Cabin temperature intelligent adjusting method and system based on deep learning

Abstract

The invention discloses an intelligent cabin temperature adjusting method and system based on deep learning, and belongs to the technical field of cabin ventilation control. The method comprises the following steps: constructing an air temperature balance model in a cabin; obtaining a plurality of groups of model parameter sets, and training a neural network model through the model parameter sets; matching sampling point positions according to the personnel positions; according to the sampling point position, an air temperature balance model and a neural network model, acquiring a human body surrounding air velocity and a human body surrounding air temperature; and according to the human body surrounding air velocity, the human body surrounding air temperature and the comfort evaluation model, the air supply velocity and the air supply temperature corresponding to the optimal comfort prediction value are output. According to the method, the air flow speed around the human body and the air temperature around the human body are obtained, the air supply speed and the air supply temperature corresponding to the optimal comfort degree predicted value are output according to the comfort degree evaluation model, intelligent control over the cabin temperature is achieved, and the state that people feel most comfortable in the cabin is maintained.

Description

Technical Field

[0001] The present invention relates to the technical field of cabin ventilation control, and in particular to a method and system for intelligently adjusting cabin temperature based on deep learning. Background Art

[0002] Cabin temperature control has a vital impact on the health and safety of the crew. At a suitable temperature, the work efficiency or entertainment experience of the crew and other personnel on board can be effectively guaranteed. In the actual cabin temperature control method, the temperature in the cabin is affected by many factors, such as changes in the external environment temperature, heat dissipation of the cabin equipment, and the density of personnel activities.

[0003] The existing cabin temperature control method is often adjusted by operators based on experience. However, the temperature inside the cabin varies due to differences in physical conditions in various parts of the cabin, and the influencing parameters of the ship's external environment and the cabin are complex and changeable. The space inside the cabin is limited and the temperature distribution is often uneven. The adjustment method based on experience relies on the experience and subjective judgment of engineers, and there is a large error between the temperature control effect and the subjective feelings of the people in the cabin. Summary of the invention

[0004] The present invention provides a method and system for intelligently adjusting cabin temperature based on deep learning to solve at least one of the problems mentioned in the above background technology.

[0005] The specific technical solutions provided by this application are as follows: A method for intelligently adjusting cabin temperature based on deep learning, comprising the steps of: Constructing an air temperature balance model inside the cabin; the air temperature balance model is used to predict the average cabin air temperature under stable working conditions based on the air supply speed, air supply temperature, environmental influencing parameters and cabin influencing parameters; Acquire several sets of model parameter sets, and train the neural network model through the model parameter sets; the model parameter sets include air supply speed, air supply temperature, average cabin air temperature, sampling point temperature parameter group and sampling point wind speed parameter group; the cabin air average temperature is acquired through an air temperature balance model; the air supply speed, air supply temperature and cabin air average temperature are input data of the neural network model; the sampling point temperature parameter group and the sampling point wind speed parameter group are output data of the neural network model; Obtain the position of the personnel inside the cabin, and match the sampling point position according to the personnel position; obtain the air flow rate and air temperature around the human body according to the sampling point position, air temperature balance model and neural network model; A comfort evaluation model is constructed based on the PMV model; and the air supply speed and air supply temperature corresponding to the optimal comfort prediction value are output according to the air flow rate around the human body, the air temperature around the human body and the comfort evaluation model.

[0006] As a preferred solution, the environmental influencing parameters include the external temperature of several heat transfer surfaces and the solar radiation heat transmittance of the lighting surface; the cabin influencing parameters include the heat generated by equipment, the heat generated by human body and the heat gained by lighting.

[0007] As a preferred solution, the air temperature balance model is expressed as: ; in, represents the average temperature of the air in the cabin; i represents the number of the heat transfer surface, and n represents the total number of heat transfer surfaces; represents the area of ​​the ith heat transfer surface, represents the heat transfer coefficient of the ith heat transfer surface, represents the external temperature of the i-th heat transfer surface; j represents the number of the illuminated surface, and m represents the number of illuminated surfaces; represents the area of ​​the j-th illuminated surface, represents the solar radiation heat transmittance of the j-th illuminated surface; Indicates the heat generated by the equipment in the cabin; Indicates the body's heat; Indicates the heat gain of lighting; is the preset constant; v is the air supply speed; is the supply air temperature.

