 people,

[0090] Specifically, the unit heat load parameters for different types of personnel need to be preset through the personnel heat load calculation model. For example, the heat generated by an adult at rest is about 100W / person, a child about 70W / person, and an elderly person about 80W / person. Based on the age distribution of the passengers, the weighted average unit heat load is calculated as (5×70+68×100+7×80)÷80=95.375W / person. Then, the total personnel heat load is calculated by combining the total number of passengers: total personnel heat load = 80×95.375=7630W=7.63kW. Finally, the ratio of the total personnel heat load to the cabin base heat load, i.e., the real-time heat load in S100, is used as the personnel heat load coefficient. For example, if the real-time heat load is 50kW, then the personnel heat load coefficient is 7.63÷50=0.1526, representing that the heat generated by personnel accounts for 15.26% of the total heat load.

[0091] Finally, the baseline control command needs to be weighted and adjusted based on the personnel heat load coefficient and the preference coefficient to obtain the standard control command.

[0092] First, it is necessary to set the weights of the preference coefficient and the personnel heat load coefficient. For example, the preference coefficient can be weighted at 0.6 and the personnel heat load coefficient at 0.4. These weights are determined based on the degree of reliance on the two parameters and can be adjusted according to the actual situation during implementation.

[0093] Then, based on the weights of the preference coefficient and the personnel heat load coefficient, the comprehensive correction coefficient needs to be calculated. In the example above, the comprehensive correction coefficient = preference coefficient × 0.6 + personnel heat load coefficient × 0.4 = 0.02 × 0.6 + 0.1526 × 0.4 = 0.012 + 0.061 = 0.073.

[0094] Next, the baseline control command is adjusted based on the comprehensive correction coefficient. If the comprehensive correction coefficient is positive and the baseline control command is in cooling mode, the set temperature is appropriately increased. For example, the correction method for the set temperature can be 25℃ + 0.073 × 10 ≈ 25.7℃. If the air conditioner supports 0.5℃ increments, the correction increment is in 0.5℃ units. That is, when the correction value for the set temperature is 25.7℃, the actual adjusted set temperature is 25.5℃. Here, 10 is a weighting value used to expand the correction range of the comprehensive correction coefficient, which can be adjusted according to the actual situation.

[0095] In step S300 of the embodiments of this application, the following is also included:

[0096] Get real-time updated boarding status information;

[0097] Based on the boarding status information, calculate and obtain the boarding time for each boarding passenger, and obtain the real-time boarding time distribution;

[0098] Based on the real-time boarding time distribution, the median boarding time is obtained, and the real-time activity correction coefficient is calculated based on the ratio of the median boarding time to the preset typical seating time.

[0099] The standard control command is compensated based on the real-time activity correction coefficient.

[0100] In this embodiment, the heat generation of passengers differs significantly between their active state during boarding and their resting state after being seated. By calculating a real-time activity correction coefficient based on the real-time boarding status, the shortcomings of previous calculations based solely on static passenger attributes can be overcome. This allows standard control commands to dynamically respond to changes in passenger status from active to resting, avoiding cabin temperature deviations caused by fluctuations in heat generation due to activity.

[0101] First, it is necessary to collect boarding data for the target flight in real time through the airport's passenger counting system, boarding pass scanning records, or seat sensors in the cabin. This includes information such as the number of passengers who have boarded and the boarding timestamps of each passenger. For example, if passenger A boards at 14:05 and passenger B boards at 14:08, a dynamically updated boarding status data stream is formed.

[0102] Then, based on the current time, the "boarding time" needs to be calculated for each boarding passenger, which is the current time minus the boarding timestamp. For example, if the current time is 14:10, passenger A's boarding time is 5 minutes, and passenger B's is 2 minutes. The boarding times of all boarding passengers are sorted in ascending order to form a real-time boarding time distribution, such as [1 minute, 2 minutes, 3 minutes, 5 minutes, 8 minutes...]. The median boarding time is taken as the median boarding time. For example, if there are 100 boarding passengers, the 50th passenger's boarding time of 6 minutes is taken as the median.

[0103] Next, a preset typical seating time is invoked. This typical seating time can be obtained through historical data statistics. For example, if passengers take an average of 8 minutes to sit down and enter a resting state after boarding, the preset typical seating time is 8 minutes. Further, a real-time activity correction coefficient is calculated using the formula: Real-time activity correction coefficient = 1 + (1 - median boarding time ÷ typical seating time) × 0.5. For example, if the median boarding time is 6 minutes and the typical seating time is 8 minutes, the real-time activity correction coefficient = 1 + (1 - 6 ÷ 8) × 0.5 = 1.125.

