Multi-temperature-zone refrigeration truck partition position intelligent adjusting method and device

Abstract

The application relates to the technical field of cold chain logistics, in particular to a multi-temperature-zone refrigerated vehicle partition position intelligent adjusting method and device, which comprises the following steps: before loading goods, inputting the thermal physical parameters and basic operation parameters of the goods, and synchronously constructing a three-dimensional digital model of a refrigerated vehicle compartment; after loading the goods, collecting real-time data of the whole scene inside and outside the compartment through a multi-source sensing network, and performing multi-physical field coupling and fusion processing; according to the input parameters, the three-dimensional digital model of the refrigerated vehicle compartment and the multi-physical field coupling and fusion data, solving the initial adjusting position of the partition through a dynamic thermal load balance algorithm; constructing a multi-objective optimization function, taking the initial adjusting position of the partition as a basic solution, performing global multi-objective optimization through a BP neural network and a genetic algorithm fusion model, and solving optimal parameters of the partition adjustment; and driving the partition to slide to a target position along a slide rail through a servo motor according to the optimal parameters of the partition adjustment. The application is helpful for realizing dynamic and accurate adjustment of the partition position.

Description

Technical Field

[0001] This application relates to the field of cold chain logistics technology, and in particular to a method and device for intelligent adjustment of the position of partitions in a multi-temperature zone refrigerated truck. Background Technology

[0002] In the cold chain transportation industry, multi-temperature zone refrigerated trucks are core equipment for transporting goods with different temperature requirements in the same vehicle. They use partitions to divide the interior space of the truck into independent temperature zones to meet the temperature-controlled storage needs of different types of goods such as fruits, vegetables, fresh produce, and pharmaceuticals. Traditional refrigerated truck partitions are mostly fixed structures or manually adjustable, which have several technical drawbacks: First, fixed partitions cannot flexibly adjust the temperature zone space according to changes in cargo load and heat load, easily leading to low utilization of the truck space, wasted cooling capacity, and significantly increased energy consumption of the refrigeration system. Second, manual adjustment of partition positions relies on experience, resulting in low adjustment accuracy and slow response speed. This cannot adapt to sudden changes in ambient temperature and dynamic changes in cargo heat load during transportation, easily leading to excessive thermal interference and temperature deviations between temperature zones, ultimately causing spoilage and loss of goods.

[0003] Therefore, as the cold chain transportation industry continues to demand higher requirements for temperature control accuracy, transportation efficiency, and energy conservation, there is an urgent need for a method that can dynamically and precisely adjust the position of the partitions to solve the aforementioned technical problems of traditional adjustment methods. Summary of the Invention

[0004] Therefore, it is necessary to provide a method and device for intelligent adjustment of the partition position of a multi-temperature zone refrigerated truck, which can achieve dynamic and precise adjustment of the partition position, in response to the above-mentioned technical problems.

[0005] In a first aspect, this application provides a method for intelligent adjustment of the position of the partition in a multi-temperature zone refrigerated truck, the method comprising: Before loading cargo, input the cargo's thermal properties and basic operating parameters, and simultaneously construct a three-dimensional digital model of the refrigerated truck compartment. After the cargo is loaded, real-time data of the entire scene inside and outside the carriage is collected through a multi-source sensing network, and the collected real-time data is processed by multi-physics coupling and fusion. Based on the input parameters, the three-dimensional digital model of the refrigerated truck compartment, and the multi-physics field coupled fusion data, the initial adjustment position of the partition is obtained by solving the dynamic heat load balance algorithm. A multi-objective optimization function is constructed, and the initial adjustment position of the partition is used as the basic solution. A global multi-objective optimization is performed through a fusion model of BP neural network and genetic algorithm to solve and output the optimal parameters for partition adjustment, so as to perform secondary global optimization on the partition position adjustment result. The optimal parameters are adjusted according to the partition, and the partition is driven by a servo motor to slide along the slide rail to the target position.

[0006] In one embodiment, the step of determining the initial adjustment position of the partition using a dynamic heat load balancing algorithm based on the input parameters, the three-dimensional digital model of the refrigerated truck compartment, and multiphysics coupled fusion data includes: Each temperature zone to be divided in the refrigerated truck compartment is defined as an independent heat load calculation node, and the real-time heat load value of each heat load calculation node is quantitatively calculated based on the input parameters and multi-physics field coupled fusion data. Based on the three-dimensional digital model of the refrigerated truck compartment, a matching relationship model between the spatial distribution of cold energy and the real-time heat load values ​​of each heat load calculation node is constructed. Based on the real-time heat load value and the matching relationship model, the spatial adjustment coefficient of each heat load calculation node is calculated by the dynamic heat load balance algorithm, wherein the spatial adjustment coefficient characterizes the spatial adjustment range and adjustment direction of the temperature zone corresponding to the node. Based on the space adjustment coefficient and the physical space constraints of the pre-set refrigerated truck compartment, the initial adjustment position of the partition is obtained.

[0007] In one embodiment, the step of calculating the spatial adjustment coefficient of each heat load calculation node using a dynamic heat load balancing algorithm based on the real-time heat load value and the matching relationship model includes: Based on the real-time heat load value and the cooling supply of the refrigerated truck compartment, the heat load supply-demand difference of each heat load calculation node is calculated. Based on the matching relationship model, the heat load supply and demand difference is associated and matched with the spatial distribution of cooling capacity in the carriage to locate the cooling capacity spatial allocation area that is compatible with each heat load calculation node. Substituting the heat load supply-demand difference and the cooling capacity spatial allocation area into the dynamic heat load balance algorithm, the spatial adjustment coefficient of each heat load calculation node is obtained.

[0008] In one embodiment, the step of associating and matching the heat load supply-demand difference with the spatial distribution of cooling capacity in the carriage based on the matching relationship model, and locating the cooling capacity spatial allocation area that is compatible with each heat load calculation node, includes: Based on the matching relationship model, the core features of the spatial distribution of cold energy in the refrigerated truck compartment are extracted and a cold energy spatial feature database is established. The core features include cold energy density distribution, spatial connectivity and cold energy adjustment redundancy in each region. Based on the positive and negative attributes and quantified values ​​of the heat load supply and demand difference, determine the type and magnitude of cooling capacity allocation requirements corresponding to each heat load calculation node; Establish cooling capacity allocation association rules, and compare and filter the cooling capacity allocation requirements of each heat load calculation node with the cooling capacity spatial feature database according to the cooling capacity allocation association rules. Eliminate areas that exceed the physical space reach range and whose cooling capacity characteristics do not match, and finally locate the cooling capacity spatial allocation area that is compatible with each heat load calculation node.

[0009] In one embodiment, before substituting the heat load supply-demand difference and the cooling capacity spatial allocation area into the dynamic heat load balance algorithm to obtain the spatial adjustment coefficient of each heat load calculation node, the method further includes: To address temperature disturbances caused by door opening for loading and unloading and sudden changes in ambient temperature, thermal inertia compensation technology and machine learning prediction models are integrated to correct the difference between heat load supply and demand in real time.

