Control method and control system of overhead air conditioner

Abstract

The invention relates to the technical field of air conditioning equipment, in particular to a control method and a control system of an overhead air conditioner, and aims to solve the problems that the overhead air conditioner only takes air temperature as a control target and neglects the influence of other factors on the comfort of passengers, so that the comfort of the passengers in a vehicle is low, and the riding experience of the passengers is influenced. In order to achieve the purpose, the control method of the overhead air conditioner comprises the following steps that an in-vehicle environment parameter set and an in-vehicle personnel parameter set are obtained; according to the in-vehicle environment parameter set and the in-vehicle personnel parameter set, an overall thermal comfort index is obtained through calculation; and the operation state of the overhead air conditioner is adjusted according to the overall thermal comfort index. According to the method, the parameter sets of two dimensions including the in-vehicle environment parameter set and the in-vehicle personnel parameter set are collected, the overall thermal comfort index is obtained through calculation, the overall thermal comfort index reflecting the real thermal feeling of the human body is used for adjusting the running state of the overhead air conditioner, and the thermal comfort of passengers in the vehicle and the riding experience can be improved.

Description

Technical Field

[0001] This invention relates to the field of air conditioning technology, specifically providing a control method and control system for a rooftop air conditioner. Background Technology

[0002] Rooftop air conditioners are widely used in the onboard environmental control systems of large commercial vehicles such as buses, coaches, long-distance coaches, and RVs. Their core function is to provide cooling, heating, ventilation, and air purification for the enclosed passenger cabin to ensure the comfort and safety of passengers. Compared with air conditioning in passenger cars and buildings, the operating environment of rooftop air conditioners in large commercial vehicles is more complex, requiring them to cope with various uncertainties such as frequent vehicle starts and stops, dynamic changes in passenger cabin density, and changes in external ambient temperature.

[0003] In existing technologies, most rooftop air conditioners use simple temperature settings and airflow adjustment methods. During operation, the air conditioner only uses air temperature as the control target, ignoring the impact of other factors such as air velocity, air humidity, and the thickness of clothing worn by passengers on passenger comfort. This results in lower passenger comfort inside the vehicle and affects the passenger's travel experience.

[0004] Accordingly, there is a need in this field for a new control method and control system for rooftop air conditioners to solve the above problems. Summary of the Invention

[0005] The present invention aims to solve the above-mentioned technical problems, namely, to solve the problem that the roof-mounted air conditioner only uses air temperature as the control target and ignores the influence of other factors on passenger comfort, resulting in low passenger comfort in the vehicle and affecting the passenger's riding experience.

[0006] In a first aspect, the present invention provides a control method for a rooftop air conditioner; the control method for the rooftop air conditioner includes the following steps: Acquire the in-vehicle environment parameter set and the in-vehicle occupant parameter set; The overall thermal comfort index is calculated based on the in-vehicle environment parameter set and the in-vehicle occupant parameter set. The operating status of the rooftop air conditioner is adjusted according to the overall thermal comfort index.

[0007] In the preferred embodiment of the control method for the above-mentioned rooftop air conditioner, the rooftop air conditioner includes multiple air outlets, and the opening degree of the air outlets is adjustable. The control method further includes: The interior space is divided into multiple zones based on the location of the air vents; Based on the in-vehicle environment parameter set and the in-vehicle occupant parameter set, the thermal comfort index of each zone is calculated. Adjust the opening degree of the corresponding air outlet according to the thermal comfort index of the zone.

[0008] In the preferred embodiment of the above-mentioned control method for roof-mounted air conditioning, the in-vehicle environmental parameter set includes air temperature, relative humidity, air velocity, and mean radiant temperature at at least one location within each zone; and / or The parameters for occupants inside the vehicle include the total number of occupants, the individual positions of all occupants, their individual clothing thickness index, and their individual activity intensity index; and / or The operating status of the rooftop air conditioner includes cooling / heating operation mode, compressor frequency, and fan speed.

[0009] In the preferred technical solution of the above-mentioned control method for roof-mounted air conditioners, "calculating the overall thermal comfort index based on the in-vehicle environment parameter set and the in-vehicle occupant parameter set" specifically includes: The overall air temperature, overall relative humidity, overall air velocity, and overall average radiant temperature are determined by weighted averaging based on the air temperature, relative humidity, air velocity, and average radiant temperature at multiple locations inside the vehicle. The overall clothing thickness index and overall activity intensity index are determined by weighted averaging based on the total number of people, the individual clothing thickness index and the individual activity intensity index of all people in the vehicle. The overall thermal comfort index is obtained by inputting the overall air temperature, overall relative humidity, overall air velocity, overall mean radiant temperature, overall clothing thickness index, and overall activity intensity index into the prediction average voting model; and / or "Based on the in-vehicle environment parameter set and the in-vehicle occupant parameter set, the thermal comfort index of each of the aforementioned zones is calculated," specifically including: The zone air temperature, zone relative humidity, zone air flow rate, and zone average radiation temperature of each zone are determined by weighted averaging of all the air temperature, relative humidity, air flow rate, and average radiation temperature of each zone. Based on the individual locations, determine the number of personnel in each of the partitions; The individual clothing thickness index and individual activity intensity index of all personnel corresponding to each partition are used to determine the partition clothing thickness index and partition activity intensity index of each partition by weighted averaging. By inputting the zone's air temperature, relative humidity, air velocity, average radiant temperature, clothing thickness index, and activity intensity index into the predictive average voting model, the zone's thermal comfort index is obtained.

[0010] In the preferred embodiment of the above-mentioned control method for a roof-mounted air conditioner, the control method further includes: The in-vehicle environment parameter set and the in-vehicle occupant parameter set are obtained again, and the overall thermal comfort index and the thermal comfort index of each zone are calculated. Obtain a set of in-vehicle state parameters, which includes the overall thermal comfort index and the zone thermal comfort index for each of the aforementioned zones. Based on the in-vehicle state parameter set, multiple air conditioning control strategies are inferred in parallel through a preset deep reinforcement learning model, and the air conditioning control strategy that can obtain the highest reward value under the current in-vehicle state is output; and / or the air conditioning control strategy includes the cooling / heating operation mode of the roof-mounted air conditioner, the compressor frequency, the fan speed and the opening degree of each of the air outlets.

