A veterinary blood sample active maintenance environmental conditioning system

Abstract

The present application relates to a kind of veterinary blood sample active maintenance environment regulation system, specifically to the field of intelligent temperature control, the scheme is realized by constructing sample digital twin and three-dimensional thermodynamic simplified digital model, the accurate digitalization characterization of each sample individualized storage demand and the high-fidelity dynamic mapping of physical microenvironment;Based on the simulation and situation analysis executed based on comprehensive data pedestal, the evolution of environment can be predicted in advance and the suitability atlas can be generated;And can simultaneously meet the optimal position layout and control instruction of the differentiated needs of multiple samples;Finally drive distributed actuator array to accurately shape local microenvironment, and continuously correct model internal parameters according to real-time feedback, form a self-improving closed loop;Fundamentally, the control target is improved from maintaining space uniform temperature to guaranteeing the individualized activity maintenance optimum of each sample unit, so as to significantly enhance the reliability and quality of sample pretreatment under complex dynamic scene.

Description

Technical Field

[0001] This invention relates to the field of intelligent temperature control, and more specifically, to a veterinary blood sample viability maintenance environment regulation system. Background Technology

[0002] In the sample receiving and pretreatment stage of veterinary diagnostic laboratories, a large number of blood samples from diverse sources and with different testing requirements are delivered in a short period of time. These samples typically come from various animals such as dogs, cats, horses, and poultry, and their testing items cover multiple categories such as complete blood count, biochemical analysis, and pathogen molecular diagnosis. Different types of anticoagulants are used for pretreatment depending on the purpose of the test. This heterogeneity leads to different short-term storage requirements for different samples before testing, including specific optimal activity maintenance temperatures and permissible standing time windows. Currently, laboratories generally use large medical refrigerators or incubators as centralized temporary storage equipment for samples. However, this scenario has inherent complexity and dynamism: on the one hand, the samples themselves differ in container specifications, liquid volume, and thermal properties; on the other hand, the frequent opening and closing of cabinet doors by laboratory personnel to access samples introduces strong airflow disturbances and heat exchange, causing intermittent temperature shocks to the microenvironment inside the cabinet. This means that the samples are actually in a dynamic environment with non-uniform temperature and constant disturbance while waiting for testing.

[0003] In existing technologies, the temperature control strategies of thermostatic devices used for temporary storage of blood samples have significant limitations. Mainstream devices typically employ feedback control mechanisms based on a single or a few monitoring points, aiming to maintain a stable setpoint or average temperature within the cabinet, such as uniformly maintaining it at 4 degrees Celsius. This homogenized control mode assumes that all samples have the same storage requirements and treats the storage space as a thermodynamically homogeneous whole. However, due to the influence of equipment structure, distribution of heat and cold sources, and airflow paths, inherent and difficult-to-eliminate temperature gradients exist at different physical locations within the cabinet. A more critical issue is that existing sample information management systems and the environmental control systems of the storage devices are independent of each other, forming information silos. The temperature control system cannot automatically identify and acquire the specific biological storage requirements of each sample unit, and therefore cannot drive the actuator to provide differentiated microenvironment protection for different samples. Its control target and evaluation standard are always the stability of the physical temperature parameters of the equipment space, rather than the optimal preservation of the biological activity of each specific sample. This leads to the risk of co-storing samples that are highly sensitive to temperature fluctuations with ordinary samples, and it is also unable to provide accurate buffering and compensation for the local environment of critical samples under frequent human interference. Therefore, how to achieve accurate and differentiated microenvironment allocation and maintenance on demand in a dynamic and non-uniform physical space according to the individual needs of samples has become a prominent technical bottleneck in improving the quality and reliability of sample pretreatment. Summary of the Invention

[0004] This invention addresses the technical problems existing in the prior art by providing a veterinary blood sample viability maintenance environment control system. This system utilizes a data acquisition and modeling module, a simulation and situation analysis module, a collaborative planning module, and an execution and calibration control module to solve the problems mentioned in the background section.

[0005] The technical solution of this invention to solve the above-mentioned technical problems is as follows: specifically including: Data acquisition and modeling module: Reads sample identifiers and obtains corresponding metadata, creates digital twins of samples; Based on real-time data from a multi-sensor network, initializes a simplified three-dimensional thermodynamic digital model in the storage device, binds the twin to the physical location, and generates a comprehensive data base; Simulation and Situation Analysis Module: Calls the data base to perform multiphysics coupling simulation, predicts the evolution of the internal environmental field, performs demand clustering analysis on the twin set, and generates a suitability map by calculating the temperature matching degree of each point in space; Collaborative planning module: Receives the suitability map and, using a rolling time-domain optimization method, jointly solves for the optimal sample location allocation scheme and equipment control command sequence; Execution and calibration control module: executes the sequence of device control commands to drive the coordinated action of the distributed actuator array, and corrects the internal parameters of the simplified three-dimensional thermodynamic digital model based on feedback data from the multi-sensor network.