[0010] In one embodiment, the construction of a multi-objective optimization function, using the initial adjustment position of the partition as the base solution, and performing global multi-objective optimization through a fusion model of BP neural network and genetic algorithm to solve for and output the optimal parameters for partition adjustment, thereby performing secondary global optimization on the partition position adjustment result, includes: With the optimization objectives of minimizing temperature deviation in the temperature zone, maximizing the utilization of the cabin space, and minimizing the energy consumption of the refrigeration system, a multi-objective optimization function is constructed by configuring dynamically adjustable weight coefficients for each optimization objective. Using the initial adjustment position of the partition as the basic solution, and combining the physical space constraints of the refrigerated truck compartment and the temperature zone isolation requirements, the optimization range is determined, and a solution space for multi-objective optimization is constructed. The optimization effect of different partition positions is predicted by a trained BP neural network prediction model. The multi-objective optimization function is used as the fitness function. With the prediction result of the BP neural network as a reference, a genetic algorithm is used to perform global multi-objective optimization in the solution space and iteratively select the optimal solution. The effectiveness of the optimal solution is verified, and the optimal parameters for the baffle adjustment are finally solved and output, thus completing the secondary global optimization of the baffle position adjustment result.

[0011] In one embodiment, the step of predicting the optimization effect of different partition positions using a trained BP neural network prediction model, and using the multi-objective optimization function as the fitness function, with the prediction results of the BP neural network as a reference, and performing global multi-objective optimization in the solution space using a genetic algorithm to iteratively select the optimal solution includes: Based on the solution space, the partition position parameters are encoded using a real-time encoding method to generate the initial population for the genetic algorithm. The partition position parameters corresponding to each individual in the initial population are input into the trained BP neural network prediction model to obtain the predicted value of the optimization effect corresponding to each partition position. The predicted optimization effect value is substituted into the multi-objective optimization function to calculate the fitness value of each individual, which serves as the basis for judging the quality of individuals in the genetic algorithm. Genetic operations of selection, crossover, and mutation are performed sequentially on the initial population to retain individuals with better fitness values, generating a new generation of population. The genetic operations are repeated until a preset iteration termination condition is met, and the best individual retained during the iteration process is taken as the optimal solution for global multi-objective optimization.

[0012] In one embodiment, the step of collecting real-time data of the entire scene inside and outside the carriage through a multi-source sensing network and performing multi-physics coupling and fusion processing on the collected real-time data includes: By using a multi-source sensing network, temperature and humidity data inside the refrigerated truck compartment, cargo heat load data, equipment operating status data, environmental conditions data outside the compartment, and vehicle driving status data are collected to form a full-dimensional raw sensing dataset. The original full-dimensional perception dataset is preprocessed by filtering, denoising, and normalization in sequence to remove abnormal data caused by sensor failure and vehicle bumps, and standardized perception data is obtained. Based on the coupling relationship between temperature field, airflow field, cargo thermal characteristic field and environmental condition field, multi-physics field coupling fusion processing is carried out on the standardized sensing data to extract the core correlation feature parameters of each physical field and form multi-physics field coupling fusion data.

[0013] Secondly, this application also provides an intelligent adjustment device for the position of the partition of a multi-temperature zone refrigerated truck. The device includes: The parameter input module is used to input the thermal properties and basic operating parameters of the cargo before loading, and simultaneously build a three-dimensional digital model of the refrigerated truck compartment. The fusion processing module is used to collect real-time data of the entire scene inside and outside the carriage through a multi-source sensing network after the cargo is loaded, and to perform multi-physics field coupling fusion processing on the collected real-time data. The initial position solution module is used to solve for the initial adjustment position of the partition by using a dynamic heat load balance algorithm based on the input parameters, the three-dimensional digital model of the refrigerated truck compartment and the multi-physics field coupled fusion data. The optimal parameter output module is used to construct a multi-objective optimization function, and with the initial adjustment position of the partition as the basic solution, it performs global multi-objective optimization through a fusion model of BP neural network and genetic algorithm, solves and outputs the optimal parameters for partition adjustment, so as to perform secondary global optimization on the partition position adjustment result; The motor drive module is used to adjust the optimal parameters according to the partition and drive the partition to slide along the slide rail to the target position via a servo motor.

[0014] In summary, this application includes the following beneficial technical effects: Before loading, the thermal properties and basic operating parameters of the cargo are input, laying a precise and comprehensive data benchmark for subsequent intelligent adjustment throughout the entire process. This avoids adjustment deviations caused by missing parameters. The simultaneously constructed 3D digital model of the cargo compartment accurately maps the physical spatial characteristics of the compartment, providing a realistic spatial model foundation for the algorithmic calculation of partition positions. After loading, real-time data from the entire interior and exterior of the compartment is collected through a multi-source sensing network. This achieves comprehensive, blind-spot-free real-time perception of the cargo's thermal load, equipment status, external environmental conditions, and vehicle driving status. Simultaneously, through multi-physics field coupling and fusion processing, abnormal data is eliminated, data redundancy and interference are removed, and core correlation features of each physical field are extracted to form fused data that accurately reflects the actual operating conditions. Dynamic thermal load leveling... The initial adjustment position of the partition is obtained by the equilibrium algorithm, which achieves a precise match between the spatial distribution of cold energy and the real-time heat load of the goods. This avoids the problem of inefficient subsequent optimization caused by excessive initial position deviation and ensures the basic accuracy of partition position adjustment. A multi-objective optimization function is constructed. Based on the initial adjustment position of the partition, a global multi-objective optimization is performed through a fusion model of BP neural network and genetic algorithm to output the optimal parameters for partition adjustment. This allows for fine optimization of the initial adjustment position and the selection of the optimal adjustment parameters that take into account the needs of all working conditions. This enables the partition position adjustment to better fit the dynamic working conditions of cold chain transportation, achieving dynamic adaptation and precise optimization of the adjustment strategy. Based on the optimal parameters for partition adjustment, the partition is driven to slide by a servo motor, achieving smooth and precise movement of the partition. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating a method for intelligent adjustment of the partition position of a multi-temperature zone refrigerated truck in one embodiment. Figure 2 This is a flowchart illustrating the intelligent adjustment method for the partition position of a multi-temperature zone refrigerated truck in another embodiment. Figure 3 This is a structural block diagram of a multi-temperature zone refrigerated truck partition position intelligent adjustment device in one embodiment. Detailed Implementation

[0016] This invention provides a method and device for intelligent adjustment of the partition position of a multi-temperature zone refrigerated truck.

[0017] The embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the accompanying drawings and embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0018] In the description of the embodiments disclosed in this invention, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0019] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the intelligent adjustment method for the position of the partition of a multi-temperature zone refrigerated truck according to the present invention includes: S100: Before loading cargo, input the cargo's thermal properties and basic operating parameters, and simultaneously construct a three-dimensional digital model of the refrigerated truck compartment.