[0008] As a preferred solution, the method of obtaining several groups of model parameter sets and training the neural network model through the model parameter sets includes the steps of: S21, obtaining environmental impact parameters and cabin impact parameters, and setting air supply speed and air supply temperature; based on the air temperature balance model, predicting the average cabin air temperature under stable working conditions according to the air supply speed, air supply temperature, environmental impact parameters and cabin impact parameters; S22, repeating step S21, constructing several groups of model parameter sets according to the obtained average temperature of the air in the cabin and the corresponding air supply speed and air supply temperature, and dividing the model parameter sets into a training parameter set and a verification parameter set; S23, constructing a neural network model; the neural network model is a multi-layer fully connected neural network; S24, during the training process, the air supply speed, air supply temperature and average cabin air temperature are input to the input layer according to the training parameter set, the weighted sum of each neuron is calculated layer by layer, and the activation function is applied to obtain the predicted values ​​of the sampling point temperature parameter group and the sampling point wind speed parameter group; S25, comparing the predicted value with the actual value in the training parameter set, and calculating the error using the loss function; and updating the parameters of the neural network model using the gradient descent method through back propagation; S26. Repeat steps S24 and S25 until the neural network model passes the performance evaluation of the verification parameter set.

[0009] As a preferred solution, the PMV model is expressed as: ; Among them, v is the air supply speed; is the air supply temperature; M represents the human metabolic rate, which is a fixed value; W represents the human body power, which is a fixed value set according to the cabin type, such as a higher value for the activity room and a lower value for the crew lounge; Indicates the air temperature around the human body, which is affected by the air supply speed and air supply temperature; Indicates the partial pressure of water vapor around the human body, in Pa, and is affected by the air temperature around the human body; It represents the ratio of the surface area of ​​the clothed body to the naked body, and is obtained based on the thermal resistance of the human body clothing; It indicates the average temperature of the outer surface of the clothed human body, in degrees Celsius, calculated by the temperature of the air surrounding the human body; Indicates the average radiation temperature in the cabin, in °C; It represents the surface heat transfer coefficient in °C and is affected by the air velocity and temperature around the human body.

[0010] As a preferred solution, the ratio of the surface area of ​​the clothed body to the naked body is expressed as: ; The surface heat transfer coefficient is expressed as: ; in, Indicates the air velocity around the human body; Represents the thermal resistance of human clothing.

[0011] As a preferred solution, the air supply speed and air supply temperature corresponding to the optimal comfort prediction value outputted by the air flow rate around the human body, the air temperature around the human body and the comfort evaluation model include the steps of: A group of particles are randomly generated, and the position and velocity of each particle are initialized; the position of the particle represents a group of supply air speed and supply air temperature; According to the air velocity around the human body, the air temperature around the human body and the comfort evaluation model, the comfort prediction value corresponding to each solution is obtained; Update the individual best position of each particle and the group best position of the particle swarm; Update the speed and position of each particle based on the individual best position and the group best position; Iterate until the maximum number of iterations or the comfort prediction value meets the output condition, and output the optimal position of the particle swarm as the supply air speed and supply air temperature corresponding to the optimal comfort prediction value.

[0012] As a preferred solution, the comfort evaluation model is expressed as: ; ; in, represents the predicted value of comfort; represents the PMV model, and PMV represents the output value of the PMV model; It indicates the predicted dissatisfaction rate; and All are preset weight parameters; and All are preset constants.

[0013] As a preferred solution, the velocity of the particle is expressed as: , The position of a particle is expressed as: , in, is the inertia weight, which represents the influence of the velocity of the previous generation of particles on the velocity of the current generation of particles; and They are individual learning factor and group learning factor respectively; is the individual best position of the particle, and g is the group best position of the group to which the particle belongs. and are all random numbers in the interval [0,1], used to provide randomness for each iteration; f is the number of the particle, represents the position of the fth particle at time t, represents the velocity of the fth particle at time t.

[0014] The present invention also provides a cabin temperature intelligent adjustment system based on deep learning, which is used to implement the aforementioned cabin temperature intelligent adjustment method based on deep learning, including: a data acquisition module, which is used to collect environmental influencing parameters and cabin influencing parameters; A data processing module is used to construct an air temperature balance model inside the cabin, obtain a model parameter set, and construct a comfort evaluation model based on the PMV model; A model training module, used to train a neural network model through a model parameter set; A comfort evaluation module, configured to output the air supply speed and air supply temperature corresponding to the optimal comfort prediction value according to the air temperature balance model, the neural network model and the comfort evaluation model; The air supply regulating module is used to control the air supply equipment according to the air supply speed and air supply temperature.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present application constructs an air temperature balance model and a neural network model, matches the sampling point position according to the personnel position, obtains the air flow rate and the air temperature around the human body according to the sampling point position, the air temperature balance model and the neural network model, thereby ensuring accurate measurement of the environment around the human body and making the comfort assessment more reliable; outputs the supply air speed and supply air temperature corresponding to the optimal comfort prediction value according to the air flow rate and the air temperature around the human body and the comfort assessment model, thereby realizing intelligent control of the cabin temperature, reducing dependence on manual experience adjustment, improving the level of automation, and maintaining the cabin in the most comfortable state for personnel.