[0104] In the initial stage of boarding, the median boarding time is much shorter than the typical sitting time, and passengers are mostly in an active state. The real-time activity correction coefficient is >1, which means that enhanced compensation is needed. As the boarding time approaches the typical sitting time, the coefficient gradually approaches 1, and then the system returns to the resting state regulation.

[0105] Finally, the real-time activity correction factor is multiplied by the core parameters of the original standard control command to complete dynamic compensation. For example, if the original standard control command is cooling mode with a power of 60% and a real-time activity correction factor of 1.125, then the compensated command will be cooling mode with a power of 60% × 1.125 = 67.5%, thus increasing the cooling power to offset the extra heat generated by passenger activities. Once all passengers are seated, if the median duration is greater than or equal to the typical seating duration, the real-time activity correction factor will be 1, and the standard control command will revert to its original standard value.

[0106] Example 2, as Figure 2 As shown, based on the same inventive concept as the intelligent control method for aircraft ground air conditioning based on multi-source data fusion provided in Embodiment 1, this embodiment of the invention also provides an intelligent control system for aircraft ground air conditioning based on multi-source data fusion, comprising:

[0107] The benchmark control command generation module 11 is used to interact with the sensor data source, acquire multi-source objective data, and combine the multi-source objective data with a preset thermodynamic model to generate benchmark control commands.

[0108] The multi-source passenger information acquisition module 12 is used to collect de-identified information from the interactive personnel information platform and acquire multi-source passenger information for the target shift. The multi-source passenger information includes at least passenger quantity information and passenger attribute information.

[0109] The standard control instruction generation module 13 is used to calculate the preference coefficient based on the multi-source passenger information and the pre-trained preference mapping model, and to perform preference compensation on the benchmark control instruction to obtain the standard control instruction.

[0110] The standard control command execution module 14 is used to perform control operations on the ground air conditioner according to the standard control command.

[0111] Furthermore, the multi-source occupant information acquisition module 12 includes the following execution steps:

[0112] Based on the preset information collection specifications, access the personnel information platform to collect de-identified information and obtain the de-identified information collection results;

[0113] Based on the de-identified information collection results, the number of passengers was statistically obtained.

[0114] The anonymized information collection results are analyzed, and the gender and age of the pilots are extracted respectively. The gender distribution and age distribution of the pilots are constructed accordingly to obtain the pilot attribute information.

[0115] The combined output of the crew attribute information and the crew quantity information constitutes the multi-source crew information.

[0116] Furthermore, the standard control instruction generation module 13 includes the following execution steps:

[0117] Interact with the personnel information platform to obtain historical crew information, historical baseline control commands, historical average cabin temperature and corresponding historical somatosensory feedback information, wherein the historical somatosensory feedback information includes thermal sensation scale;

[0118] Based on the historical passenger information, the gender distribution and age distribution of historical passengers were extracted and constructed respectively.

[0119] Based on the historical somatosensory feedback information, the thermal sensation scale is extracted, the average thermal sensation scale is calculated accordingly, and the first target preference coefficient is calculated and obtained by combining the historical average cabin temperature.

[0120] By combining the thermal comfort knowledge graph, the prior comfort temperature corresponding to the historical occupant information is obtained, and the second target preference coefficient is calculated based on the historical average cabin temperature and the prior comfort temperature.

[0121] Using the historical crew gender distribution, the historical crew age distribution, and the historical baseline control instructions as inputs, and with the mean of the first target preference coefficient and the second target preference coefficient as supervision, the preference mapping model is constructed and trained.

[0122] Input the gender distribution of the pilots, the age distribution of the pilots, and the baseline control command into the preference mapping model to obtain the preference coefficients;

[0123] Based on the multi-source occupancy information and the personnel heat load calculation model, the personnel heat load coefficient is generated.

[0124] The baseline control command is weighted and adjusted based on the personnel heat load coefficient and the preference coefficient to obtain the standard control command.

[0125] Get real-time updated boarding status information;

[0126] Based on the boarding status information, calculate and obtain the boarding time for each boarding passenger, and obtain the real-time boarding time distribution;

[0127] Based on the real-time boarding time distribution, the median boarding time is obtained, and the real-time activity correction coefficient is calculated based on the ratio of the median boarding time to the preset typical seating time.