[0006] In a preferred embodiment, the process of creating a sample digital twin in the data acquisition and modeling module is as follows: The sample identifier is obtained by reading the RFID tag attached to the sample container. Based on the sample identifier, the corresponding sample metadata is retrieved from the associated laboratory information management system. The sample metadata includes the optimal activity retention temperature, the allowable temperature deviation threshold, the preset total settling time, and the sample priority weight. Based on sample metadata, a structured data object is instantiated in the virtual space as a sample digital twin. The sample digital twin includes sample identifier, optimal activity holding temperature, allowable temperature deviation threshold, preset total settling time, real-time remaining settling time calculated based on the current time and sample collection timestamp, and sample priority weight. It also introduces a derived parameter, namely the urgency of storage demand, which is determined by the sample priority weight and the ratio of the real-time remaining idle time to the preset total idle time.

[0007] In a preferred embodiment, the process of initializing the simplified three-dimensional thermodynamic digital model is as follows: It continuously receives real-time monitoring data streams from a pre-deployed multi-sensor network inside the storage device. The multi-sensor network consists of temperature sensors, humidity sensors, and micro-airflow wind speed sensors that are uniformly distributed on three-dimensional spatial grid nodes inside the storage device. Maintain a discretized digital model that includes the internal geometry of the storage device, mesh generation, material thermal properties, and dynamic boundary conditions, i.e., a three-dimensional thermodynamic simplified digital model; During initialization, the latest real-time monitoring data from the multi-sensor network is mapped to the corresponding grid nodes of the simplified three-dimensional thermodynamic digital model through a spatial interpolation algorithm, thereby synchronizing the current state of the simplified three-dimensional thermodynamic digital model with the real environmental state of the physical world.

[0008] In a preferred embodiment, the process of generating the integrated data base is as follows: When the sample container is placed in the physical shelf location of the storage device, obtain the coordinate information of that physical shelf location; Perform the binding operation to associate and record the sample digital twin with the coordinate information; Finally, all the created sample digital twins and their associated coordinate information, along with the current state snapshot of the completed state-synchronized three-dimensional thermodynamic simplified digital model, are encapsulated into a unified data structure, namely the integrated data base.

[0009] In a preferred embodiment, the process of performing multiphysics coupling simulation and demand clustering analysis in the simulation and situation analysis module is as follows: Extract the current state of the simplified 3D thermodynamic digital model from the integrated data base, as well as the physical coordinates and thermal properties of all sample digital twins; Using the current state as the initial condition, and superimposing a dynamic heat source term formed by the set of physical coordinates associated with each sample digital twin, the coupled fluid motion and energy conservation equations are solved transiently. The three-dimensional temperature field and airflow velocity field inside the storage device are predicted in rolling predictions for multiple consecutive time slices in the future, forming a predicted environmental field sequence containing temperature field data for each future moment. Meanwhile, based on the two attributes of optimal activity retention temperature and storage urgency contained in each sample digital twin in the sample digital twin set, density-based clustering analysis is performed to group samples with similar optimal activity retention temperature and similar storage urgency into the same demand cluster. Subsequently, for each point inside the storage device, based on its distance to the physical coordinates of each sample digital twin, the urgency of storage needs of each sample digital twin, and the optimal activity retention temperature of each sample digital twin, a spatial demand temperature field representing the comprehensive expected temperature value of the location point is calculated.

[0010] In a preferred embodiment, the process of generating the suitability map is as follows: For each future moment and each spatial location in the predicted environmental field sequence, the comprehensive expected temperature value of the corresponding location is obtained from the spatial demand temperature field. Based on the urgency of storage demand and the allowable temperature deviation threshold of each sample digital twin around the location, a matching tolerance characterizing the temperature tolerance range of the location is dynamically calculated. The predicted temperature values ​​of the corresponding future time and spatial location points recorded in the predicted environmental field sequence are compared with the comprehensive expected temperature value of the location point. The deviation relative to the normalized result of the matching tolerance is mapped to a spatiotemporal temperature matching degree value between zero and one through a nonlinear decay function. After completing this calculation for all spatiotemporal points, a four-dimensional data field with the same dimension as the predicted environmental field sequence is obtained. This four-dimensional data field is the microenvironment suitability map.