[0016] The embodiment of the present application obtains the comprehensive index of the average cabin air temperature by preprocessing the environmental influencing parameters and the cabin influencing parameters in the air temperature balance model, which can effectively narrow the range of input data, reduce the complexity of the neural network model, and improve the convergence speed and prediction effect of the neural network. In addition, the average cabin air temperature can more accurately and stably reflect the actual situation in the cabin, avoiding the interference of a single environmental influencing parameter or cabin influencing parameter fluctuations on the final prediction result, making the training and prediction of the neural network model more universal and stable.

[0017] In the embodiment of the present application, by combining PMV and PPD for weighted calculation, the comfort evaluation model can more comprehensively reflect the actual comfort and dissatisfaction rate of the current environment, and reduce the deviation that may be caused by single indicator evaluation; based on the comfort prediction value output by the comfort evaluation model, the particle swarm optimization algorithm can be used to quickly determine the air supply equipment settings to achieve the best indoor comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0020] Figure 1 A schematic diagram of a process flow of a method for intelligently adjusting cabin temperature based on deep learning provided by an embodiment of the present invention; Figure 2A schematic diagram of a process for obtaining several groups of model parameter sets and training a neural network model through the model parameter sets provided in an embodiment of the present invention; Figure 3 A schematic flow chart of outputting the air supply speed and air supply temperature corresponding to the optimal comfort prediction value according to the air flow rate around the human body, the air temperature around the human body and the comfort evaluation model provided in an embodiment of the present invention; Figure 4 A schematic diagram of the structure of intelligent adjustment of cabin temperature based on deep learning provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0023] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0024] With the rapid development of the global economy and the improvement of people's quality of life, more and more people choose ships as a way of entertainment and sightseeing. In order to enrich the user's activity experience, ships can provide automatic driving and assisted manual driving for users to use. The introduction of automatic driving mode not only improves the safety of navigation, but also allows users to avoid fatigue during long-term navigation; at the same time, the existence of assisted manual driving mode provides users with more autonomy and fun.

[0025] However, this approach also places higher demands on the ship's driving experience and safety. For ship engines, the ship control instructions that control their working status need to be able to smoothly switch between different driving modes; for the engine push rods that interact directly with the user, it is necessary to ensure that they are easy to operate and avoid the risk of misoperation, so as to ensure the safe operation of the ship while satisfying the user's driving experience.

[0026] To solve the above problems, please refer to Figure 1 The present invention provides a method for intelligently adjusting cabin temperature based on deep learning, comprising the steps of: S1. Constructing an air temperature balance model inside the cabin; the air temperature balance model is used to predict the average cabin air temperature under stable working conditions according to the air supply speed, air supply temperature, environmental influencing parameters and cabin influencing parameters; In this embodiment, the air temperature balance model predicts the average cabin air temperature under stable working conditions based on the air supply speed and air supply temperature set manually or by a computer program, as well as the environmental influencing parameters and cabin influencing parameters acquired in real time.

[0027] Among them, the stable working condition of the air temperature balance model refers to the working condition when the system reaches thermal equilibrium and the temperature state of each part no longer changes. Under this working condition, the heat input and output in the system are balanced, and the air temperature distribution in the entire environment is stable. Specifically, the stable working condition has the following characteristics: Under stable operating conditions, the temperature value at any position in the system does not change with time, that is, the temperature at each point reaches a steady state. Although the temperatures at different positions may be different, the temperature at each point remains constant. The heat input and output of the entire system are equal, that is, the heat generated by the external heat source (environmental influencing parameters) and the internal heat source (cabin influencing parameters) and the heat discharged by the system are equal, and there is no net heat accumulation or loss. The supply and return air volumes are stable, and the various parameters of air flow (such as wind speed and flow direction) remain constant without significant changes. All surfaces of objects in the system reach a balance through radiation exchange of heat, and the heat radiated to the outside by each surface is equal to the radiation heat received from the outside.

[0028] In one embodiment, the environmental influencing parameters include the external temperature of several heat transfer surfaces and the solar radiation heat transmittance of the lighting surface; the cabin influencing parameters include the heat generated by equipment, the heat generated by human body and the heat gained by lighting.

[0029] Among them, for the external temperature of the heat transfer surface, the higher the external temperature, the more heat is transferred to the cabin through the heat transfer surface, and more air conditioning and cooling capacity is required to maintain a comfortable temperature in the cabin. For the solar radiation heat transmittance of the illuminated surface, some surfaces of the cabin may be directly exposed to sunlight, such as the deck and portholes; light radiation will increase the heat load and affect the temperature in the cabin; the higher the solar radiation heat transmittance, the more heat is absorbed by the illuminated surface.