[0011] In the preferred embodiment of the above-mentioned control method for roof-mounted air conditioning, the in-vehicle state parameter set also includes the total energy consumption of the roof-mounted air conditioning. The reward function of the deep learning model is: Reward=-α*|PMVtarget-PMVcurrent|-β*Power consumption-γ*Nuncomfortable; Where Reward is the reward value; α, β, and γ are all weighting coefficients and are preset values; PMVtarget is the target overall thermal comfort index and is a preset value. PMVcurrent is the overall thermal comfort index; Power consumption refers to the total energy consumption of the rooftop air conditioner. N uncomfortable represents the number of uncomfortable zones, which is the total number of zones whose thermal comfort index is outside the preset range.

[0012] In the preferred embodiment of the above-mentioned control method for roof-mounted air conditioners, the in-vehicle state parameter set further includes a vehicle state parameter set; The control method further includes: When any parameter in the vehicle state parameter set reaches any predetermined state, the predicted in-vehicle state parameter set is obtained. Based on the predicted in-vehicle state parameter set, multiple air conditioning control strategies are inferred in parallel using a preset deep reinforcement learning model. The resulting air conditioning control strategy, which yields the highest reward value under the predicted in-vehicle state, is then used to pre-adjust the operating state of the roof-mounted air conditioner to cope with any changes in the in-vehicle state. And / or The predetermined state includes a target geographical location range, a target time range, a closed door, and an open door; and / or Preferably, the predicted in-vehicle state parameter set includes a predicted overall thermal comfort index and a predicted zone thermal comfort index for each of the zones. Vehicles communicate with the ticketing system; "Obtaining the predicted in-vehicle state parameter set" specifically includes: Based on the ticketing system, obtain the number of passengers boarding and alighting at the next stop; The thermal comfort correction coefficient is determined based on the difference between the number of passengers boarding and alighting at the next station. The predicted overall thermal comfort index is determined based on the overall thermal comfort index and the thermal comfort correction coefficient. The predicted thermal comfort index for each zone is determined based on the zone thermal comfort index and the thermal comfort correction coefficient.

[0013] In the preferred embodiment of the above-mentioned control method for a roof-mounted air conditioner, the control method further includes: The roof-mounted air conditioner records the in-vehicle state parameter set, the corresponding air conditioning control strategy, and the reward value in real time, and performs offline training on the cloud or local machine periodically to update the air conditioning control strategy of the deep reinforcement learning model, so that the air conditioning control strategy output by the deep reinforcement learning model can adapt to the operation pattern of the fixed route.

[0014] Secondly, the present invention also provides a control system for a rooftop air conditioner; the control system for the rooftop air conditioner includes a data acquisition module, a data processing module, a core control module, and an execution module; The data acquisition module is used to collect the in-vehicle environmental parameters, in-vehicle occupant parameters, vehicle status parameters, and total energy consumption of the roof-mounted air conditioner during vehicle operation, and transmit them to the data processing module and the core control module. The data processing module can determine the set of occupant parameters based on the received occupant parameters and transmit it to the core control module. The core control module can store program instructions, predict average voting models, and deep reinforcement learning models, and execute the control method for the rooftop air conditioner as described above when running the program instructions; and / or The data acquisition module includes a CAN bus interface, an energy consumption acquisition device, a vision sensor, multiple temperature sensors, a humidity sensor, and an airflow sensor. The CAN bus interface is used to transmit the vehicle's geographical location, date, time, and door opening / closing status to the core control module. The energy consumption acquisition device is used to collect the total energy consumption of the rooftop air conditioner and transmit it to the core control module; The visual sensor is used to collect in-vehicle video information and transmit it to the data processing module and the core control module. The temperature sensor is used to collect temperature information within each of the partitions and transmit it to the core control module. The humidity sensor is used to collect humidity information in each of the partitions and transmit it to the core control module. The airflow velocity sensor is used to collect airflow velocity information in each of the zones and transmit it to the core control module; and / or The data processing module includes an AI chip that integrates a neural network processing unit. The AI ​​chip can determine the set of parameters of the people in the vehicle based on the video information collected by the visual sensor and transmit it to the core control module.

[0015] In summary, the present invention has at least the following beneficial effects: 1. This invention calculates the overall thermal comfort index based on a set of in-vehicle environmental parameters and a set of in-vehicle occupant parameters, and then adjusts the operation of the roof-mounted air conditioner according to the overall thermal comfort index. Compared with the limitations of using only air temperature as a single control target, this invention collects parameters from two dimensions—the in-vehicle environmental parameters and the in-vehicle occupant parameters—to calculate the overall thermal comfort index. This overall thermal comfort index, which more accurately reflects the true thermal sensation of the human body, is used to adjust the operation of the roof-mounted air conditioner, thereby improving the thermal comfort of passengers and enhancing their travel experience. 2. The interior space is divided into multiple zones based on the location of the air vents. The thermal comfort index of each zone is calculated based on the set of interior environmental parameters and the set of occupant parameters. The opening of the corresponding air vent is adjusted according to the thermal comfort index of each zone, thereby achieving precise control of the air volume within the zone and further improving the thermal comfort of passengers. 3. Based on the in-vehicle state parameter set, a pre-set deep reinforcement learning model performs parallel reasoning for multiple air conditioning control strategies, outputting the air conditioning control strategy that obtains the highest reward value under the current in-vehicle state, ensuring good thermal comfort for most passengers. Furthermore, the deep reinforcement learning model is continuously trained and updates the air conditioning control strategy, resulting in a strategy that becomes increasingly suitable for the operational patterns of fixed routes, further improving passenger thermal comfort. Attached Figure Description

[0016] The preferred embodiments of the present invention are described below with reference to the accompanying drawings, in which: Figure 1 This is a flowchart of the main steps of the control method for a rooftop air conditioner in Embodiment 1 of the present invention; Figure 2 This is a flowchart illustrating the specific steps of the control method for a rooftop air conditioner in Embodiment 1 of the present invention; Figure 3This is a flowchart illustrating the specific steps of the control method for the roof-mounted air conditioner in Embodiment 2 of the present invention; Figure 4 This is a flowchart illustrating the specific steps of step S92 in Embodiment 2 of the present invention. Detailed Implementation

[0017] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the invention and are not intended to limit the scope of protection of the invention. Those skilled in the art can make adjustments as needed to adapt to specific applications.