[0011] In a preferred embodiment, the process of constructing the optimization problem and decision variables in the collaborative planning module is as follows: Using the received microenvironment suitability map as input, a rolling optimization time domain with a fixed future step size is defined; The first type of decision variable is defined as a binary decision matrix that encodes the sample location allocation scheme, representing the allocation relationship between each sample digital twin and each physical shelf location; The second type of decision variable is defined as a continuous decision vector sequence of all distributed actuators constituting the device control command sequence at each time step within the rolling optimization time domain; The overall optimization objective function is the sum of two parts. The first part is the future adaptive benefit, which is obtained by weighted integration of the microenvironment suitability map in the spatial and temporal dimensions within the rolling optimization time domain. The weight of each spatial location point is proportional to the urgency of the storage demand of the sample digital twin located at that spatial location point. The second part is the control energy consumption cost, which is obtained by multiplying the square norm of the continuous decision vector sequence that constitutes the equipment control command sequence by a preset positive real number penalty coefficient, aiming to minimize the energy consumption of control actions.

[0012] In a preferred embodiment, the process of finding the optimal solution is as follows: A decomposition-coordination iterative optimization algorithm is used to solve the overall optimization objective function. Each iteration contains two alternating sub-optimization processes. The first sub-optimization process is location planning. Under the condition of fixing the current continuous decision vector sequence that constitutes the equipment control command sequence, the binary decision matrix used to encode the sample location allocation scheme is updated by solving an integer programming sub-problem with the binary decision matrix used to encode the sample location allocation scheme as the decision variable. The second sub-optimization process is control optimization. Under the condition of fixing the current sample position allocation scheme, a convex quadratic programming subproblem is solved, with the continuous decision vector sequence constituting the equipment control command sequence as the decision variable. The continuous decision vector sequence constituting the equipment control command sequence is updated. The two sub-optimization processes are executed alternately until the change in the overall optimization objective function value is less than a preset threshold. The final jointly optimized optimal sample position allocation scheme and optimal equipment control command sequence are output, and the equipment control command corresponding to the first moment in the rolling optimization time domain of the optimal equipment control command sequence is issued for execution.

[0013] In a preferred embodiment, the process of executing control commands and collecting feedback data in the execution and calibration control module is as follows: Receive the device control command sequence and immediately parse the control command vector of the first moment in the device control command sequence, convert it into a specific control signal to drive each independent execution unit in the distributed actuator array. The control signal is converted into a physical quantity by the power driver, driving each actuator to work together to change the local microenvironment inside the storage device. Within a preset time period after the control command takes effect, real-time monitoring data from all sensors in the multi-sensor network are collected synchronously at a frequency higher than the control cycle. Timestamp alignment and outlier cleaning are performed on these feedback data to generate a measured environmental field data sequence that corresponds to the output dimension of the three-dimensional thermodynamic simplified digital model and is time-synchronized.

[0014] In a preferred embodiment, the process of correcting the internal parameters of the simplified three-dimensional thermodynamic digital model is as follows: Based on the measured environmental field data sequence, the predicted environmental field sequence obtained by simulation prediction using the current three-dimensional thermodynamic simplified digital model under the same control command input is compared, and the prediction error vector between the two is calculated. At the same time, a parameter sensitivity matrix is ​​maintained, which describes the degree of influence of the changes of each correctable physical parameter in the three-dimensional thermodynamic simplified digital model on the predicted values ​​corresponding to each measuring point of the sensor network. Based on the prediction error vector and parameter sensitivity matrix, and by introducing a weight matrix that is positively correlated with the urgency of storage requirements for samples near the sensor location, a set of correction amounts for the internal correctable physical parameters of the simplified three-dimensional thermodynamic digital model is calculated using a least squares estimation algorithm with regularization terms. Finally, the calculated correction values ​​are used to update the corresponding internal parameters of the simplified three-dimensional thermodynamic digital model, and the state of the simplified three-dimensional thermodynamic digital model is reinitialized using the latest measured environmental field data, thus completing the online calibration.