[0030] For equipment heat generation, the usage status of electrical equipment in the cabin will affect its heat generation, and thus affect the average temperature in the cabin; for human body heat generation, the average heat generation of each person is about 100 watts (changes according to human activity status, such as rest, work, and exercise). The total human body heat generation can be estimated based on the number of crew members and their activity level; for lighting heat generation, the lighting system in the cabin will generate significant heat, especially high-power lighting equipment.

[0031] Based on the above embodiment, the air temperature balance model is expressed as: ; in, represents the average temperature of the air in the cabin; i represents the number of the heat transfer surface, and n represents the total number of heat transfer surfaces; represents the area of ​​the ith heat transfer surface, represents the heat transfer coefficient of the ith heat transfer surface, represents the external temperature of the i-th heat transfer surface; j represents the number of the illuminated surface, and m represents the number of illuminated surfaces; represents the area of ​​the j-th illuminated surface, represents the solar radiation heat transmittance of the j-th illuminated surface; Indicates the heat generated by the equipment in the cabin; Indicates the body's heat; Indicates the heat gain of lighting; is a preset constant, usually 0.288×1.163=0.334944; v is the air supply speed, in CMH; is the supply air temperature.

[0032] Based on the above content, the air temperature balance model can calculate the average cabin air temperature under different set values ​​of supply air speed and supply air temperature through the real-time acquired environmental influencing parameters and cabin influencing parameters.

[0033] S2. Obtain several sets of model parameter sets, and train the neural network model through the model parameter sets; the model parameter sets include air supply speed, air supply temperature, average cabin air temperature, sampling point temperature parameter group and sampling point wind speed parameter group; the cabin air average temperature is obtained through the air temperature balance model; the air supply speed, air supply temperature and cabin air average temperature are input data of the neural network model; the sampling point temperature parameter group and the sampling point wind speed parameter group are output data of the neural network model; Among them, the sampling point temperature parameter group and the sampling point wind speed parameter group are obtained by temperature sensors and wind speed sensors respectively set at several sampling points in the cabin. The sampling point locations are arranged evenly or focused according to the experimental design; in one embodiment, the sampling points are preferentially selected from areas where personnel activities are frequent, such as sofas and beds in the rest area, office chairs in the work area and other facilities. The environmental parameters in these places have a greater impact on the comfort of the occupants. It can be understood that this application does not need to limit the number and specific locations of the sampling points, but it is necessary to divide the facility areas, including: Define different facility areas within the experimental area (such as sofas and beds in the rest area, chairs and tables in the office area, etc.), and decorate each facility area with a unique logo.

[0034] Arrange one or more sampling points in each facility area according to the experimental design, and ensure that the sampling points cover locations where people frequently move around, such as sofas, beds, office chairs, etc.

[0035] Record precise location information (calibrated using a 3D coordinate system) for each area and sampling point.

[0036] Based on the foregoing, the average cabin air temperature in this step is obtained by inputting the supply air speed, supply air temperature, environmental impact parameters and cabin impact parameters into the air temperature balance model, so the average cabin air temperature is associated with the environmental impact parameters and cabin impact parameters. This step sets the supply air speed, supply air temperature and cabin air average temperature as the input data of the neural network model instead of directly using the supply air speed, supply air temperature, environmental impact parameters and cabin impact parameters. The reason is that the neural network model may face the problem of overfitting or low training efficiency when processing high-dimensional data. The number of parameters of environmental impact parameters and cabin impact parameters is large and the types are complex. Directly using these parameters as input data will cause the input dimension of the neural network model to be too high, thereby increasing the complexity of model training and calculation.

[0037] This embodiment pre-processes the environmental influencing parameters and the cabin influencing parameters in the air temperature balance model to obtain the comprehensive indicator of the average cabin air temperature, which can effectively narrow the range of input data, reduce the complexity of the neural network model, and improve the convergence speed and prediction effect of the neural network. In addition, the average cabin air temperature can more accurately and stably reflect the actual situation in the cabin, avoiding the interference of a single environmental influencing parameter or the fluctuation of the cabin influencing parameter on the final prediction result. This makes the training and prediction of the neural network model more universal and stable.

[0038] In this embodiment, a neural network model is trained based on the air supply speed, air supply temperature, average air temperature in the cabin, sampling point temperature parameter group and sampling point wind speed parameter group. After the training is completed, the neural network model can be deployed on the ship control host computer. In the actual application process, the host computer inputs the currently set air supply speed, air supply temperature and average air temperature in the cabin into the neural network model, and the sampling point temperature parameters and sampling points corresponding to each sampling point position can be predicted.