[0018] It should be noted that in the description of this invention, terms such as "upper," "lower," "left," and "right," indicating directional or positional relationships, are based on the directional or positional relationships shown in the accompanying drawings. These are merely for ease of description and do not indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation; therefore, they should not be construed as limitations on this invention. Furthermore, the terms "first," "second," "third," and "fourth" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0019] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "connected," "linked," and "connected" should be interpreted broadly. For example, they can refer to a fixed connection or a detachable connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances. Example 1

[0020] To address the issue that overhead air conditioning systems only control air temperature and ignore the impact of other factors on passenger comfort, resulting in lower passenger comfort and a negative impact on the passenger travel experience.

[0021] like Figure 1-2As shown, this embodiment discloses a control method for a roof-mounted air conditioner, commonly used in large commercial vehicles such as buses, coaches, and long-distance coaches. The roof-mounted air conditioner includes a control system, refrigerant circulation piping, a compressor, a condenser, an electronic expansion valve, an evaporator, a fan, ductwork, multiple air outlets, and a motor. The compressor, fan, and multiple motors are all electrically connected to the control system, enabling the control system to control the compressor frequency, fan speed, and motor rotation angle. The compressor, condenser, electronic expansion valve, and evaporator are installed on the refrigerant circulation piping. The fan is located on one side of the evaporator, and the ductwork is connected to the fan outlet. Multiple air outlets are installed on the ductwork. The fan forces air to flow through the evaporator, cooling or heating the air, which is then delivered into the vehicle interior through the ductwork and air outlets. The air outlets are connected to the output end of the motor, allowing the motor to adjust the opening of the air outlets and change the airflow into the vehicle interior.

[0022] The rooftop air conditioner's control system includes a data acquisition module, a data processing module, and a core control module. The data acquisition module is electrically connected to both the data processing module and the core control module. The data acquisition module includes a CAN bus interface, an energy consumption sensor, a vision sensor, multiple temperature sensors, humidity sensors, and airflow sensors. The CAN bus interface collects vehicle status parameters and transmits them to the data processing module and the core control module. These vehicle status parameters include the vehicle's geographical location, date, time, and door open / closed status. The energy consumption sensor is a power meter that transmits the rooftop air conditioner's electrical power, i.e., its total energy consumption, to the core control module.

[0023] Multiple temperature, humidity, and airflow sensors are used to collect in-vehicle environmental parameters. The vehicle interior is divided into multiple zones based on the location of the air vents. Each zone contains at least one temperature, humidity, and airflow sensor. In this embodiment, the visual sensor is a wide-angle camera capable of capturing full-vehicle video; the temperature, humidity, and airflow sensors are all temperature and humidity sensors, respectively. The visual sensor collects in-vehicle video information to gather occupant parameters and uploads it to the data processing module. The temperature, humidity, and airflow sensors collect temperature information from each zone and transmit it to the core control module. The data acquisition module also includes multiple average radiation temperature measurement devices, such as spherical thermometers or surface temperature sensors, which transmit the collected temperature values ​​to the core control module.

[0024] The data processing module includes an AI chip. This AI chip determines the set of occupant parameters based on video information collected by visual sensors and date, time, and geographic location information transmitted via the CAN bus, and transmits this information to the core control module. Specifically, the AI ​​chip includes a first processing unit, a first storage unit, and a neural network processing unit. The first storage unit stores a target detection neural network model, a clothing recognition model, and a human posture estimation model. The first processing unit receives video information, date, and time information, and controls the neural network processing unit to call the target detection neural network model, outputting a matrix of passenger number and location distribution. For each detected person, the human posture estimation model is called to identify their posture and output the corresponding individual activity intensity index, such as 1 Mt for sitting and 1.2 Mt for standing. The clothing recognition model is also called, combined with date, time, and geographic location information, to output the corresponding individual clothing thickness index. Based on the date, time, and geographic location information, a base value is determined, such as 0.5 in Hainan summer and 1 in Harbin winter. The clothing recognition model identifies clothing, with 0.4 for short sleeves and 0.5 for long sleeves. Based on the recognition results, a correction value is determined, and the final individual clothing thickness index is the base value plus the correction value. Semantic segmentation can also be used to identify the skin area covered by clothing and determine the individual's clothing thickness index.

[0025] The core control module includes a processing unit and a storage unit. The storage unit stores program instructions, a prediction average voting model, and a deep reinforcement learning model. The processing unit can receive the in-vehicle environment parameter set, the in-vehicle occupant parameter set, and the vehicle state parameter set, calculate the average radiant temperature, run program instructions, execute the control method of the roof-mounted air conditioner, and call the prediction average voting model and the deep reinforcement learning model according to the control method. It can also control the compressor, fan, and motor to execute the air conditioning control strategy.

[0026] like Figure 1 As shown, the main steps of the control method for a rooftop air conditioner include: S1. Obtain the in-vehicle environment parameter set and the in-vehicle occupant parameter set; S2. Calculate the overall thermal comfort index based on the in-vehicle environmental parameter set and the in-vehicle occupant parameter set; S3. Adjust the operating status of the rooftop air conditioner according to the overall thermal comfort index.

[0027] Based on the in-vehicle environmental parameter set and the in-vehicle occupant parameter set, an overall thermal comfort index is calculated. Then, the operation of the overhead air conditioner is adjusted according to this overall thermal comfort index. Compared to the limitations of using only air temperature as a single control target, this invention collects parameter sets from two dimensions—the in-vehicle environmental parameter set and the in-vehicle occupant parameter set—to calculate the overall thermal comfort index. This overall thermal comfort index, which more accurately reflects the actual thermal sensation of the human body, is used to adjust the operation of the overhead air conditioner, thereby improving the thermal comfort of passengers and enhancing their overall travel experience.

[0028] like Figure 2 As shown, the specific steps of the control method for a rooftop air conditioner include: S1. The interior space is divided into multiple zones according to the location of the air vents. In this embodiment, each air vent corresponds to one zone.