[0020] Specifically, before loading, a standardized parameter input port is set up through the vehicle's intelligent terminal, supporting multiple modes of input, including manual input and batch import, for cargo thermophysical parameters and basic operating parameters. The cargo thermophysical parameters cover a full range of characteristics, such as the cargo's preset storage temperature and humidity range, heat sensitivity coefficient, specific heat capacity, actual loading volume, and volume distribution, accurately depicting the temperature control requirements of different types of cargo. The basic operating parameters integrate key equipment operation indicators such as the refrigeration system's rated power, air supply mode, core threshold for temperature zone isolation, and partition adjustment accuracy requirements. After all parameters are entered, they are automatically verified and categorized for storage, forming a structured parameter database. This avoids the adjustment deviation problems caused by incomplete parameter input and lack of verification in traditional methods. Simultaneously, relying on 3D laser scanning and digital modeling technology, a full-scale inspection of the refrigerated truck compartment is conducted. Dimensional spatial modeling constructs a 1:1 precise 3D digital model of the refrigerated truck compartment that matches the actual physical structure of the compartment. This model is not simply a replication of the compartment size, but integrates core spatial attributes such as the sliding rail layout trajectory, the limit boundary of partition movement, the natural distribution characteristics of cold energy within the compartment, and the potential temperature zone division areas. It can also achieve linkage mapping with the parameter database, digitally integrating cargo loading distribution, equipment operating parameters, and the physical space of the compartment. This breaks through the technical limitations of traditional refrigerated trucks that lack digital spatial model support and rely solely on manual experience to determine temperature zone divisions. It lays a digital foundation for the algorithmic and precise calculation of subsequent partition positions, enabling partition position adjustments to conform to the personalized temperature control needs of the cargo and the actual spatial characteristics of the compartment from the source, achieving a digital, standardized, and precise upgrade in the preparation for adjustment.

[0021] After the cargo is loaded, the S200 collects real-time data of the entire scene inside and outside the carriage through a multi-source sensing network, and performs multi-physics field coupling and fusion processing on the collected real-time data.

[0022] Specifically, after the cargo is loaded, a multi-source sensing network deployed inside and outside the vehicle is activated. This network integrates high-precision sensors for temperature, humidity, displacement, pressure, flow, and other types of data to achieve real-time, comprehensive data collection of cargo thermal load, equipment operating status, vehicle environmental conditions, and vehicle driving status. It can comprehensively capture various operating condition data during cold chain transportation. At the same time, the collected raw data undergoes multi-physics field coupling and fusion processing. Based on the inherent coupling relationship between temperature field, airflow field, cargo thermal characteristic field, and environmental condition field, abnormal data caused by sensor failure and vehicle vibration is eliminated, and the core correlation features of each physical field are extracted to form fused data that accurately reflects the actual operating conditions. This provides high-quality and high-reliability input for subsequent algorithmic decisions and prevents partition adjustment errors caused by single data dimensions or data distortion.

[0023] S300 uses a dynamic heat load balancing algorithm to solve for the initial adjustment position of the partition based on the input parameters, the three-dimensional digital model of the refrigerated truck compartment, and the multi-physics field coupled fusion data.

[0024] Specifically, using the core parameters, the 3D digital model of the refrigerated truck compartment, and multi-physics field coupled fusion data as comprehensive inputs, the initial adjustment position of the partition is solved through a dynamic heat load balancing algorithm. This algorithm breaks through the traditional mindset of static cold energy distribution in refrigerated trucks. It can quantitatively calculate the heat load supply and demand relationship of each temperature zone to be divided in the compartment according to real-time operating conditions, realizing the dynamic matching of cold energy spatial distribution and real-time heat load of goods. At the same time, with the core objective of minimizing thermal interference between temperature zones, combined with the spatial constraints of the 3D digital model of the compartment, the initial adjustment position of the partition is scientifically calculated. This makes the determination of the initial position of the partition more in line with the dynamic heat load changes of cold chain transportation, laying a scientific and reasonable foundation for subsequent refined optimization and avoiding the problem of inefficient subsequent optimization caused by excessive initial position deviation.

[0025] S400 constructs a multi-objective optimization function, and uses the initial adjustment position of the partition as the basic solution. It performs global multi-objective optimization through a fusion model of BP neural network and genetic algorithm, solves and outputs the optimal parameters for partition adjustment, and performs secondary global optimization on the partition position adjustment result.

[0026] Specifically, a multi-objective optimization function is constructed based on the actual needs of cold chain transportation, breaking through the limitations of traditional refrigerated trucks' single temperature control objective. It simultaneously considers three core objectives: minimizing temperature deviation in the temperature zones, maximizing the utilization of the cargo compartment space, and minimizing the energy consumption of the refrigeration system. Weighting coefficients are assigned to each objective to adapt to different operating conditions. Using the initial adjustment position of the partitions as the basic solution, an intelligent optimization model integrating BP neural networks and genetic algorithms is introduced for global multi-objective optimization. Leveraging the learning ability of BP neural networks on historical and real-time data, the multi-objective optimization effect corresponding to different partition positions is accurately predicted. Then, through the global optimization characteristics of the genetic algorithm, the optimal parameters for partition adjustment that take into account multiple objectives are selected within the preset solution space, completing a secondary global optimization of the initial partition adjustment position. This makes the partition position adjustment strategy more suitable for the complex dynamic conditions of cold chain transportation, achieving an upgrade from single adjustment to multi-objective collaborative optimization.

[0027] The S500 adjusts the optimal parameters according to the partition and drives the partition to slide along the slide rail to the target position via a servo motor.

[0028] Specifically, using the optimal parameters for partition adjustment derived from the solution as the basis for precise execution, the parameter commands are transmitted in real time to the servo motor drive system via the vehicle-mounted intelligent control bus. This servo motor is equipped with a high-precision position closed-loop control module, which can precisely adjust the output power and operating speed according to the target position and sliding accuracy requirements in the optimal parameters, driving the partition to slide along the slide rail. At the same time, high-resolution displacement sensors deployed at the partition and slide rail continuously feed back real-time position data during the sliding process to the control module. The control module compares the actual position with the target position. If a position deviation is detected, it immediately sends a correction command to the servo motor to achieve real-time dynamic compensation for the deviation, ensuring that the partition can accurately fit the optimal position calculated by the algorithm and stop. The entire driving and sliding process does not require manual intervention, achieving seamless connection and precise matching from algorithm decision parameters to mechanical execution actions. This breaks through the technical drawbacks of low precision, slow response, and large adjustment errors caused by traditional manual adjustment and ordinary motor drive, allowing the optimal position of the partition after algorithm optimization to be accurately landed.

[0029] In one embodiment, such as Figure 2 As shown, S300 includes: S310 defines each temperature zone to be divided in the refrigerated truck compartment as an independent heat load calculation node, and quantifies the real-time heat load value of each heat load calculation node based on the input parameters and multi-physics field coupled fusion data. S320, combined with the three-dimensional digital model of the refrigerated truck compartment, constructs a matching relationship model between the spatial distribution of cold energy and the real-time heat load values ​​of each heat load calculation node; S330 calculates the spatial adjustment coefficient of each heat load calculation node based on the real-time heat load value and the matching relationship model through a dynamic heat load balancing algorithm. S340, based on the space adjustment coefficient and combined with the preset physical space constraints of the refrigerated truck compartment, the initial adjustment position of the partition is obtained.