[0015] The beneficial effects of this invention are as follows: By constructing a digital twin of the sample and a simplified three-dimensional thermodynamic digital model, it achieves accurate digital representation of the personalized storage needs of each sample and high-fidelity dynamic mapping of the physical microenvironment; based on simulation and situational analysis performed by the integrated data base, it can proactively predict environmental evolution and generate suitability maps; it can simultaneously meet the optimal location layout and control commands for the differentiated needs of multiple samples; ultimately, it drives the distributed actuator array to accurately shape the local microenvironment and continuously correct the internal parameters of the model based on real-time feedback, forming a self-improving closed loop; fundamentally, it elevates the control objective from maintaining a uniform temperature to ensuring that the individualized activity of each sample unit remains optimal, thereby significantly enhancing the reliability and quality of sample preprocessing in complex dynamic scenarios. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a block diagram of the system structure of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0018] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0019] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0020] Example 1: This example provides the following... Figures 1-2 The system shown is a veterinary blood sample viability maintenance environment control system, specifically comprising: Data acquisition and modeling module: Reads sample identifiers and obtains corresponding metadata, creates digital twins of samples; Based on real-time data from a multi-sensor network, initializes a simplified three-dimensional thermodynamic digital model in the storage device, binds the twin to the physical location, and generates a comprehensive data base; Simulation and Situation Analysis Module: Calls the data base to perform multiphysics coupling simulation, predicts the evolution of the internal environmental field, performs demand clustering analysis on the twin set, and generates a suitability map by calculating the temperature matching degree of each point in space; Collaborative planning module: Receives the suitability map and, using a rolling time-domain optimization method, jointly solves for the optimal sample location allocation scheme and equipment control command sequence; Execution and calibration control module: executes the sequence of device control commands to drive the coordinated action of the distributed actuator array, and corrects the internal parameters of the simplified three-dimensional thermodynamic digital model based on feedback data from the multi-sensor network.