[0029] S2. Obtain the in-vehicle environmental parameter set and the in-vehicle occupant parameter set. The in-vehicle environmental parameter set includes air temperature, relative humidity, air velocity, and mean radiant temperature at at least one location within each zone. The in-vehicle occupant parameters include the total number of occupants, the individual location of each occupant, individual clothing thickness index, and individual activity intensity index.

[0030] S3. Calculate the overall thermal comfort index based on the in-vehicle environmental parameter set and the in-vehicle occupant parameter set. Determine the overall air temperature, overall relative humidity, overall air velocity, and overall average radiant temperature using a weighted average method based on the air temperature, relative humidity, air velocity, mean radiant temperature at various locations within the vehicle, and the number of corresponding detection locations. Determine the overall clothing thickness index and overall activity intensity index using a weighted average method based on the total number of occupants, the individual clothing thickness index, and the individual activity intensity index for all occupants. Input the overall air temperature, overall relative humidity, overall air velocity, overall average radiant temperature, overall clothing thickness index, and overall activity intensity index into the predictive average voting model to obtain the overall thermal comfort index.

[0031] S4. The overall thermal comfort index is set in accordance with the cooling / heating operation mode, compressor frequency and fan speed. Based on the calculated overall thermal comfort index, the cooling / heating operation mode, compressor frequency and fan speed of the rooftop air conditioner are adjusted.

[0032] S5. Based on the in-vehicle environmental parameter set and the in-vehicle occupant parameter set, calculate the thermal comfort index for each zone. Based on all air temperatures, relative humidity, air velocity, mean radiant temperature, and the number of detection locations for each zone, determine the zone's air temperature, relative humidity, air velocity, and mean radiant temperature using a weighted average. Determine the number of occupants in each zone based on individual locations. Based on the individual clothing thickness index and activity intensity index of all occupants in each zone, determine the zone's clothing thickness index and activity intensity index using a weighted average. Input the zone's air temperature, relative humidity, air velocity, mean radiant temperature, clothing thickness index, and activity intensity index into the predictive average voting model to obtain the zone's thermal comfort index.

[0033] S6. The zoned thermal comfort index is set in relation to the corresponding air outlet opening. Based on the calculated zoned thermal comfort index, the opening of the corresponding air outlet is adjusted. First, the overall thermal comfort index is used to adjust the cooling / heating operation mode, compressor frequency, and fan speed of the rooftop air conditioner to determine the air temperature and airflow velocity. Then, the zoned thermal comfort index is used to individually adjust the opening of each air outlet and the airflow within each zone. Compared to using only air temperature as a single control target, controlling the air conditioner's operation based on overall and zoned thermal comfort is more consistent with the actual thermal sensation of the human body, resulting in better user comfort. Furthermore, due to the large interior space, controlling the airflow within each zone individually allows for precise control of the airflow within that zone, ensuring a better match between the airflow in that zone and the thermal comfort of the passengers in that zone, further improving passenger thermal comfort and enhancing the overall passenger experience.

[0034] S7. Obtain the in-vehicle environment parameter set, in-vehicle occupant parameter set, and vehicle status parameter set. The vehicle status parameter set includes the vehicle's geographical location, date, time, and door open / closed status. Calculate the overall thermal comfort index and the thermal comfort index for each zone using a predictive average voting model.

[0035] S8. Obtain the in-vehicle state parameter set, which includes the in-vehicle environmental parameter set, the in-vehicle occupant parameter set, the vehicle state parameter set, the overall thermal comfort index, and the zonal thermal comfort index for each zone. Specifically, the in-vehicle state parameter set includes air temperature, relative humidity, air velocity, mean radiant temperature, total energy consumption of the roof-mounted air conditioner, total number of occupants, individual occupant positions, individual clothing thickness index, individual activity intensity index for all occupants, vehicle geographical location, date, time, and door open / closed status.

[0036] S9. Based on the in-vehicle state parameter set, perform parallel reasoning of multiple air conditioning control strategies through a preset deep reinforcement learning model, and output the air conditioning control strategy that can obtain the highest reward value under the current in-vehicle state. The air conditioning control strategy includes cooling / heating operation mode, compressor frequency, fan speed and opening degree of each air outlet.

[0037] The reward function for a deep learning model is: Reward = -α * |PMVtarget - PMVcurrent| - β * Power consumption - γ * Nuncomfortable; where α, β, and γ are weighting coefficients, preset values, and their magnitudes indicate the degree of emphasis on overall thermal comfort, energy consumption, and the number of uncomfortable zones; larger values ​​indicate greater emphasis. PMVtarget is the target overall thermal comfort index, a preset value, usually set to 0. PMVcurrent is the overall thermal comfort index, calculated using a predictive average voting model. Power consumption is the total energy consumption of the overhead air conditioner, collected by a data acquisition module. Nuncomfortable is the number of uncomfortable zones, the sum of the number of zones whose thermal comfort indices are outside the preset range. Specifically, comfort is defined as the range between the first and second preset thermal comfort indices; thermal comfort indices outside this range are considered uncomfortable, and the number of uncomfortable zones is counted. The reward function guides the deep learning model to output the air conditioning control strategy, ensuring that the air conditioning control strategy is an efficient strategy that guarantees the comfort of most passengers, considers energy costs, and does not neglect the experience of passengers in any zone.

[0038] After the roof-mounted air conditioner is started, the state inside the vehicle changes after steps S1-S6. The state parameter set inside the vehicle is re-acquired and then input into the deep reinforcement learning model. The deep reinforcement learning model then provides the air conditioning control strategy.

[0039] The deep reinforcement learning model integrates multiple factors, including in-vehicle environmental parameters, total number of people, individual positions, individual clothing thickness index of all people in the vehicle, individual activity intensity index, vehicle geographical location, date, time, and door opening / closing status. It effectively solves the problem of uneven temperature field caused by sunlight and uneven passenger distribution in large commercial vehicles, avoids local overheating or undercooling, improves the comfort of all passengers, and achieves precise energy allocation and conservation by reducing air supply to vacant areas.