[0030] Specifically, firstly, based on the spatial calibration capability of the 3D digital model of the refrigerated truck compartment, all temperature zones to be divided in the compartment are defined as independent heat load calculation nodes. Each node corresponds to a specific physical space area within the compartment and is associated with the spatial coordinates, volume parameters, and cold air flow characteristics in the 3D digital model, realizing digital node management of temperature zones. Then, combined with the previously input cargo thermal property parameters, basic operating parameters, and real-time operating data after multi-physics field coupling and fusion processing, the real-time heat load value of each heat load calculation node is quantitatively calculated through heat conduction and heat radiation related thermal calculation models. This real-time heat load value is a core quantitative indicator representing the total amount of heat that the refrigeration system needs to remove per unit time in a single temperature zone. It can accurately reflect the actual heat load demand of each temperature zone to be divided, providing a precise quantitative basis for subsequent matching of cold air and space. Secondly, using the 3D digital model of the refrigerated truck compartment as a spatial carrier, the real-time heat load values ​​of each heat load calculation node are deeply integrated with the spatial distribution characteristics of cold energy within the compartment. This constructs a matching relationship model between the spatial distribution of cold energy and the real-time heat load values ​​of each heat load calculation node. This model is not a simple numerical correspondence, but rather realizes the digital linkage mapping between spatial distribution characteristics such as cold energy density, cold energy flow path, and cold energy adjustment capability, and the real-time heat load values, heat load change trends, and cold energy demand types of each node. It can clearly present the current matching status of cold energy space and heat load demand within the compartment, identify spatial areas with redundant cold energy and node areas with excessive heat load and insufficient cold energy supply, and build a model foundation for the subsequent algorithm calculation of partition positions to accurately match cold energy and heat load. This ensures that the adjustment of partition positions always revolves around the dynamic adaptation of cold energy and heat load. Then, using the real-time heat load value of each heat load calculation node as the core quantitative indicator, and combined with the matching relationship model constructed above, the spatial adjustment coefficient of each heat load calculation node is obtained through iterative calculation using a dynamic heat load balancing algorithm. This spatial adjustment coefficient is the core quantitative parameter characterizing the spatial adjustment range and direction of the corresponding temperature zone of the node. Its calculation process takes the precise matching of cold space and heat load demand and the minimization of thermal interference between temperature zones as dual optimization objectives. At the same time, it incorporates actual working condition factors such as the natural distribution law of cold energy in the compartment and the cooling efficiency of the refrigeration system, so that the calculated spatial adjustment coefficient not only meets the actual demand of heat load, but also fits the actual refrigeration and spatial characteristics of refrigerated trucks, avoiding theoretical calculations that are divorced from actual working conditions, and ensuring that the spatial adjustment range and direction of each temperature zone can achieve dynamic balance between cold energy and heat load.Finally, based on the spatial adjustment coefficients of each heat load calculation node, and combined with the preset physical space constraints of the refrigerated truck compartment, the initial adjustment position of the partition is solved. The physical space constraints of the compartment include the motion limit boundary of the slide rail, the structural size limit of the partition, the minimum space threshold of each temperature zone, and the space avoidance requirements between the refrigeration air outlet and the return air outlet. During the solution process, the algorithm converts the spatial adjustment coefficients into actual spatial movement parameters of the partition in the three-dimensional digital model, and strictly verifies whether the parameters meet the above physical space constraints. Adjustment parameters that exceed the constraint range are corrected in real time. Finally, the initial adjustment position of the partition is obtained that conforms to the actual physical structure of the compartment, matches the real-time heat load requirements, and meets the cooling balance requirements.

[0031] In one embodiment, the spatial adjustment coefficient of each heat load calculation node is calculated using a dynamic heat load balancing algorithm based on real-time heat load values ​​and a matching relationship model, including: Based on the real-time heat load value and the cooling supply of the refrigerated truck compartment, the heat load supply-demand difference of each heat load calculation node is calculated. Based on the matching relationship model, the heat load supply-demand difference is associated with the spatial distribution of cooling in the compartment to locate the cooling spatial allocation area that is suitable for each heat load calculation node. The heat load supply-demand difference and the cooling spatial allocation area are substituted into the dynamic heat load balance algorithm to obtain the spatial adjustment coefficient of each heat load calculation node.

[0032] Specifically, firstly, using the real-time heat load value of each heat load calculation node as the core quantitative indicator, and combining it with the actual cooling capacity of the refrigerated truck's refrigeration system to each node, the heat load supply-demand difference of each node is calculated. This value can accurately reflect the imbalance between the cooling capacity supply and actual cooling demand in a single temperature zone, and quantitatively determine the specific degree of insufficient or excessive cooling capacity supply at each node. Next, based on a pre-built matching model between the spatial distribution of cooling capacity and the real-time heat load value of each heat load calculation node, the heat load supply-demand difference calculated by each node is deeply correlated and matched with the actual spatial distribution characteristics of cooling capacity within the truck compartment. Through the digital mapping of the model, it can... The algorithm precisely locates the cooling space allocation area that matches the cooling demand of each heat load calculation node, clearly defining the adjustable cooling space range, cooling density, and adjustment capability of each node. This ensures that the allocation of cooling space accurately aligns with the actual supply-demand imbalance at each node. Finally, the heat load supply-demand difference of each node, along with relevant parameters of the located cooling space allocation area, are substituted into the dynamic heat load balancing algorithm. With the optimization objectives of precise matching between cooling space and heat load and minimizing thermal interference within temperature ranges, the algorithm iteratively calculates the space adjustment coefficient for each heat load calculation node by combining the natural distribution law of cooling in the carriage with the operating characteristics of the refrigeration system. The entire calculation process is progressive, from the quantitative determination of supply and demand relationships to the precise location of cooling space, and finally to the algorithm's solution. This ensures that the determination of the space adjustment coefficient is always based on the actual operating conditions of the carriage and the real cooling demand of each node, abandoning the traditional empirical estimation method for cooling allocation and space adjustment. It achieves full-process digitalization and precision from numerical quantification to spatial matching and algorithm solution, ensuring the scientific and accurate nature of the space adjustment coefficient for each node.

[0033] In one embodiment, based on a matching relationship model, the difference between heat load supply and demand is correlated and matched with the spatial distribution of cooling capacity in the carriage to locate the cooling capacity spatial allocation area that is compatible with each heat load calculation node, including: Based on the matching relationship model, the core features of the spatial distribution of cold energy in refrigerated truck compartments are extracted and a cold energy spatial feature database is established. The core features include cold energy density distribution, spatial connectivity, and cold energy adjustment redundancy in each area. According to the positive and negative attributes and quantified values ​​of the heat load supply and demand difference, the type and magnitude of cold energy allocation demand corresponding to each heat load calculation node are determined. Cold energy allocation association rules are established, and the cold energy allocation demand of each heat load calculation node is compared and filtered with the cold energy spatial feature database according to the cold energy allocation association rules. Areas that exceed the physical space reach or have mismatched cold energy characteristics are eliminated, and finally, the cold energy spatial allocation area that is compatible with each heat load calculation node is located.