[0128] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0129] Those skilled in the art will understand that embodiments of the present invention can provide methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0130] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0131] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0132] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0133] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0134] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for intelligent control of aircraft ground air conditioning based on multi-source data fusion, characterized in that, include: Interactive sensing data source acquires multi-source objective data, and combines the multi-source objective data with a preset thermodynamic model to generate a benchmark control command; The interactive personnel information platform collects de-identified information to obtain multi-source passenger information for the target shift, wherein the multi-source passenger information includes at least passenger quantity information and passenger attribute information; Based on the multi-source passenger information and the pre-trained preference mapping model, the preference coefficients are calculated, and the baseline control command is compensated for the preference to obtain the standard control command. The ground air conditioning system is controlled according to the standard control instructions. The multi-source objective data includes at least cabin interior environmental data, external environmental data, and intrinsic performance parameters of the ground air conditioning system. The cabin interior environmental data includes the cabin interior temperature distribution, the external environmental data includes the external ambient temperature and external solar radiation intensity, and the intrinsic performance parameters include the adjustable performance range of the ground air conditioning system. Specifically, by combining the multi-source objective data with a preset thermodynamic model, a baseline control command is generated, including: Based on the cabin interior environmental data and the external environmental data, the real-time heat load of the cabin is calculated using the thermodynamic model. Based on the real-time heat load, the intrinsic performance parameters, and the preset standard cabin temperature, a reference control command is generated to control the operation status of the ground air conditioning system. The interactive personnel information platform collects anonymized information to obtain multi-source passenger information for the target shift. This multi-source passenger information includes at least passenger quantity information and passenger attribute information, including: Based on the preset information collection specifications, access the personnel information platform to collect de-identified information and obtain the de-identified information collection results; Based on the de-identified information collection results, the number of passengers was statistically obtained. The anonymized information collection results are analyzed, and the gender and age of the pilots are extracted respectively. The gender distribution and age distribution of the pilots are constructed accordingly to obtain the pilot attribute information. The combined output of the crew attribute information and the crew quantity information constitutes the multi-source crew information.

2. The intelligent control method for aircraft ground air conditioning based on multi-source data fusion as described in claim 1, characterized in that, The acquisition of the preference mapping model includes: Interact with the personnel information platform to obtain historical crew information, historical baseline control commands, historical average cabin temperature and corresponding historical somatosensory feedback information, wherein the historical somatosensory feedback information includes thermal sensation scale; Based on the historical passenger information, the gender distribution and age distribution of historical passengers were extracted and constructed respectively. Based on the historical somatosensory feedback information, the thermal sensation scale is extracted, the average thermal sensation scale is calculated accordingly, and the first target preference coefficient is calculated and obtained by combining the historical average cabin temperature. By combining the thermal comfort knowledge graph, the prior comfort temperature corresponding to the historical occupant information is obtained, and the second target preference coefficient is calculated based on the historical average cabin temperature and the prior comfort temperature. Using the historical crew gender distribution, the historical crew age distribution, and the historical baseline control instructions as inputs, and with the mean of the first target preference coefficient and the second target preference coefficient as supervision, the preference mapping model is constructed and trained.

3. The intelligent control method for aircraft ground air conditioning based on multi-source data fusion as described in claim 2, characterized in that, To obtain a standard control command by performing preference compensation on the benchmark control command, the method further includes: Input the gender distribution of the crew members, the age distribution of the crew members, and the baseline control command into the preference mapping model to obtain the preference coefficients; Based on the multi-source occupancy information and the personnel heat load calculation model, the personnel heat load coefficient is generated. The baseline control command is weighted and adjusted based on the personnel heat load coefficient and the preference coefficient to obtain the standard control command.

4. The intelligent control method for aircraft ground air conditioning based on multi-source data fusion as described in claim 1, characterized in that, Also includes: Get real-time updated boarding status information; Based on the boarding status information, calculate and obtain the boarding time for each boarding passenger, and obtain the real-time boarding time distribution; Based on the real-time boarding time distribution, the median boarding time is obtained, and the real-time activity correction coefficient is calculated based on the ratio of the median boarding time to the preset typical seating time. The standard control command is compensated based on the real-time activity correction coefficient.

5. An intelligent control system for aircraft ground air conditioning based on multi-source data fusion, characterized in that, The system is used to implement the intelligent control method for aircraft ground air conditioning based on multi-source data fusion as described in any one of claims 1-4, including: The benchmark control command generation module is used to interact with the sensor data source, acquire multi-source objective data, and combine the multi-source objective data with a preset thermodynamic model to generate benchmark control commands. The multi-source passenger information acquisition module is used to collect de-identified information from the interactive personnel information platform and acquire multi-source passenger information for the target shift. The multi-source passenger information includes at least passenger quantity information and passenger attribute information. The standard control instruction generation module is used to calculate the preference coefficient based on the multi-source passenger information and the pre-trained preference mapping model, and to perform preference compensation on the benchmark control instruction to obtain the standard control instruction; The standard control command execution module is used to perform control operations on the ground air conditioner according to the standard control commands.

Images