[0021] In this embodiment, it is necessary to specifically explain the operation process of creating a sample digital twin in the data acquisition and modeling module as follows: The sample identifier is obtained by reading the RFID tag attached to the sample container. Based on the sample identifier, the corresponding sample metadata is retrieved from the associated laboratory information management system. The sample metadata includes at least the optimal activity maintenance temperature, the allowable temperature deviation threshold, the preset total settling time, and the sample priority weight. The typical setting of the allowable temperature deviation threshold is ±0.5 degrees Celsius, which is used to limit the acceptable temperature fluctuation range for maintaining sample activity. The sample priority weight is preset according to the clinical urgency or research value of the sample. For example, the weight of emergency samples is set to 1.0, while the weight of routine samples is set to 0.5. Based on sample metadata, a structured data object is instantiated in virtual space as a sample digital twin. This sample digital twin includes a sample identifier, optimal activity maintenance temperature, allowable temperature deviation threshold, preset total resting time, a real-time remaining resting time calculated based on the current time and the sample acquisition timestamp, and a sample priority weight. The specific instantiation process of the sample digital twin is as follows: a data structure containing specific fields is dynamically created in memory, and the sample identifier, optimal activity maintenance temperature, allowable temperature deviation threshold, preset total resting time, and sample priority weight are directly filled into the corresponding fields as initial attribute values. The value of the real-time remaining resting time field is obtained by subtracting the sample acquisition timestamp from the current system timestamp to get an already rested time, and then calculating it by subtracting the already rested time from the preset total resting time. A background timer is started to decrement the value of this real-time remaining resting time field at a fixed period. A derived parameter, namely the storage demand urgency, is introduced, which is jointly determined by the sample priority weight and the ratio of the real-time remaining idle time to the preset total idle time. The calculation process of the storage demand urgency is as follows: First, calculate the ratio of the real-time remaining idle time to the preset total idle time; then, set a decay coefficient greater than zero, with a typical value range between 0.5 and 2; then, calculate the exponential function value with the natural constant as the base and the product of the negative decay coefficient and the above ratio as the exponent; finally, multiply the calculated exponential function value by the sample priority weight, and the resulting product is the storage demand urgency. This calculation process makes the storage demand urgency increase nonlinearly and rapidly as the remaining idle time decreases, with higher priority samples having a higher urgency base, thus dynamically and quantitatively characterizing the real-time demand intensity of the samples for environmental stability. The process of initializing a simplified three-dimensional thermodynamic digital model is as follows: It continuously receives real-time monitoring data streams from a pre-deployed multi-sensor network inside the storage device. The multi-sensor network consists of temperature sensors, humidity sensors, and micro-airflow wind speed sensors uniformly distributed on three-dimensional spatial grid nodes inside the storage device. The specific process of receiving the real-time monitoring data stream is as follows: through a serial communication bus or wireless sensor network protocol, the output of each sensor in the multi-sensor network is polled or monitored at a fixed sampling period. The read analog or digital signals are converted and calibrated into temperature, humidity, and wind speed values ​​with standard units, and a unified timestamp and three-dimensional spatial coordinate label are added to form an environmental parameter data package with spatiotemporal attributes. Maintaining a discretized digital model, i.e., a three-dimensional thermodynamic simplified digital model, includes the internal geometry of the storage device, mesh generation, material thermal properties, and dynamic boundary conditions. The specific process of maintaining the discretized digital model is as follows: a three-dimensional array or voxel mesh data structure is established in computer memory, with its dimensions matching the number of meshes in the internal space of the storage device; each mesh cell stores material properties representing that location, including but not limited to the specific heat capacity and density of air, and the thermal conductivity of the surrounding walls; dynamic boundary conditions include the temperature of the evaporator surface, the flow rate and temperature of the air outlet, and the distribution of internal heat sources consisting of the sample container and the shelf. During initialization, the latest real-time monitoring data from the multi-sensor network is mapped to the corresponding grid nodes of the simplified 3D thermodynamic digital model using a spatial interpolation algorithm. This synchronizes the current state of the simplified 3D thermodynamic digital model with the real environmental state of the physical world. The specific process of the spatial interpolation algorithm is as follows: For each grid node in the simplified 3D thermodynamic digital model, a preset number of sensor nodes closest to the node are first located in the 3D space. Then, the environmental parameter values ​​of the grid node are calculated using the inverse distance weighting method. The calculation process of the inverse distance weighting method is as follows: The real-time monitoring data corresponding to each sensor node are weighted and summed using the reciprocal of the spatial Euclidean distance between the target grid node and each selected sensor node. The closer the sensor node is, the greater its data weight. This is used to calculate the estimated values ​​of the missing temperature, humidity, and wind speed at the grid node, thus completing the initialization and assignment of the spatial field of the entire digital model. The process of generating a comprehensive data base is as follows: When a sample container is placed in a physical shelf location of a storage device, the coordinate information of that physical shelf location is obtained. The specific process for obtaining the coordinate information is as follows: Each storage compartment of the physical shelf location is equipped with a unique QR code or near-field communication tag. When a sample container is placed in, the tag of the compartment is identified by a scanner or reader installed on the shelf. After decoding, its predefined three-dimensional spatial coordinates are obtained. The coordinate information is represented in the form of a floating-point array of length three, which corresponds to the X-axis, Y-axis and Z-axis positions in the internal coordinate system of the storage device, respectively. Perform a binding operation to associate and record the sample digital twin with coordinate information. The specific process of the binding operation is as follows: add a physical coordinate field to the data structure of the sample digital twin; assign the obtained coordinate information to this physical coordinate field, thereby logically establishing a one-to-one correspondence between the sample digital twin and a specific three-dimensional spatial location, and persistently store this association in the mapping table of the relational database. Finally, all created sample digital twins and their associated coordinate information, along with a snapshot of the current state of the synchronized 3D thermodynamic simplified digital model, are encapsulated into a unified data structure, namely the integrated data base. The encapsulation process involves creating an integrated data base data structure object containing three core members: the first member is a list storing references to all currently active sample digital twin objects; the second member is a list corresponding to the first list in order, storing the coordinate information associated with each sample digital twin; and the third member is a model state object storing the latest environmental parameter values ​​of all mesh nodes in the 3D thermodynamic simplified digital model after initialization. Finally, this integrated data base data structure object is serialized into a specific binary or text format for direct use by subsequent simulation and situation analysis modules. This data structure ensures strict logical and spatiotemporal alignment between sample requirement attributes, spatial location information, and physical environment state.