[0040] S10. Has the parameter acquisition time exceeded the preset time? If not, execute S9. If yes, return to S7. After executing the air conditioning control strategy in S9, changes in solar radiation intensity, passenger position changes, and changes in passenger clothing cause the vehicle's interior state to change dynamically in real time. Every preset time, execute S7-S9. This cycle repeats, and during vehicle operation, the deep reinforcement learning model continuously and dynamically outputs the air conditioning control strategy.

[0041] Specifically, the system acquires a set of in-vehicle state parameters. A deep reinforcement learning model then performs parallel reasoning for various air conditioning control strategies based on the in-vehicle state, outputting the air conditioning control strategy that yields the highest reward value under the current in-vehicle state. When any parameter in the in-vehicle environment parameter set, the in-vehicle occupant parameter set, or the vehicle state parameter set reaches any predetermined state (including target geographical location interval, target time interval, door closed, and door open), the air conditioning control strategy that yields the highest reward value is used to pre-adjust the operating state of the overhead air conditioner to cope with the changing in-vehicle state. This reduces drastic fluctuations in passenger thermal comfort caused by sudden changes in the in-vehicle state, maintaining thermal comfort at a stable, high level.

[0042] For example, the space inside a bus is divided into four zones. The bus is traveling to a train station, and the overall thermal comfort level is "warm." Zone 1 is "warm" with 8 people; Zone 2 is "warm" with 6 people; Zone 3 is "comfortable" with 5 people; and Zone 4 is "warm" with 5 people. The control strategy output by the deep reinforcement learning model is: cooling operation mode, compressor frequency set to 60Hz, fan speed set to 1600 rpm, and air vent openings for Zones 1 through 4 set to 80%, 75%, 65%, and 70% respectively. After the parameter acquisition time exceeds a preset time, the in-vehicle state parameter set is acquired again, and the deep reinforcement learning model outputs the control strategy again, and this cycle repeats.

[0043] When any parameter in the in-vehicle environment parameter set, in-vehicle personnel parameter set, or vehicle status parameter set reaches any predetermined state, the predetermined state includes the target geographical location interval, the target time interval, door closed, and door open. The target geographical location interval includes multiple location intervals, such as the first location interval corresponding to a distance of 1 kilometer from station A, the second location interval corresponding to a distance of 1 kilometer from station B, etc. The target time interval includes multiple time intervals, such as the first time interval corresponding to a distance of 5 minutes from station A, the second time interval being 18:00-7:00, etc., pre-set according to the travel time.

[0044] The bus continues its journey. When the vehicle reaches the target geographical location, such as 1 kilometer from the train station, a large number of passengers will board. At this time, the air conditioning control strategy is to operate in cooling mode, with the compressor frequency set to 70Hz, the fan speed at 1800 rpm, and the air outlet openings of the first to fourth zones set to 80%, 75%, 65%, and 70% respectively. By strengthening the cooling system before the large number of passengers board at the train station, and before the changes in the interior conditions, the impact on overall thermal comfort is minimal, even if the number of passengers increases after the bus arrives at the train station.

[0045] Alternatively, if the current time reaches the target time interval (set to ten minutes before sunset), where the average radiant temperature of the in-vehicle environmental parameters is zero, adjust the compressor power, fan speed, and air vent opening in advance. Or, if passenger density is highest between 8:00 and 9:00 AM daily on the route from station A to station B, increase the compressor power, fan speed, and air vent opening in advance. Alternatively, if the doors close and open, indicating passenger boarding and alighting, causing changes in the in-vehicle heat load and air turbulence, increase the fan speed after the doors close to quickly stabilize the environment.

[0046] In addition, the rooftop air conditioner records the in-vehicle state parameter set, corresponding air conditioning control strategies, and reward values ​​in real time, and periodically performs offline training in the cloud or locally to update the air conditioning control strategy of the deep reinforcement learning model. This ensures that the air conditioning control strategy output by the deep reinforcement learning model adapts to the operating patterns of fixed routes, resulting in increasingly better thermal comfort for passengers. Furthermore, the rooftop air conditioner also records the outside temperature, vehicle speed, and solar azimuth. The outside temperature is detected by an outside temperature sensor and transmitted to the core control module. Vehicle speed is transmitted to the core control module via the CAN bus, and solar azimuth is determined by vehicle location and time. During the training process of the deep reinforcement learning model, the model learns the relationship between solar azimuth, outside temperature, vehicle speed, and mean radiant temperature. During vehicle operation, the model acquires solar azimuth, outside temperature, and vehicle speed in real time and outputs the mean radiant temperature. This mean radiant temperature is then input into a predictive average voting model to obtain the overall thermal comfort index and the regional thermal comfort index. Compared to collecting parameters through a mean radiant temperature measuring device and then calculating the mean radiant temperature through the core control module, this method saves on the cost of the mean radiant temperature measuring device. Example 2

[0047] like Figure 3 and Figure 4As shown, the difference between this embodiment and Embodiment 1 is that when the vehicle is used in a scenario where advance ticket purchase is required and the ticket information can be obtained through the ticketing system, such as long-distance buses, short-distance buses, or trains, the control system of the roof-mounted air conditioner also includes a ticketing information collection module. The ticketing information collection module is connected to the ticketing system and obtains the number of passengers boarding and alighting at the next station through the ticketing information collection module and transmits it to the processing unit.

[0048] The processing unit can also determine a thermal comfort correction coefficient based on the difference between the number of passengers boarding and alighting at the next stop. Each difference corresponds to a thermal comfort correction coefficient. When the difference is greater than 0, the thermal comfort correction coefficient is greater than 1; when the difference is less than or equal to 0, the thermal comfort correction coefficient is less than or equal to 1. Preferably, the thermal comfort correction coefficient is between 0.85 and 1.15. The processing unit can also determine a predicted overall thermal comfort index based on the overall thermal comfort index and the thermal comfort correction coefficient, and determine a predicted zone thermal comfort index for each zone based on the zone thermal comfort index and the thermal comfort correction coefficient. Specifically, the predicted overall thermal comfort index = overall thermal comfort index * thermal comfort correction coefficient. When the difference is greater than 0, it means that the total number of people in the vehicle has increased, and the overall thermal comfort index will increase; conversely, when the difference is less than or equal to 0, the overall thermal comfort index will decrease. The predicted zone thermal comfort index for each zone = zone thermal comfort index * thermal comfort correction coefficient. When the overall number of people inside the vehicle changes, the number of people in each compartment is likely to change synchronously, so the same thermal comfort correction coefficient is used for both the overall system and each compartment.