[0034] Specifically, based on the matching relationship model between the spatial distribution of cold energy and the real-time heat load values ​​of each heat load calculation node, the core features of the spatial distribution of cold energy within the refrigerated truck compartment are deeply extracted, including key attributes such as cold energy density distribution, spatial connectivity, and the redundancy of cold energy adjustment in each area. These core features are then digitally integrated to establish a structured cold energy spatial feature database, providing standardized feature basis for accurate matching of cold energy space. Next, based on the positive or negative attributes of the heat load supply-demand difference at each heat load calculation node, the type of cold energy allocation demand at each node is accurately determined, i.e., whether the node has a cold energy supplementation demand or a cold energy reduction demand. Simultaneously, based on the quantified value of the supply-demand difference, the specific characteristics of each node are further determined. The corresponding cooling capacity allocation level for each node enables both qualitative and quantitative definition of cooling capacity allocation needs. Finally, based on the actual spatial characteristics and cooling requirements of the carriage, a scientific cooling capacity allocation association rule is established. According to this rule, the cooling capacity allocation needs of each heat load calculation node are comprehensively compared and screened with various features in the cooling capacity spatial feature database. Areas that exceed the physical space reach or whose cooling capacity characteristics do not match the node requirements are strictly eliminated. Ultimately, the cooling capacity spatial allocation area that is highly compatible with the cooling requirements of each heat load calculation node is accurately located, allowing the allocation of cooling capacity space to be precisely matched with the actual needs of each node, laying a solid spatial foundation for the accurate calculation of subsequent spatial adjustment coefficients.

[0035] In one embodiment, before substituting the heat load supply-demand difference and the cooling capacity spatial allocation area into the dynamic heat load balance algorithm to obtain the spatial adjustment coefficient of each heat load calculation node, the algorithm further includes: To address temperature disturbances caused by door opening for loading and unloading and sudden changes in ambient temperature, thermal inertia compensation technology and machine learning prediction models are integrated to correct the difference between heat load supply and demand in real time.

[0036] Specifically, before substituting the heat load supply-demand difference and the cold energy spatial allocation area into the dynamic heat load balance algorithm to solve for the spatial adjustment coefficient, a real-time correction step for the heat load supply-demand difference under temperature disturbances is added. This addresses the sudden temperature fluctuations in cold chain transportation caused by door openings during loading and unloading, and abrupt changes in ambient temperature. It integrates thermal inertia compensation technology and machine learning prediction models to dynamically correct the heat load supply-demand difference, overcoming the limitations of traditional algorithms that rely solely on static data and cannot adapt to sudden changes in operating conditions. This allows the heat load supply-demand difference to better reflect real-time operating conditions. Specifically, the multi-source sensing network captures the door opening / closing status during loading and unloading, the rate of temperature change and the magnitude of temperature difference during sudden changes in ambient temperature, and identifies temperature disturbances. The system identifies the type, timing, and extent of the temperature disturbance's impact on the carriage. Based on a trained machine learning prediction model, and combined with historical data on heat load changes and cooling supply response characteristics under similar temperature disturbances, it predicts the impact trend of the current temperature disturbance on the heat load values ​​and cooling supply of each heat load calculation node, as well as the dynamic changes in the heat load supply-demand difference. Simultaneously, it integrates thermal inertia compensation technology. Based on the prediction results, and considering the carriage's insulation characteristics and the thermal inertia of cooling transfer, it provides targeted compensation and correction to the heat load supply-demand difference of each heat load calculation node. This eliminates numerical deviations caused by temperature disturbances, ensuring that the corrected heat load supply-demand difference accurately reflects the true cooling supply-demand imbalance at each node under the disturbance condition. The corrected heat load supply-demand difference is then substituted into the dynamic heat load balance algorithm along with the cold energy space allocation area. This effectively avoids the problem of heat load supply-demand difference distortion caused by temperature disturbances, which in turn leads to deviations in the calculation of space adjustment coefficients and inaccurate partition position adjustments. This allows the space adjustment coefficients solved by the algorithm to adapt to sudden changes in working conditions during cold chain transportation, further improving the calculation accuracy of the initial partition adjustment position and ensuring that the partition position adjustment is always highly matched with the actual cold energy supply and demand requirements of each node.

[0037] In one embodiment, a multi-objective optimization function is constructed, and the initial adjustment position of the partition is used as the basic solution. A fusion model of BP neural network and genetic algorithm is used to perform global multi-objective optimization, solve for and output the optimal parameters for partition adjustment, and perform secondary global optimization on the partition position adjustment result, including: With the optimization objectives of minimizing temperature deviation in the refrigerated compartment, maximizing the utilization of the compartment space, and minimizing the energy consumption of the refrigeration system, a multi-objective optimization function is constructed by assigning dynamically adjustable weight coefficients to each objective. The initial adjustment position of the partition is used as the basic solution, and the optimization range is determined by combining the physical space constraints of the refrigerated truck compartment and the temperature zone isolation requirements, thus constructing a solution space for multi-objective optimization. A trained BP neural network prediction model is used to predict the optimization effect of different partition positions, and the multi-objective optimization function is used as the fitness function. With the prediction results of the BP neural network as a reference, a genetic algorithm is used to perform global multi-objective optimization within the solution space, iteratively selecting the optimal solution. The effectiveness of the optimal solution is verified, and finally, the optimal parameters for partition adjustment are solved and output, completing the secondary global optimization of the partition position adjustment results.

[0038] Specifically, the three core optimization objectives are minimizing temperature deviation in the refrigerated truck compartment, maximizing the utilization of the cargo space, and minimizing the energy consumption of the refrigeration system. Each objective is assigned a dynamically adjustable weight coefficient, and the priority of each optimization objective is adjusted through weight assignment, thus constructing a multi-objective optimization function that considers multiple dimensions of needs. Next, the initial adjustment position of the partitions, obtained through a dynamic heat load balancing algorithm, is used as the basic solution. Combined with the physical space constraints of the refrigerated truck compartment and the core threshold for temperature zone isolation, the adjustable range of the partition positions is defined, thereby constructing a solution space for multi-objective optimization. Subsequently, historical partition adjustment data and real-time operating condition data are used as training samples to train and validate the BP neural network prediction model. The parameters of different partition positions within the solution space are then input into the trained BP neural network. The network prediction model accurately predicts the optimization effects of temperature deviation in each location, cabin space utilization, and refrigeration system energy consumption. This prediction result is then substituted into a preset multi-objective optimization function to calculate the fitness value corresponding to each partition position parameter. This fitness value serves as the core basis for the genetic algorithm to determine the quality of individual partitions. Finally, using the fitness value as a reference, the genetic algorithm performs global multi-objective optimization in the solution space. Through multiple rounds of selection, crossover, and mutation, the solution with the optimal fitness value is iteratively selected. The optimal solution is then validated, and solutions that do not meet the physical constraints of the cabin, the operating requirements of the refrigeration system, and the temperature zone isolation standards are eliminated. Finally, the optimal partition adjustment parameters that balance the three core optimization objectives are solved and output, completing the secondary global optimization of the initial partition adjustment position. The entire process combines the accurate prediction capability of the BP neural network with the global optimization characteristics of the genetic algorithm. Guided by the multi-objective optimization function, it achieves refined iterative optimization from the basic solution to the optimal solution. This ensures that the output optimal partition adjustment parameters not only guarantee accurate and stable temperature in the temperature zones but also improve cabin space utilization and reduce refrigeration system energy consumption.