[0039] For further information, see Figure 2 The method of obtaining a plurality of sets of model parameter sets and training a neural network model by using the model parameter sets comprises the steps of: S21, obtaining environmental impact parameters and cabin impact parameters, and setting air supply speed and air supply temperature; based on the air temperature balance model, predicting the average cabin air temperature under stable working conditions according to the air supply speed, air supply temperature, environmental impact parameters and cabin impact parameters; S22, repeating step S21, constructing several groups of model parameter sets according to the obtained average temperature of the air in the cabin and the corresponding air supply speed and air supply temperature, and dividing the model parameter sets into a training parameter set and a verification parameter set; S23, constructing a neural network model; in one embodiment, the neural network model is a multi-layer fully connected neural network, including 3 fully connected layers, each layer including 64 to 128 neurons; S24, during the training process, the air supply speed, air supply temperature and average cabin air temperature are input to the input layer according to the training parameter set, the weighted sum of each neuron is calculated layer by layer, and the activation function is applied to obtain the predicted values ​​of the sampling point temperature parameter group and the sampling point wind speed parameter group; S25, comparing the predicted value with the actual value in the training parameter set, calculating the error using a loss function (e.g., mean square error); updating the parameters of the neural network model using back propagation using the gradient descent method; S26. Repeat steps S24 and S25 until the neural network model passes the performance evaluation of the verification parameter set.

[0040] This embodiment uses a neural network model to adaptively predict and control the environment in the cabin, and can maintain high-precision prediction under a variety of working conditions to improve the accuracy and stability of environmental control.

[0041] S3, obtaining the position of the personnel inside the cabin, matching the sampling point position according to the personnel position; obtaining the air velocity and air temperature around the human body according to the sampling point position, the air temperature balance model and the neural network model; The positions of people inside the cabin can be obtained by arranging multiple position acquisition devices (such as infrared sensors, cameras, etc.) inside the cabin to monitor the activity positions of people in real time.

[0042] Based on the above, the sampling points are evenly or focusedly arranged according to the experimental design, and the precise coordinate position of each sampling point in the cabin is used as the sampling point position. Accordingly, the sampling point position is matched according to the personnel position, specifically: Determine the facility area to which the personnel location belongs; based on the nearest neighbor principle, match the sampling point location closest to the personnel location within the facility area; For example, the facility area includes a rest area, a work area, an aisle, etc. After identifying that the person's position is the aisle, the distance between the person's position and several sampling point positions in the aisle is calculated, and the sampling point position closest to the person's position in the facility area is determined accordingly.

[0043] In the closed environment of the cabin, the temperature field and wind speed field are usually affected by the facilities in the cabin and show obvious regional characteristics. Since there are different heat sources, cold sources and facilities in each area that affect the convection mode, the environmental parameters of each area tend to a local equilibrium within it. For example, even if some sampling points are physically closer to the personnel, if these sampling points belong to different facility areas, their temperature and wind speed may not represent the situation in the area where the personnel are actually located. For example, there may be airflow barriers such as sofas between the rest area and the work area, resulting in obvious differences in wind speed and temperature between two points in two facility areas with close spatial distances.

[0044] This embodiment first determines the facility area where the personnel are located and selects the nearest sampling point in the area, thereby limiting the sampling results to an environment with relatively uniform conditions, thereby ensuring the accuracy and representativeness of the data.

[0045] S4. Construct a comfort evaluation model based on the PMV model; output the supply air speed and supply air temperature corresponding to the optimal comfort prediction value according to the air flow rate around the human body, the air temperature around the human body and the comfort evaluation model.

[0046] Among them, the PMV model is constructed by combining prior knowledge and the actual application scenario of this application, and the average evaluation of the crew's thermal comfort prediction in a specific environment is obtained by comprehensively considering multiple factors.

[0047] In one embodiment, the PMV model is expressed as: ; Among them, v is the air supply speed; is the air supply temperature; M represents the human metabolic rate, which is a fixed value; W represents the human body power, which is a fixed value set according to the cabin type, such as a higher value for the activity room and a lower value for the crew lounge; Indicates the air temperature around the human body, which is affected by the air supply speed and air supply temperature; Indicates the partial pressure of water vapor around the human body, in Pa, and is affected by the air temperature around the human body; It represents the ratio of the surface area of ​​the clothed body to the naked body, and is obtained based on the thermal resistance of the human body clothing; It indicates the average temperature of the outer surface of the clothed human body, in degrees Celsius, calculated by the temperature of the air surrounding the human body; Indicates the average radiation temperature in the cabin, in °C; It represents the surface heat transfer coefficient in °C and is affected by the air velocity and temperature around the human body.

[0048] In one embodiment, the ratio of the surface area of ​​the clothed body to the naked body is expressed as: ; The surface heat transfer coefficient is expressed as: ; in, Indicates the air velocity around the human body; It represents the thermal resistance of human clothing, which refers to the ability of clothing to prevent heat from being transferred from the body surface to the external environment under steady-state conditions.