[0049] like Figure 3 As shown, the specific steps of the control method for a rooftop air conditioner include: S1. Divide the interior space into multiple zones based on the location of the air vents.

[0050] S2. Obtain the in-vehicle environment parameter set and the in-vehicle personnel parameter set.

[0051] S3. Calculate the overall thermal comfort index based on the in-vehicle environmental parameter set and the in-vehicle occupant parameter set. The specific method is the same as in Example 1, and will not be repeated here.

[0052] S4. The overall thermal comfort index is set in accordance with the cooling / heating operation mode, compressor frequency and fan speed. Based on the calculated overall thermal comfort index, the cooling / heating operation mode, compressor frequency and fan speed of the rooftop air conditioner are adjusted.

[0053] S5. Based on the in-vehicle environmental parameter set and the in-vehicle occupant parameter set, calculate the thermal comfort index of each zone. The specific method is the same as in Example 1, and will not be repeated here.

[0054] S6. The zoned thermal comfort index is set in relation to the corresponding air outlet opening. Based on the calculated zoned thermal comfort index, the opening of the corresponding air outlet is adjusted. First, the overall thermal comfort index is used to adjust the cooling / heating operation mode, compressor frequency, and fan speed of the rooftop air conditioner to determine the air temperature and airflow velocity. Then, the zoned thermal comfort index is used to individually adjust the opening of each air outlet and the airflow within each zone. Compared to using only air temperature as a single control target, controlling the air conditioner's operation based on overall and zoned thermal comfort is more consistent with the actual thermal sensation of the human body, resulting in better user comfort. Furthermore, due to the large interior space, controlling the airflow within each zone individually allows for precise control of the airflow within that zone, ensuring a better match between the airflow in that zone and the thermal comfort of the passengers in that zone, further improving passenger thermal comfort and enhancing the overall passenger experience.

[0055] S7. Obtain the in-vehicle environment parameter set, in-vehicle personnel parameter set, and vehicle status parameter set, and calculate the overall thermal comfort index and the thermal comfort index of each zone through the prediction average voting model.

[0056] S8. Obtain the in-vehicle state parameter set, which includes the in-vehicle environment parameter set, the in-vehicle occupant parameter set, the vehicle state parameter set, the overall thermal comfort index, and the zonal thermal comfort index of each zone.

[0057] S9. Determine whether any parameter in the vehicle state parameter set has reached any predetermined state. The predetermined states include the target geographical location interval, the target time interval, the door being closed, and the door being open. If not, proceed to S91; if yes, proceed to S92.

[0058] S91. Based on the in-vehicle state parameter set, perform parallel reasoning of multiple air conditioning control strategies through a preset deep reinforcement learning model, and output the air conditioning control strategy that can obtain the highest reward value under the current in-vehicle state. The air conditioning control strategy includes cooling / heating operation mode, compressor frequency, fan speed and opening degree of each air outlet.

[0059] S92. Obtain the predicted in-vehicle state parameter set. Based on the predicted in-vehicle state parameter set, perform parallel reasoning of multiple air conditioning control strategies through a preset deep reinforcement learning model. Output the air conditioning control strategy that can obtain the highest reward value under the predicted in-vehicle state. The air conditioning control strategy that obtains the highest reward value is used to pre-adjust the operating state of the roof air conditioner to cope with the changing in-vehicle state.

[0060] The reward function for a deep learning model is: Reward = -α * |PMVtarget - PMVcurrent| - β * Power consumption - γ * Nuncomfortable; where α, β, and γ are weighting coefficients, preset values, and their magnitudes indicate the degree of emphasis on overall thermal comfort, energy consumption, and the number of uncomfortable zones; larger values ​​indicate greater emphasis. PMVtarget is the target overall thermal comfort index, a preset value, usually set to 0. PMVcurrent is the overall thermal comfort index, calculated using a predictive average voting model. Power consumption is the total energy consumption of the overhead air conditioner, collected by a data acquisition module. Nuncomfortable is the number of uncomfortable zones, the sum of the number of zones whose thermal comfort indices are outside the preset range. Specifically, comfort is defined as the range between the first and second preset thermal comfort indices; thermal comfort indices outside this range are considered uncomfortable, and the number of uncomfortable zones is counted. The reward function guides the deep learning model to output the air conditioning control strategy, ensuring that the air conditioning control strategy is an efficient strategy that guarantees the comfort of most passengers, considers energy costs, and does not neglect the experience of passengers in any zone.

[0061] The deep reinforcement learning model integrates multiple factors, including in-vehicle environmental parameters, total number of people, individual positions, individual clothing thickness index of all people in the vehicle, individual activity intensity index, vehicle geographical location, date, time, and door opening / closing status. It effectively solves the problem of uneven temperature field caused by sunlight and uneven passenger distribution in large commercial vehicles, avoids local overheating or undercooling, improves the comfort of all passengers, and achieves precise energy allocation and conservation by reducing air supply to vacant areas.

[0062] When any parameter in the vehicle state parameter set fails to reach any predetermined state, the deep reinforcement learning model outputs the air conditioning control strategy that yields the highest reward value under the current in-vehicle state, based on the in-vehicle state parameter set. This air conditioning control strategy represents the optimal immediate response to the current in-vehicle state, ensuring good thermal comfort for most passengers in each zone and low overall energy consumption. For example, consider a short-distance bus divided into four zones. The bus travels between two stations, and the overall thermal comfort level is slightly warm. Zone 1 is warm with 8 passengers, Zone 2 is slightly warm with 6 passengers, Zone 3 is comfortable with 5 passengers, and Zone 4 is slightly warm with 5 passengers. The control strategy output by the deep reinforcement learning model is: cooling operation mode, compressor frequency set to 60Hz, fan speed 1600r / min, and air outlet openings for Zones 1 through 4 set to 80%, 75%, 65%, and 70%, respectively.