[0039] In one embodiment, a trained BP neural network prediction model predicts the optimization effect of different partition positions, and a multi-objective optimization function is used as the fitness function. With the prediction results of the BP neural network as a reference, a genetic algorithm is used to perform global multi-objective optimization in the solution space, iteratively selecting the optimal solution, including: Based on the solution space, a real-time encoding method is used to encode the partition position parameters to generate the initial population of the genetic algorithm. The partition position parameters corresponding to each individual in the initial population are input into the trained BP neural network prediction model to obtain the predicted optimization effect value corresponding to each partition position. The predicted optimization effect value is substituted into the multi-objective optimization function to calculate the fitness value of each individual, which serves as the basis for judging the quality of individuals in the genetic algorithm. The initial population is subjected to selection, crossover, and mutation genetic operations in sequence, retaining individuals with better fitness values ​​to generate a new generation population. The genetic operations are repeated until the preset iteration termination condition is met. The best individual retained during the iteration process is taken as the optimal solution for global multi-objective optimization.

[0040] Specifically, firstly, based on a pre-defined multi-objective optimization solution space, and combining parameters such as the adjustment accuracy and sliding range of the partition positions, the partition position parameters are digitally encoded using a real-time encoding method. This transforms the physical partition positions into individuals that the genetic algorithm can recognize and process, thus generating the initial population of the genetic algorithm, providing the basic computational units for subsequent global optimization. Next, the partition position parameters corresponding to each individual in the initial population are input one by one into a BP neural network prediction model trained on historical data and verified for accuracy. Based on the learned correlation between partition positions and temperature control effects, space utilization, and energy consumption levels, the model accurately outputs multi-dimensional optimization effect prediction values ​​for each partition position, including temperature deviation of the temperature zone, cabin space utilization, and refrigeration system energy consumption, providing accurate data support for subsequent fitness calculations. Finally, the optimization effect prediction values ​​for each partition position are... Substituting the constructed multi-objective optimization function, the fitness value corresponding to each individual is obtained through function operation. This value comprehensively reflects the degree to which the corresponding partition position satisfies the three core optimization objectives and directly serves as the core basis for the genetic algorithm to determine the quality of individuals. Finally, the classic genetic operations of selection, crossover, and mutation are sequentially performed on the initial population. In the selection stage, individuals with better fitness values ​​are retained; in the crossover stage, parameters between different individuals are fused; and in the mutation stage, parameters are randomly adjusted to ensure population diversity. A new generation of population is generated through a single genetic operation, and the above process of effect prediction, fitness calculation, and genetic operation is repeated on the new generation of population until the preset iteration termination conditions are met (such as the number of iterations reaching a threshold or the optimal fitness value stabilizing). Finally, the individual with the best fitness value retained throughout the iteration process is taken as the optimal solution for this global multi-objective optimization. The entire process achieves a deep integration of prediction and optimization. The predictive ability of the BP neural network effectively reduces the blindness of the genetic algorithm's optimization and greatly improves the optimization efficiency. Meanwhile, the global optimization characteristic of the genetic algorithm can break through the limitation of local optima and select the global optimal solution that takes into account minimizing the temperature deviation of the temperature zone, maximizing the utilization of the carriage space, and minimizing the energy consumption of the refrigeration system. This provides reliable algorithmic support for the subsequent output of the optimal parameters for partition adjustment and the realization of precise secondary optimization of the partition position. Compared with traditional single optimization algorithms, the optimization results of this process are more scientific and more adaptable, further ensuring the accuracy and rationality of the intelligent adjustment of the partition position.

[0041] In one embodiment, collecting real-time data of the entire scene inside and outside the carriage through a multi-source sensing network and performing multi-physics coupling and fusion processing on the collected real-time data includes: A multi-source sensing network is used to collect temperature and humidity data inside the refrigerated truck compartment, cargo heat load data, equipment operating status data, environmental conditions data outside the compartment, and vehicle driving status data to form a full-dimensional raw sensing dataset. The full-dimensional raw sensing dataset is then preprocessed by filtering, denoising, and normalization to remove abnormal data caused by sensor malfunctions and vehicle bumps, resulting in standardized sensing data. Based on the coupling relationship between the temperature field, airflow field, cargo thermal characteristic field, and environmental condition field, multi-physics field coupling fusion processing is carried out on the standardized sensing data to extract the core correlation feature parameters of each physical field, forming multi-physics field coupling fusion data.

[0042] Specifically, the first step is to activate a distributed multi-source sensing network deployed inside and outside the refrigerated truck compartment. This network integrates multiple types and locations of sensing terminals, including high-precision temperature and humidity sensors, cargo heat load monitoring sensors, refrigeration equipment operation status sensors, environmental condition sensors, and vehicle driving status sensors. This enables the synchronous and real-time collection of temperature and humidity data inside the refrigerated truck compartment, real-time cargo heat load data, and operation status data of the refrigeration system and partition drive equipment. It also allows for the collection of environmental condition data outside the compartment, such as temperature, humidity, and atmospheric pressure, as well as driving status data, including vehicle speed, start / stop status, and road conditions. This comprehensively captures all-scenario operating condition information inside and outside the truck compartment during cold chain transportation, forming a full-dimensional raw sensing dataset covering cargo, equipment, environment, and vehicle. This breaks the limitation of traditional refrigerated trucks only collecting single temperature and humidity data, achieving full-dimensional sensing of operating condition data. Next, a systematic preprocessing operation is performed on the collected full-dimensional raw sensing dataset, sequentially employing professional filtering algorithms, outlier removal rules, and... The data normalization process filters, denoises, and normalizes the raw data, effectively eliminating abnormal data caused by sensor malfunctions, signal interference, vehicle vibrations, and other factors. This eliminates data redundancy and errors, resulting in standardized sensing data and preventing interference from the raw data from affecting the accuracy of subsequent algorithm calculations. Finally, based on the inherent coupling relationship and dynamic interaction law between temperature field, airflow field, cargo thermal characteristic field, and environmental condition field during cold chain transportation, a multi-physics field fusion algorithm is used to conduct deep multi-physics field coupling fusion processing on the standardized sensing data. Through feature extraction, correlation analysis, and data fusion, the inherent correlation features between various physical field data are mined, and core correlation feature parameters of each physical field that can accurately reflect the actual working conditions inside and outside the vehicle are extracted. The scattered single physical field data are fused into interrelated and mutually supportive multi-physics field coupling fusion data, realizing an upgrade from single data collection to multi-field feature fusion. This allows the processed data to accurately and comprehensively reflect the actual dynamic working conditions of cold chain transportation.

[0043] In one embodiment, such as Figure 3As shown, a multi-temperature zone refrigerated truck partition position intelligent adjustment device is provided, including: a parameter input module 10, a fusion processing module 20, an initial position solving module 30, an optimal parameter output module 40, and a motor drive module 50, wherein: The parameter input module 10 is used to input the thermal properties parameters and basic operating parameters of the cargo before loading, and simultaneously construct a three-dimensional digital model of the refrigerated truck compartment. The fusion processing module 20 is used to collect real-time data of the entire scene inside and outside the carriage through a multi-source sensing network after the cargo is loaded, and to perform multi-physics field coupling fusion processing on the collected real-time data. The initial position solving module 30 is used to solve for the initial adjustment position of the partition by using a dynamic heat load balance algorithm based on the input parameters, the three-dimensional digital model of the refrigerated truck compartment and the multi-physics field coupled fusion data. The optimal parameter output module 40 is used to construct a multi-objective optimization function, and with the initial adjustment position of the partition as the basic solution, it performs global multi-objective optimization through a fusion model of BP neural network and genetic algorithm, solves and outputs the optimal parameters for partition adjustment, so as to perform secondary global optimization on the partition position adjustment result; The motor drive module 50 is used to adjust the optimal parameters according to the partition and drive the partition to slide along the slide rail to the target position via a servo motor.