[0049] As a preferred embodiment, please refer to Figure 3 The method of outputting the air supply speed and air supply temperature corresponding to the optimal comfort prediction value according to the air flow rate around the human body, the air temperature around the human body and the comfort evaluation model comprises the steps of: S41, randomly generate a group of particles, and initialize the position and speed of each particle; the position of the particle represents a group of air supply speed and air supply temperature; A group of particles is randomly generated, including determining the size of the particle group, such as 30 particles, and setting the position constraints of the particles according to the performance of the air supply equipment, that is, constraining the value range of the air supply speed and air supply temperature; the position of the particle corresponds to a point in the solution space, representing a potential solution to the optimization problem. For this embodiment, the position of a particle represents a set of air supply speed and air supply temperature. The speed of the particle is used to determine the moving direction and moving speed of the particle in the solution space.

[0050] S42, obtaining a comfort prediction value corresponding to each solution according to the air velocity around the human body, the air temperature around the human body and the comfort evaluation model; Based on the air temperature balance model and the neural network model, the air flow rate and air temperature around the human body are calculated, and the comfort evaluation model is used to obtain the comfort prediction value corresponding to the particle. In this embodiment, the comfort prediction value is used as the objective function or fitness function of the particle swarm optimization algorithm to ensure that the optimization direction of the solution can accurately reflect the actual needs.

[0051] S43, updating the personal best position of each particle and the group best position of the particle group; The personal best position refers to the position where the particle has the smallest objective function value among the positions it has experienced in history. Each particle will save this information during the search process as its best position in history. In each iteration of the particle swarm algorithm, the current particle position is compared with the personal best position. If the current position is better, the personal best position is updated to the current position.

[0052] The group best position refers to the position with the smallest objective function value among the individual best positions of all particles in the entire particle swarm. The group best position is global information, and all particles in the particle swarm will use the group best position to update their own speed and position. Similar to the individual best position, if a better position is found in each iteration, the group best position will be updated.

[0053] S44, updating the speed and position of each particle according to the individual best position and the group best position; Wherein, the velocity of the particle is expressed as: , The position of the particle is expressed as: , in, is the inertia weight, which represents the influence of the velocity of the previous generation of particles on the velocity of the current generation of particles. In one embodiment, it is set to 0.5. and They are individual learning factor and group learning factor, the individual learning factor is used to control the amplitude of particle movement to the individual optimal position, and the group learning factor is used to control the amplitude of particle movement to the group optimal position. In one embodiment, and All are set to 1.5; is the individual best position of the particle, and g is the group best position of the group to which the particle belongs. and are all random numbers in the interval [0,1], used to provide randomness for each iteration. f is the particle number, represents the position of the fth particle at time t, represents the velocity of the fth particle at time t.

[0054] S45, iterating until the maximum number of iterations or the comfort prediction value meets the output condition, outputting the optimal position of the particle swarm as the supply air speed and supply air temperature corresponding to the optimal comfort prediction value.

[0055] Among them, iteration refers to repeatedly executing steps S42-S44; the maximum number of iterations is preset at the time of initialization, and when the number of iterations reaches this value, the algorithm terminates. For the output condition of the comfort prediction value, a comfort threshold can be preset. For example, when the comfort prediction value exceeds a certain predetermined value, it is considered that a sufficiently good solution has been found, thereby terminating the iteration.

[0056] Assume that the maximum number of iterations in one round of optimization is set to 100 and the comfort threshold is 0.95: In each iteration, particles calculate their own comfort predictions and update their individual best positions and group best positions.

[0057] When the particle swarm has iterated 100 times, or a gbest comfort prediction value reaches 0.95 (for example, this value is reached at the 40th iteration), the iteration process stops.

[0058] The final group best position is considered to be the optimal solution, and the supply air speed and supply air temperature corresponding to the optimal solution are output.

[0059] In this embodiment, step S41-35 uses a particle swarm algorithm (PSO) to optimize the air supply settings to achieve the best comfort. Specifically, this embodiment combines the air temperature balance model and the neural network model to improve the accuracy of reflecting the actual environment and comfort requirements, and quickly determines the optimal wind speed and temperature combination by evaluating the comfort of different setting combinations to ensure that passengers can enjoy the best comfort.

[0060] As a preferred embodiment, the comfort evaluation model is expressed as: ; ; in, represents the predicted value of comfort; represents the PMV model, and PMV represents the output value of the PMV model; It represents the predicted dissatisfaction rate. The predicted dissatisfaction rate calculated based on PMV indicates the proportion of people who feel uncomfortable in the current environment. The value ranges from 0% (completely satisfied) to 100% (completely dissatisfied). and All are preset weight parameters; and are all preset constants. In one embodiment, , .