[0022] In this embodiment, it is specifically necessary to explain the operation process of performing multiphysics coupling simulation and demand clustering analysis in the simulation and situation analysis module as follows: The current state of the simplified three-dimensional thermodynamic digital model and the physical coordinates and thermal properties of all sample digital twins are extracted from the comprehensive data base. The specific meaning of the thermal properties is: the heat capacity coefficient per unit time, which is calculated based on the standard volume of the sample container and the specific heat capacity of blood, and is used to quantify the rate at which the sample releases or absorbs heat when it acts as a heat source or heat sink in the simulation. Using the current state as initial conditions and superimposing a dynamic heat source term formed by the set of physical coordinates associated with each sample digital twin, the coupled fluid motion and energy conservation equations are solved transiently using numerical methods. This allows for rolling prediction of the three-dimensional temperature and airflow velocity fields inside the storage device over multiple consecutive time slices, forming a predicted environmental field sequence containing temperature field data for each future time step. The specific process of the transient numerical solution is as follows: a pressure-velocity coupled solver based on the finite volume method is used to discretize the control equations including the turbulence model; within each time step, the equations are solved sequentially. The momentum equation is used to obtain the predicted velocity field. The pressure correction equation is solved to correct the velocity field and satisfy mass conservation. Finally, the energy equation is solved to update the temperature field. The boundary conditions are set as follows: the air supply outlet is a fixed velocity and temperature, the return air outlet is a pressure outlet, the equipment wall is an adiabatic or non-slip boundary, and the sample container surface is a convective heat transfer boundary calculated based on its thermal properties. Rolling prediction means that after completing the calculation of the current time step, the calculation result is used as the initial condition for the next time step, and so on, until the preset total future duration is covered, for example, the step size is one minute for the next thirty minutes. Simultaneously, based on the two attributes of optimal activity retention temperature and storage demand urgency contained in each sample digital twin in the sample digital twin set, density-based clustering analysis is performed to group samples with similar optimal activity retention temperatures and similar storage demand urgency into the same demand cluster. The specific process of density-based clustering analysis is as follows: First, each sample digital twin is regarded as a two-dimensional data point, with the horizontal axis representing the optimal activity retention temperature and the vertical axis representing the normalized storage demand urgency. Then, using a density clustering algorithm, with setting the neighborhood radius parameter and the minimum number of points parameter, high-density sample point regions separated by low-density regions are found, and all sample points in each high-density region are grouped into a demand cluster. The neighborhood radius parameter is usually set according to the typical value range of the optimal activity retention temperature and the normalized storage demand urgency, for example, the temperature neighborhood radius is 0.5 degrees Celsius, and the urgency neighborhood radius is 0.1 degrees Celsius. Subsequently, for each point within the storage device's internal space, based on its distance to the physical coordinates of each sample digital twin, the storage urgency of each sample digital twin, and the optimal activity retention temperature of each sample digital twin, a spatial demand temperature field representing the comprehensive expected temperature value of the location point is calculated. The specific process for calculating the spatial demand temperature field is as follows: For each node in the spatial grid model, the following steps are performed: First, calculate the Euclidean distance from the node to the physical coordinates of each sample digital twin; Second, calculate a spatial influence weight for each sample, which is equal to the storage urgency of the sample multiplied by a Gaussian kernel function value. The Gaussian kernel function uses the distance from the node to the sample as the independent variable and a preset influence decay length as the scale parameter; Third, multiply the optimal activity retention temperature of all samples by their corresponding spatial influence weights to obtain a weighted temperature value; Fourth, sum the weighted temperature values ​​of all samples and divide them by the sum of the spatial influence weights of all samples. The quotient obtained is the comprehensive expected temperature value of the spatial node. After traversing all grid nodes, the complete spatial demand temperature field is obtained. The process of generating a suitability map is as follows: For each future moment and each spatial location in the predicted environmental field sequence, the comprehensive expected temperature value of the corresponding location is obtained from the spatial demand temperature field. Based on the urgency of storage demand and the allowable temperature deviation threshold of each sample digital twin around the location, a matching tolerance characterizing the temperature tolerance range of the location is dynamically calculated. The specific process of dynamically calculating the matching tolerance is as follows: For a given spatial location, the following steps are performed: First, calculate the distance from the point to the physical coordinates of each sample digital twin; Second, using the same Gaussian kernel function and influence decay length as when calculating the spatial demand temperature field, calculate the spatial influence weight of each sample on the location; Third, multiply the allowable temperature deviation threshold of each sample by its spatial influence weight to obtain the weighted allowable deviation; Fourth, sum the weighted allowable deviations of all samples and divide by the sum of the spatial influence weights of all samples. The quotient is the dynamic matching tolerance of the location, which reflects the strict requirements of samples with high urgency and low allowable deviation for the temperature stability of the local area. The predicted temperature values ​​for future times and spatial locations recorded in the predicted environmental field sequence are compared with the comprehensive expected temperature value of the location. The deviation relative to the normalized result of the matching tolerance is mapped to a spatiotemporal temperature matching degree value between zero and one through a nonlinear decay function. The specific process for calculating the spatiotemporal temperature matching degree value is as follows: First, the absolute value of the difference between the predicted temperature value and the comprehensive expected temperature value is calculated to obtain the absolute temperature deviation. Next, the absolute temperature deviation is divided by the dynamic matching tolerance to obtain a normalized relative deviation. Finally, the normalized relative deviation is used as input to calculate the exponential function value with the natural constant as the base and the negative relative deviation as the exponent. This exponential function value is the spatiotemporal temperature matching degree value. When the predicted temperature equals the comprehensive expected temperature, the matching degree is a perfect one. When the predicted temperature deviates from the comprehensive expected temperature, the matching degree decays exponentially from one. When the relative deviation is equal to one, the matching degree is approximately 0.368, indicating a significant decrease in the matching degree. After performing this calculation for all spatiotemporal points, a four-dimensional data field with the same dimensions as the predicted environmental field sequence is obtained. This four-dimensional data field is the microenvironment suitability map. The closer the value is to one, the better the corresponding spatiotemporal point can meet the overall needs of the sample set. The data structure of the microenvironment suitability map is: a four-dimensional array, where the first three dimensions correspond to the spatial three-dimensional grid coordinates, the fourth dimension corresponds to the sequence number of the future time series, and each element of the array stores the spatiotemporal temperature matching degree value of the corresponding spatiotemporal point. This map presents the distribution of the ability of the internal microenvironment of the storage device to meet the overall needs of the sample in a numerical way over a period of time.