[0063] When any parameter in the vehicle state parameter set fails to reach any predetermined state, a predicted in-vehicle state parameter set is obtained. Based on this set, a deep reinforcement learning model outputs the air conditioning control strategy that yields the highest reward value under the predicted in-vehicle state. This control strategy pre-adjusts the operation of the overhead air conditioner to cope with the changing in-vehicle state. By predicting the in-vehicle state and pre-adjusting the overhead air conditioner, drastic fluctuations in passenger thermal comfort caused by sudden changes in the in-vehicle state can be reduced, maintaining thermal comfort at a stable, high level. For example, when a short-distance bus reaches its target geographical location, such as 1 kilometer from the next stop, a predicted in-vehicle state parameter set is obtained. Based on this set, a deep reinforcement learning model outputs an air conditioning control strategy to pre-adjust the overhead air conditioner's operation to cope with the changing in-vehicle state.

[0064] S10. Has the parameter acquisition time exceeded the preset time? If not, execute S9. If yes, return to S7. After executing the air conditioning control strategy in S9, changes in solar radiation intensity, passenger position changes, and changes in passenger clothing cause the vehicle's interior state to change dynamically in real time. Every preset time, execute S7-S9. This cycle repeats. During vehicle operation, the deep reinforcement learning model continuously and dynamically outputs the air conditioning control strategy, adjusting the cooling / heating operation mode of the roof-mounted air conditioner, compressor frequency, fan speed, and the opening of each air outlet to ensure good thermal comfort for most passengers.

[0065] like Figure 4 As shown, the specific steps of step S92 include: S921. Obtain the number of passengers boarding and alighting at the next stop based on the ticketing system; S922. Determine the thermal comfort correction coefficient based on the difference between the number of passengers boarding and alighting at the next station. S923. Determine the predicted overall thermal comfort index based on the overall thermal comfort index and the thermal comfort correction coefficient; determine the predicted thermal comfort index of each zone based on the zone thermal comfort index and the thermal comfort correction coefficient. S924. Obtain the predicted in-vehicle state parameter set, which includes the predicted overall thermal comfort index and the predicted thermal comfort index of each zone. S925. Based on the predicted in-vehicle state parameter set, perform parallel reasoning of multiple air conditioning control strategies through a preset deep reinforcement learning model, and output the air conditioning control strategy that can obtain the highest reward value under the predicted in-vehicle state. At this time, the air conditioning control strategy that obtains the highest reward value is used to pre-adjust the operating state of the roof air conditioner to cope with the changing in-vehicle state.

[0066] Because the deep reinforcement learning model outputs air conditioning control strategies at preset intervals, the overall number of passengers significantly influences the predicted overall thermal comfort index and the predicted thermal comfort indices for each zone. Based on the ticketing system, the number of passengers boarding and alighting at the next stop is accurately obtained. Then, based on the difference in passenger numbers, a thermal comfort correction coefficient is determined, which in turn determines the predicted overall thermal comfort index and the predicted thermal comfort indices for each zone. The deep reinforcement learning model outputs air conditioning control strategies based on these indices. This results in a more precise adaptation of the predicted in-vehicle state—the state of the vehicle upon arrival at the next stop—maintaining a stable and high level of thermal comfort for passengers. During the journey from the predetermined state to the stop, currently boarding passengers do not experience sudden changes in temperature, while newly boarded passengers experience immediate and uniform comfort, thus enhancing the overall passenger experience.

[0067] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A control method of a ceiling air conditioner, characterized by, The control method comprises the following steps: acquiring a set of in-vehicle environment parameters and a set of in-vehicle personnel parameters; calculating an overall thermal comfort index according to the set of in-vehicle environment parameters and the set of in-vehicle personnel parameters; adjusting the operating state of the roof air conditioner according to the overall thermal comfort index.

2. The control method of the ceiling air conditioner according to claim 1, characterized by, The roof air conditioner comprises a plurality of air outlets, the opening degree of the air outlets being adjustable; The control method further comprises: dividing the in-vehicle space into a plurality of sub-zones according to the positions of the air outlets; calculating a sub-zone thermal comfort index of each of the sub-zones according to the set of in-vehicle environment parameters and the set of in-vehicle personnel parameters; adjusting the opening degree of the corresponding air outlet according to the sub-zone thermal comfort index.

3. The control method of the ceiling air conditioner according to claim 2, characterized by, The set of in-vehicle environment parameters comprises the air temperature, relative humidity, air flow rate and mean radiant temperature at at least one position in each of the sub-zones; and / or The in-vehicle personnel parameters comprise the overall number of personnel, the individual positions of all the personnel in the vehicle, individual clothing thickness index and individual activity intensity index; and / or The operating state of the roof air conditioner comprises the cooling / heating operating mode, compressor frequency and fan speed.

4. The control method of the ceiling air conditioner according to claim 3, characterized by, "Calculating an overall thermal comfort index according to the set of in-vehicle environment parameters and the set of in-vehicle personnel parameters" specifically comprises: determining the overall air temperature, overall relative humidity, overall air flow rate and overall mean radiant temperature by weighted average according to the air temperature, relative humidity, air flow rate and mean radiant temperature at a plurality of positions in the vehicle; determining the overall clothing thickness index and overall activity intensity index by weighted average according to the overall number of personnel, individual clothing thickness index and individual activity intensity index of all the personnel in the vehicle; inputting the overall air temperature, overall relative humidity, overall air flow rate, overall mean radiant temperature, overall clothing thickness index and overall activity intensity index into a predicted average voting model to obtain the overall thermal comfort index; and / or "Calculating a sub-zone thermal comfort index of each of the sub-zones according to the set of in-vehicle environment parameters and the set of in-vehicle personnel parameters" specifically comprises: determining the sub-zone air temperature, sub-zone relative humidity, sub-zone air flow rate and sub-zone mean radiant temperature of each of the sub-zones by weighted average according to all the air temperature, relative humidity, air flow rate and mean radiant temperature corresponding to each of the sub-zones; determining the sub-zone number of personnel of each of the sub-zones according to the individual positions; determining the sub-zone clothing thickness index and sub-zone activity intensity index of each of the sub-zones by weighted average according to the individual clothing thickness index and individual activity intensity index of all the personnel corresponding to each of the sub-zones; inputting the sub-zone air temperature, sub-zone relative humidity, sub-zone air flow rate, sub-zone mean radiant temperature, sub-zone clothing thickness index and sub-zone activity intensity index into a predicted average voting model to obtain the sub-zone thermal comfort index.