[0044] In one embodiment, the initial position solving module 30 is further configured to define each temperature zone to be divided within the refrigerated truck compartment as an independent heat load calculation node, and to quantify the real-time heat load value of each heat load calculation node based on the input parameters and multi-physics field coupled fusion data; to construct a matching relationship model between the spatial distribution of cold energy and the real-time heat load value of each heat load calculation node by combining the three-dimensional digital model of the refrigerated truck compartment; to calculate the spatial adjustment coefficient of each heat load calculation node by using a dynamic heat load balance algorithm based on the real-time heat load value and the matching relationship model, wherein the spatial adjustment coefficient characterizes the spatial adjustment amplitude and adjustment direction of the temperature zone corresponding to the node; and to solve for the initial adjustment position of the partition by combining the spatial adjustment coefficient with the preset physical space constraints of the refrigerated truck compartment.

[0045] In one embodiment, the initial location solving module 30 is further configured to calculate the heat load supply-demand difference of each heat load calculation node based on the real-time heat load value and the cooling supply of the refrigerated truck compartment; based on the matching relationship model, associate and match the heat load supply-demand difference with the cooling space distribution of the compartment to locate the cooling space allocation area that is compatible with each heat load calculation node; and substitute the heat load supply-demand difference and the cooling space allocation area into the dynamic heat load balance algorithm to solve for the spatial adjustment coefficient of each heat load calculation node.

[0046] In one embodiment, the initial location solving module 30 is further used to extract the core features of the spatial distribution of cold energy in the refrigerated truck compartment based on the matching relationship model and establish a cold energy spatial feature database. The core features include cold energy density distribution, spatial connectivity, and cold energy adjustment redundancy in each area. Based on the positive and negative attributes and quantified values ​​of the heat load supply and demand difference, the module determines the type and magnitude of cold energy allocation demand corresponding to each heat load calculation node. It establishes cold energy allocation association rules and compares and filters the cold energy allocation demand of each heat load calculation node with the cold energy spatial feature database according to the cold energy allocation association rules, eliminating areas that exceed the physical space reach and have mismatched cold energy characteristics, and finally locates the cold energy spatial allocation area that is compatible with each heat load calculation node.

[0047] In one embodiment, the intelligent adjustment device for the partition position of the multi-temperature zone refrigerated truck also includes a correction module, which is used to correct the difference between heat load supply and demand in real time for temperature disturbances caused by sudden changes in ambient temperature during door opening for loading and unloading and loading and unloading.

[0048] In one embodiment, the optimal parameter output module 40 is further configured to construct a multi-objective optimization function by configuring dynamically adjustable weight coefficients for each optimization objective, taking the minimization of temperature deviation in the temperature zone, the maximization of cargo space utilization, and the minimization of refrigeration system energy consumption as optimization objectives; taking the initial adjustment position of the partition as the basic solution, and combining the physical space constraints of the refrigerated truck cargo compartment and the temperature zone isolation requirements to determine the optimization range, thus constructing a solution space for multi-objective optimization; predicting the optimization effect of different partition positions through a trained BP neural network prediction model, and using the multi-objective optimization function as the fitness function, with the prediction results of the BP neural network as a reference, performing global multi-objective optimization in the solution space through a genetic algorithm, iteratively selecting the optimal solution; verifying the effectiveness of the optimal solution, and finally solving and outputting the optimal parameters for partition adjustment, thus completing the secondary global optimization of the partition position adjustment result.

[0049] In one embodiment, the optimal parameter output module 40 is further configured to encode the partition position parameters based on the solution space using a real-time encoding method to generate an initial population for the genetic algorithm; input the partition position parameters corresponding to each individual in the initial population into the trained BP neural network prediction model to obtain the predicted optimization effect value corresponding to each partition position; substitute the predicted optimization effect value into the multi-objective optimization function to calculate the fitness value of each individual, which serves as the basis for judging the quality of individuals in the genetic algorithm; perform genetic operations of selection, crossover, and mutation sequentially on the initial population, retain individuals with better fitness values ​​to generate a new generation population, and repeat the genetic operations until the preset iteration termination condition is met, and use the best individual retained during the iteration process as the optimal solution for global multi-objective optimization.

[0050] In one embodiment, the fusion processing module 20 is further configured to collect temperature and humidity data, cargo heat load data, equipment operating status data, environmental conditions data, and vehicle driving status data inside the refrigerated truck compartment via a multi-source sensing network, forming a full-dimensional raw sensing dataset; perform filtering, noise reduction, and normalization preprocessing on the full-dimensional raw sensing dataset in sequence, remove abnormal data caused by sensor failures and vehicle bumps, and obtain standardized sensing data; based on the coupling relationship between the temperature field, airflow field, cargo thermal characteristic field, and environmental conditions field, perform multi-physics field coupling fusion processing on the standardized sensing data, extract the core correlation feature parameters of each physical field, and form multi-physics field coupling fusion data.

[0051] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for intelligent adjustment of the partition position in a multi-temperature zone refrigerated truck, characterized in that, include: Before loading cargo, input the cargo's thermal properties and basic operating parameters, and simultaneously construct a three-dimensional digital model of the refrigerated truck compartment. After the cargo is loaded, real-time data of the entire scene inside and outside the carriage is collected through a multi-source sensing network, and the collected real-time data is processed by multi-physics coupling and fusion. Based on the input parameters, the three-dimensional digital model of the refrigerated truck compartment, and the multi-physics field coupled fusion data, the initial adjustment position of the partition is obtained by solving the dynamic heat load balance algorithm. A multi-objective optimization function is constructed, and the initial adjustment position of the partition is used as the basic solution. A global multi-objective optimization is performed through a fusion model of BP neural network and genetic algorithm to solve and output the optimal parameters for partition adjustment, so as to perform secondary global optimization on the partition position adjustment result. The optimal parameters are adjusted according to the partition, and the partition is driven by a servo motor to slide along the slide rail to the target position.

2. The intelligent adjustment method for the position of the partition of a multi-temperature zone refrigerated truck according to claim 1, characterized in that, The process of determining the initial adjustment position of the partition based on the input parameters, the three-dimensional digital model of the refrigerated truck compartment, and multiphysics coupled fusion data through a dynamic heat load balancing algorithm includes: Each temperature zone to be divided in the refrigerated truck compartment is defined as an independent heat load calculation node, and the real-time heat load value of each heat load calculation node is quantitatively calculated based on the input parameters and multi-physics field coupled fusion data. Based on the three-dimensional digital model of the refrigerated truck compartment, a matching relationship model between the spatial distribution of cold energy and the real-time heat load values ​​of each heat load calculation node is constructed. Based on the real-time heat load value and the matching relationship model, the spatial adjustment coefficient of each heat load calculation node is calculated by the dynamic heat load balance algorithm, wherein the spatial adjustment coefficient characterizes the spatial adjustment range and adjustment direction of the temperature zone corresponding to the node. Based on the space adjustment coefficient and the physical space constraints of the pre-set refrigerated truck compartment, the initial adjustment position of the partition is obtained.