[0023] In this embodiment, it is necessary to specifically explain the operation process of constructing the optimization problem and decision variables in the collaborative planning module as follows: Using the received microenvironment suitability map as input, a rolling optimization time domain with a fixed future step size is set; the length of the rolling optimization time domain is usually set to ten to thirty time steps, and each time step corresponds to the time resolution of the simulation prediction, such as one minute. The first type of decision variable is defined as a binary decision matrix that encodes the sample location allocation scheme, representing the allocation relationship between each sample digital twin and each physical shelf location. The rows of this matrix correspond to the sample digital twins, and the columns correspond to the physical shelf locations. A value of one in the matrix element indicates that the corresponding sample is allocated to the corresponding shelf location, and a value of zero indicates that it is not allocated. The second type of decision variable is defined as a continuous decision vector sequence of all distributed actuators constituting the device control command sequence at each moment in the rolling optimization time domain; each vector in this sequence contains the set value of all controllable actuators at the same moment, such as the power percentage of heating elements in each area, the speed setting of each fan, and the angle value of each guide plate. The overall optimization objective function is the sum of two parts. The first part is the future adaptive benefit, which is obtained by weighted integration of the microenvironment suitability map in the spatial and temporal dimensions within the rolling optimization time domain. The weight of each spatial location point is proportional to the urgency of the storage demand of the digital twin of the sample located at that spatial location point. The specific process of calculating the future adaptive benefit is as follows: at each future time in the rolling optimization time domain, all grid nodes in the three-dimensional space inside the storage device are traversed. For each node, it is first determined whether the node is occupied by a sample according to the currently assumed sample location allocation scheme. If it is, the urgency of the storage demand of the digital twin of the sample is used as the weight value of the node. If no sample is occupied, the weight is zero. Then, the weight value of the node is multiplied by the spatiotemporal temperature matching degree value corresponding to this spatial node at this time, which is found in the microenvironment suitability map. Finally, the weighted matching degree values ​​of all nodes are summed and then summed over all future time points to obtain the first part of the benefit value. This calculation method ensures that the optimization process will prioritize ensuring the high environmental matching degree of the location of high-urgency samples in the future time. The second part is the control energy consumption cost, which is obtained by multiplying the square norm of the continuous decision vector sequence that constitutes the equipment control command sequence by a preset positive real number penalty coefficient. The aim is to minimize the energy consumption of the control action. The preset positive real number penalty coefficient typically ranges from 0.01 to 0.1, and is used to adjust the system's trade-off between "pursuing environmental adaptability" and "reducing control energy consumption". The final value of the overall optimization objective function is the first part of future adaptive benefits minus the second part of control energy consumption cost. The optimization objective is to maximize this objective function value. The process of finding the optimal solution is as follows: A decomposition-coordinated iterative optimization algorithm is used to solve the overall optimization objective function. Each iteration contains two alternating sub-optimization processes. A convergence threshold, such as 0.001, is preset to determine whether the optimization has terminated. The first sub-optimization process is location planning. Under the condition of a fixed continuous decision vector sequence constituting the current equipment control command sequence, an integer programming sub-problem is solved, with the binary decision matrix used to encode the sample location allocation scheme as the decision variable. This update of the binary decision matrix, i.e., the sample location allocation scheme, is the result of the sample location allocation scheme. The specific process of solving the location planning sub-problem is as follows: Given a fixed control command sequence, the contribution of placing each sample digital twin in each available physical shelf location to the future adaptive benefit component of the overall optimization objective function is evaluated. Then, using simulated annealing or tabu search algorithms, under the constraint of "one sample, one location; one location, at most one sample," the sample location allocation scheme that maximizes the future adaptive benefit component is searched. Essentially, this process finds storage locations with higher matching values ​​in the future microenvironment suitability map for samples with high storage urgency. The second sub-optimization process is control optimization. Under the condition of a fixed sample location allocation scheme, a convex quadratic programming subproblem is solved, with the continuous decision vector sequence constituting the equipment control command sequence as the decision variable. This update updates the continuous decision vector sequence constituting the equipment control command sequence. The specific process of solving the control optimization subproblem is as follows: Given a fixed sample location, the overall optimization objective function is expressed as a quadratic function of the control command vector sequence. Simultaneously, two types of linear constraints must be satisfied. The first type is the equipment physical constraint, meaning that the control command value of each actuator must be between its allowed minimum and maximum values, and the change in command between adjacent time points must not exceed the maximum rate of change. The second type is the system dynamic approximation constraint, meaning that a linear time-varying approximation model describing how the control command affects the temperature field evolution must be satisfied. This model links the control command vector with future temperature field changes (and thus with the microenvironment suitability map value), ensuring that the optimized command is physically realizable and effective. Then, a convex optimization solver such as the interior-point method is used to solve this convex quadratic programming problem to obtain the optimal control command vector sequence. Two sub-optimization processes are executed alternately until the change in the overall optimization objective function value is less than a preset threshold. The final jointly optimized optimal sample location allocation scheme and optimal equipment control command sequence are then output, and the equipment control command corresponding to the first moment in the rolling optimization time domain within the optimal equipment control command sequence is issued for execution. The specific process of the alternating iteration is as follows: First, a set of sample location allocation schemes is initialized; then, an iterative loop is entered. In each iteration, the control optimization sub-process is executed with the fixed position to obtain a new control command sequence; then, the location planning sub-process is executed with the new control command sequence to obtain a new location scheme; the overall optimization objective function value after the new iteration is calculated and compared with the value of the previous round. If the difference is less than a preset convergence threshold, convergence is determined, the iteration stops, and the current location scheme and command sequence are output as the optimal solution; if convergence is not achieved, the next iteration continues; finally, the control command corresponding to the first moment in the rolling optimization time domain, i.e., the command to be executed immediately, is extracted from the output optimal equipment control command sequence.