5. The control method of the ceiling air conditioner according to claim 4, characterized by, The control method further comprises: acquiring again the set of in-vehicle environment parameters and the set of in-vehicle personnel parameters, and calculating the overall thermal comfort index and the sub-zone thermal comfort index of each of the sub-zones; acquire an in-vehicle state parameter set, the in-vehicle state parameter set comprising an overall thermal comfort index and a zonal thermal comfort index of each of the zones; inference, output an air conditioning control strategy capable of obtaining the highest reward value under the current in-vehicle state; and / or the air conditioning control strategy comprises a refrigeration / heating operation mode of the roof-mounted air conditioner, a compressor frequency, a fan speed, and an opening degree of each of the air outlets.

6. The control method of the ceiling air conditioner according to claim 5, characterized by, the in-vehicle state parameter set further comprises total energy consumption of the roof-mounted air conditioner; the reward function of the deep learning model is: Reward = -α * |PMVtarget-PMVcurrent| - β * Power consumption - γ * N uncomfortable; wherein, Reward is a reward value; α, β, and γ are weight coefficients, which are preset values; PMVtarget is a target overall thermal comfort index, which is a preset value; PMVcurrent is an overall thermal comfort index; Power consumption is total energy consumption of the roof-mounted air conditioner; N uncomfortable is the number of uncomfortable zones, which is the sum of the number of zones whose zonal thermal comfort index is not within a preset range.

7. The control method of the ceiling air conditioner according to claim 5, characterized by, the in-vehicle state parameter set further comprises a vehicle state parameter set; the control method further comprises: when any one of the parameters in the vehicle state parameter set reaches any one of the predetermined states, inference, according to the in-vehicle state parameter set, is performed on multiple air conditioning control strategies in parallel through a preset deep reinforcement learning model, and an air conditioning control strategy capable of obtaining the highest reward value under the current in-vehicle state is output; at this time, the air conditioning control strategy capable of obtaining the highest reward value is used to adjust the operating state of the roof-mounted air conditioner in advance to cope with the in-vehicle state that will change; and / or the vehicle state parameter set comprises a vehicle geographic location, a date, a time, and an opening / closing state of a vehicle door; and / or the predetermined states comprise a target geographic location interval, a date, a time, and an opening / closing state of a vehicle door.

8. The control method of the ceiling air conditioner according to claim 5, characterized by, the in-vehicle state parameter set further comprises a vehicle state parameter set; the control method further comprises: when any one of the parameters in the vehicle state parameter set reaches any one of the predetermined states, acquire a predicted in-vehicle state parameter set; inference, according to the predicted in-vehicle state parameter set, is performed on multiple air conditioning control strategies in parallel through a preset deep reinforcement learning model, and an air conditioning control strategy capable of obtaining the highest reward value under the predicted in-vehicle state is output; at this time, the air conditioning control strategy capable of obtaining the highest reward value is used to adjust the operating state of the roof-mounted air conditioner in advance to cope with the in-vehicle state that will change; and / or the predetermined states comprise a target geographic location interval, a target time interval, a closed vehicle door, and an open vehicle door; and / or Preferably, the predicted in-vehicle state parameter set comprises a predicted overall thermal comfort index and a predicted zonal thermal comfort index of each of the zones. the vehicle is in communication connection with a ticketing system; "acquiring a predicted in-vehicle state parameter set" specifically comprises: According to the ticket system, the number of passengers boarding and the number of passengers alighting at the next station are obtained; According to the difference between the number of passengers boarding and the number of passengers alighting at the next station, a thermal comfort correction coefficient is determined; According to the overall thermal comfort index and the thermal comfort correction coefficient, a predicted overall thermal comfort index is determined; According to the partition thermal comfort index and the thermal comfort correction coefficient, a predicted partition thermal comfort index of each partition is determined.

9. The control method of the ceiling air conditioner according to claim 5, characterized by, The control method further comprises: The roof-mounted air conditioner records the set of in-vehicle state parameters, the corresponding air conditioner control strategy and the reward value in real time, and periodically performs offline training in the cloud or locally to update the air conditioner control strategy of the deep reinforcement learning model, so that the air conditioner control strategy output by the deep reinforcement learning model adapts to the operation rules of fixed lines.

10. A control system of a ceiling air conditioner, characterized by, The control system comprises a data acquisition module, a data processing module, a core control module and an execution module; The data acquisition module is configured to acquire the in-vehicle environmental parameters, the in-vehicle personnel parameters, the vehicle state parameters and the total energy consumption of the roof-mounted air conditioner during the operation of the vehicle, and transmit them to the data processing module and the core control module; The data processing module is configured to determine the set of in-vehicle personnel parameters based on the received in-vehicle personnel parameters, and transmit them to the core control module; The core control module is configured to store program instructions, a prediction average voting model and a deep reinforcement learning model, and execute the control method of the roof-mounted air conditioner according to any one of claims 1-9 when the program instructions are executed; and / or the data acquisition module comprises a CAN bus interface, an energy consumption acquisition unit, a visual sensing unit, a plurality of temperature sensing units, a humidity sensing unit and an air flow rate sensing unit, the CAN bus interface being configured to transmit the geographical position of the vehicle, the date, the time and the opening and closing state of the vehicle door to the core control module; The energy consumption acquisition unit is configured to acquire the total energy consumption of the roof-mounted air conditioner, and transmit it to the core control module; The visual sensing unit is configured to collect in-vehicle video information and transmit it to the data processing module and the core control module; The temperature sensing unit is configured to collect temperature information in each partition, and transmit it to the core control module; The humidity sensing unit is configured to collect humidity information in each partition, and transmit it to the core control module; The air flow rate sensing unit is configured to collect air flow rate information in each partition, and transmit it to the core control module; and / or The data processing module comprises an AI chip integrated with a neural network processing unit, which is configured to determine the set of in-vehicle personnel parameters based on the video information collected by the visual sensing unit, and transmit them to the core control module.

Images