[0061] In this embodiment, by combining PMV and PPD for weighted calculation, the comfort evaluation model can more comprehensively reflect the actual comfort and dissatisfaction rate of the current environment and reduce the deviation that may be caused by single indicator evaluation; based on the comfort prediction value output by the comfort evaluation model, the particle swarm optimization algorithm can quickly determine the air supply equipment settings to achieve the best indoor comfort.

[0062] See also Figure 4 Based on the aforementioned cabin temperature intelligent adjustment method based on deep learning, the present invention also provides a cabin temperature intelligent adjustment system based on deep learning, comprising: Data acquisition module, used to collect environmental impact parameters and cabin impact parameters; A data processing module is used to construct an air temperature balance model inside the cabin, obtain a model parameter set, and construct a comfort evaluation model based on the PMV model; A model training module, used to train a neural network model through a model parameter set; A comfort evaluation module, configured to output the air supply speed and air supply temperature corresponding to the optimal comfort prediction value according to the air temperature balance model, the neural network model and the comfort evaluation model; The air supply regulating module is used to control the air supply equipment according to the air supply speed and air supply temperature.

[0063] The present application constructs an air temperature balance model and a neural network model, matches the sampling point position according to the personnel position, obtains the air flow rate and the air temperature around the human body according to the sampling point position, the air temperature balance model and the neural network model, thereby ensuring accurate measurement of the environment around the human body and making the comfort assessment more reliable; outputs the supply air speed and supply air temperature corresponding to the optimal comfort prediction value according to the air flow rate and the air temperature around the human body and the comfort assessment model, thereby realizing intelligent control of the cabin temperature, reducing dependence on manual experience adjustment, improving the level of automation, and maintaining the cabin in the most comfortable state for personnel.

[0064] In the embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the modules can be electrical, mechanical or other forms.

[0065] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0066] In addition, each functional module in each embodiment of the present application can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or software functional modules.

[0067] If the integrated module is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, read-only memory), random access memory (RAM, random access memory), disk or optical disk and other media that can store program code.

[0068] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for intelligently adjusting cabin temperature based on deep learning, characterized in that: Includes steps: Constructing an air temperature balance model inside the cabin; the air temperature balance model is used to predict the average cabin air temperature under stable working conditions based on the air supply speed, air supply temperature, environmental influencing parameters and cabin influencing parameters; Acquire several sets of model parameter sets, and train the neural network model through the model parameter sets; the model parameter sets include air supply speed, air supply temperature, average cabin air temperature, sampling point temperature parameter group and sampling point wind speed parameter group; the cabin air average temperature is acquired through an air temperature balance model; the air supply speed, air supply temperature and cabin air average temperature are input data of the neural network model; the sampling point temperature parameter group and the sampling point wind speed parameter group are output data of the neural network model; Obtain the position of the personnel inside the cabin, and match the sampling point position according to the personnel position; obtain the air flow rate and air temperature around the human body according to the sampling point position, air temperature balance model and neural network model; A comfort evaluation model is constructed based on the PMV model; and the air supply speed and air supply temperature corresponding to the optimal comfort prediction value are output according to the air flow rate around the human body, the air temperature around the human body and the comfort evaluation model.

2. The method for intelligently adjusting cabin temperature based on deep learning according to claim 1 is characterized in that: The environmental influencing parameters include the external temperature of several heat transfer surfaces and the solar radiation heat transmittance of the lighting surface; the cabin influencing parameters include the heat generated by equipment, the heat generated by human body and the heat gained by lighting.

3. The method for intelligently adjusting cabin temperature based on deep learning according to claim 2 is characterized in that: The air temperature balance model is expressed as: ； in, represents the average temperature of the air in the cabin; i represents the number of the heat transfer surface, and n represents the total number of heat transfer surfaces; represents the area of ​​the ith heat transfer surface, represents the heat transfer coefficient of the ith heat transfer surface, represents the external temperature of the i-th heat transfer surface; j represents the number of the illuminated surface, and m represents the number of illuminated surfaces; represents the area of ​​the j-th illuminated surface, represents the solar radiation heat transmittance of the j-th illuminated surface; Indicates the heat generated by the equipment in the cabin; Indicates the body's heat; Indicates the heat gain of lighting; is the preset constant; v is the air supply speed; is the supply air temperature.