3. The intelligent adjustment method for the position of the partition of a multi-temperature zone refrigerated truck according to claim 2, characterized in that, The step of calculating the spatial adjustment coefficient of each heat load calculation node using a dynamic heat load balancing algorithm based on the real-time heat load value and the matching relationship model includes: Based on the real-time heat load value and the cooling supply of the refrigerated truck compartment, the heat load supply-demand difference of each heat load calculation node is calculated. Based on the matching relationship model, the heat load supply and demand difference is associated and matched with the spatial distribution of cooling capacity in the carriage to locate the cooling capacity spatial allocation area that is compatible with each heat load calculation node. Substituting the heat load supply-demand difference and the cooling capacity spatial allocation area into the dynamic heat load balance algorithm, the spatial adjustment coefficient of each heat load calculation node is obtained.

4. The intelligent adjustment method for the position of the partition of a multi-temperature zone refrigerated truck according to claim 3, characterized in that, The process of associating and matching the heat load supply-demand difference with the spatial distribution of cooling capacity in the carriage based on the matching relationship model, and locating the cooling capacity spatial allocation area that is compatible with each heat load calculation node, includes: Based on the matching relationship model, the core features of the spatial distribution of cold energy in the refrigerated truck compartment are extracted and a cold energy spatial feature database is established. The core features include cold energy density distribution, spatial connectivity and cold energy adjustment redundancy in each region. Based on the positive and negative attributes and quantified values ​​of the heat load supply and demand difference, determine the type and magnitude of cooling capacity allocation requirements corresponding to each heat load calculation node; Establish cooling capacity allocation association rules, and compare and filter the cooling capacity allocation requirements of each heat load calculation node with the cooling capacity spatial feature database according to the cooling capacity allocation association rules. Eliminate areas that exceed the physical space reach range and whose cooling capacity characteristics do not match, and finally locate the cooling capacity spatial allocation area that is compatible with each heat load calculation node.

5. The intelligent adjustment method for the position of the partition of a multi-temperature zone refrigerated truck according to claim 3, characterized in that, Before substituting the heat load supply-demand difference and the cooling capacity spatial allocation area into the dynamic heat load balance algorithm to obtain the spatial adjustment coefficient of each heat load calculation node, the method further includes: To address temperature disturbances caused by door opening for loading and unloading and sudden changes in ambient temperature, thermal inertia compensation technology and machine learning prediction models are integrated to correct the difference between heat load supply and demand in real time.

6. The intelligent adjustment method for the position of the partition of a multi-temperature zone refrigerated truck according to claim 1, characterized in that, The process of constructing a multi-objective optimization function, using the initial adjustment position of the partition as the basic solution, and performing global multi-objective optimization through a fusion model of BP neural network and genetic algorithm to solve for and output the optimal parameters for partition adjustment, and performing secondary global optimization on the partition position adjustment results, includes: With the optimization objectives of minimizing temperature deviation in the temperature zone, maximizing the utilization of the cabin space, and minimizing the energy consumption of the refrigeration system, a multi-objective optimization function is constructed by configuring dynamically adjustable weight coefficients for each optimization objective. Using the initial adjustment position of the partition as the basic solution, and combining the physical space constraints of the refrigerated truck compartment and the temperature zone isolation requirements, the optimization range is determined, and a solution space for multi-objective optimization is constructed. The optimization effect of different partition positions is predicted by a trained BP neural network prediction model. The multi-objective optimization function is used as the fitness function. With the prediction result of the BP neural network as a reference, a genetic algorithm is used to perform global multi-objective optimization in the solution space and iteratively select the optimal solution. The effectiveness of the optimal solution is verified, and the optimal parameters for the baffle adjustment are finally solved and output, thus completing the secondary global optimization of the baffle position adjustment result.

7. The intelligent adjustment method for the position of the partition of a multi-temperature zone refrigerated truck according to claim 6, characterized in that, The method involves using a trained BP neural network prediction model to predict the optimization effect of different partition positions, and using the multi-objective optimization function as the fitness function. With the prediction results of the BP neural network as a reference, a genetic algorithm is used to perform global multi-objective optimization within the solution space, iteratively selecting the optimal solution, including: Based on the solution space, the partition position parameters are encoded using a real-time encoding method to generate the initial population for the genetic algorithm. The partition position parameters corresponding to each individual in the initial population are input into the trained BP neural network prediction model to obtain the predicted value of the optimization effect corresponding to each partition position. The predicted optimization effect value is substituted into the multi-objective optimization function to calculate the fitness value of each individual, which serves as the basis for judging the quality of individuals in the genetic algorithm. Genetic operations of selection, crossover, and mutation are performed sequentially on the initial population to retain individuals with better fitness values, generating a new generation of population. The genetic operations are repeated until a preset iteration termination condition is met, and the best individual retained during the iteration process is taken as the optimal solution for global multi-objective optimization.

8. The intelligent adjustment method for the position of the partition of a multi-temperature zone refrigerated truck according to claim 1, characterized in that, The process of collecting real-time data from the entire interior and exterior of the vehicle through a multi-source sensing network and performing multi-physics coupling and fusion processing on the collected real-time data includes: By using a multi-source sensing network, temperature and humidity data inside the refrigerated truck compartment, cargo heat load data, equipment operating status data, environmental conditions data outside the compartment, and vehicle driving status data are collected to form a full-dimensional raw sensing dataset. The original full-dimensional perception dataset is preprocessed by filtering, denoising, and normalization in sequence to remove abnormal data caused by sensor failure and vehicle bumps, and standardized perception data is obtained. Based on the coupling relationship between temperature field, airflow field, cargo thermal characteristic field and environmental condition field, multi-physics field coupling fusion processing is carried out on the standardized sensing data to extract the core correlation feature parameters of each physical field and form multi-physics field coupling fusion data.

9. A smart adjustment device for the position of the partition of a multi-temperature zone refrigerated truck, characterized in that, include: The parameter input module is used to input the thermal properties and basic operating parameters of the cargo before loading, and simultaneously build a three-dimensional digital model of the refrigerated truck compartment. The fusion processing module is used to collect real-time data of the entire scene inside and outside the carriage through a multi-source sensing network after the cargo is loaded, and to perform multi-physics field coupling fusion processing on the collected real-time data. The initial position solution module is used to solve for the initial adjustment position of the partition by using a dynamic heat load balance algorithm based on the input parameters, the three-dimensional digital model of the refrigerated truck compartment and the multi-physics field coupled fusion data. The optimal parameter output module is used to construct a multi-objective optimization function, and with the initial adjustment position of the partition as the basic solution, it performs global multi-objective optimization through a fusion model of BP neural network and genetic algorithm, solves and outputs the optimal parameters for partition adjustment, so as to perform secondary global optimization on the partition position adjustment result; The motor drive module is used to adjust the optimal parameters according to the partition and drive the partition to slide along the slide rail to the target position via a servo motor.

Images