4. The method for intelligently adjusting cabin temperature based on deep learning according to claim 1 is characterized in that: The method of obtaining a plurality of sets of model parameter sets and training a neural network model by using the model parameter sets comprises the following steps: S21, obtaining environmental impact parameters and cabin impact parameters, and setting air supply speed and air supply temperature; based on the air temperature balance model, predicting the average cabin air temperature under stable working conditions according to the air supply speed, air supply temperature, environmental impact parameters and cabin impact parameters; S22, repeating step S21, constructing several groups of model parameter sets according to the obtained average temperature of the air in the cabin and the corresponding air supply speed and air supply temperature, and dividing the model parameter sets into a training parameter set and a verification parameter set; S23, constructing a neural network model; the neural network model is a multi-layer fully connected neural network; S24, during the training process, the air supply speed, air supply temperature and average cabin air temperature are input to the input layer according to the training parameter set, the weighted sum of each neuron is calculated layer by layer, and the activation function is applied to obtain the predicted values ​​of the sampling point temperature parameter group and the sampling point wind speed parameter group; S25, comparing the predicted value with the actual value in the training parameter set, and calculating the error using the loss function; and updating the parameters of the neural network model using the gradient descent method through back propagation; S26. Repeat steps S24 and S25 until the neural network model passes the performance evaluation of the verification parameter set.

5. The method for intelligently adjusting cabin temperature based on deep learning according to claim 1, characterized in that: The PMV model is expressed as: ； Among them, v is the air supply speed; is the air supply temperature; M represents the human metabolic rate, which is a fixed value; W represents the human body power, which is a fixed value set according to the cabin type, such as a higher value for the activity room and a lower value for the crew lounge; Indicates the air temperature around the human body, which is affected by the air supply speed and air supply temperature; Indicates the partial pressure of water vapor around the human body, in Pa, and is affected by the air temperature around the human body; It represents the ratio of the surface area of ​​the clothed body to the naked body, and is obtained based on the thermal resistance of the human body clothing; It indicates the average temperature of the outer surface of the clothed human body, in degrees Celsius, calculated by the temperature of the air surrounding the human body; Indicates the average radiation temperature in the cabin, in °C; It represents the surface heat transfer coefficient in °C and is affected by the air velocity and temperature around the human body.

6. The method for intelligently adjusting cabin temperature based on deep learning according to claim 5 is characterized in that: The ratio of the surface area of ​​the clothed body to the naked body is expressed as: ； The surface heat transfer coefficient is expressed as: ； in, Indicates the air velocity around the human body; Represents the thermal resistance of human clothing.

7. The method for intelligently adjusting cabin temperature based on deep learning according to claim 1 is characterized in that: The method outputs the air supply speed and air supply temperature corresponding to the optimal comfort prediction value according to the air flow rate around the human body, the air temperature around the human body and the comfort evaluation model, comprising the steps of: A group of particles are randomly generated, and the position and velocity of each particle are initialized; the position of the particle represents a group of supply air speed and supply air temperature; According to the air velocity around the human body, the air temperature around the human body and the comfort evaluation model, the comfort prediction value corresponding to each solution is obtained; Update the individual best position of each particle and the group best position of the particle swarm; Update the speed and position of each particle based on the individual best position and the group best position; Iterate until the maximum number of iterations or the comfort prediction value meets the output condition, and output the optimal position of the particle swarm as the supply air speed and supply air temperature corresponding to the optimal comfort prediction value.

8. The method for intelligently adjusting cabin temperature based on deep learning according to claim 7 is characterized in that: The comfort evaluation model is expressed as: ； ； in, represents the predicted value of comfort; represents the PMV model, and PMV represents the output value of the PMV model; It indicates the predicted dissatisfaction rate; and All are preset weight parameters; and All are preset constants.

9. The method for intelligently adjusting cabin temperature based on deep learning according to claim 7 is characterized in that: The velocity of the particle is expressed as: ， The position of a particle is expressed as: ， in, is the inertia weight, which represents the influence of the velocity of the previous generation of particles on the velocity of the current generation of particles; and They are individual learning factor and group learning factor respectively; is the individual best position of the particle, g is the group best position of the group to which the particle belongs, and are all random numbers in the interval [0,1], used to provide randomness for each iteration; f is the number of the particle, represents the position of the fth particle at time t, represents the velocity of the fth particle at time t.

10. A cabin temperature intelligent adjustment system based on deep learning, characterized in that: The method for intelligently adjusting cabin temperature based on deep learning as described in any one of claims 1 to 9 is applied, comprising: a data acquisition module for collecting environmental influencing parameters and cabin influencing parameters; A data processing module is used to construct an air temperature balance model inside the cabin, obtain a model parameter set, and construct a comfort evaluation model based on the PMV model; A model training module, used to train a neural network model through a model parameter set; A comfort evaluation module, configured to output the air supply speed and air supply temperature corresponding to the optimal comfort prediction value according to the air temperature balance model, the neural network model and the comfort evaluation model; The air supply regulating module is used to control the air supply equipment according to the air supply speed and air supply temperature.

Images