[0024] In this embodiment, it is necessary to specifically explain the operation process of executing control commands and collecting feedback data in the execution and calibration control module as follows: The system receives a sequence of device control commands and immediately parses the control command vector at the first moment in the sequence, converting it into a specific control signal that drives each independent execution unit in the distributed actuator array. This control signal is then converted into a physical quantity by a power driver, driving the actuators to work together to change the local microenvironment inside the storage device. Examples of the conversion relationship between specific control signals and physical quantities are as follows: For a local heating film, the control signal is the duty cycle of a pulse-width modulation signal, which is amplified by the power driver and converted into heating power in watts; for a variable frequency fan, the control signal is a digital frequency value, which is converted into AC frequency by the variable frequency driver, thereby controlling the fan speed in revolutions per minute; for micro-guide vanes, the control signal is the number of pulses, which is converted into the vane rotation angle by a stepper motor driver. Within a preset time period after the control command takes effect, real-time monitoring data from all sensors in the multi-sensor network are synchronously collected at a frequency higher than the control cycle. Timestamp alignment and outlier cleaning are performed on these feedback data to generate a measured environmental field data sequence that strictly corresponds to the output dimension of the simplified three-dimensional thermodynamic digital model and is time-synchronized. The preset time period is typically set to the duration of three to five control cycles in the future, for example, three to five minutes. The data acquisition frequency is typically five to ten times the control cycle frequency, for example, once per second. Timestamp alignment refers to uniformly correcting data collected from different nodes of the multi-sensor network with slight time differences to the same standard time point using interpolation methods. Outlier cleaning involves using a sliding window to detect the collected data; when a sensor reading changes abruptly within a short period and is significantly inconsistent with the readings of surrounding sensors, it is identified as an outlier and replaced with its historical average or interpolated value from adjacent sensors. The generated measured environmental field data sequence is a set of temperature, humidity, and wind speed data arranged in chronological order, corresponding one-to-one with the grid nodes of the simplified three-dimensional thermodynamic digital model. The procedure for correcting the internal parameters of a simplified three-dimensional thermodynamic digital model is as follows: Based on the measured environmental field data sequence, the predicted environmental field sequence obtained by simulation prediction using the current three-dimensional thermodynamic simplified digital model under the same control command input is compared, and the prediction error vector between the two is calculated. The calculation process of the prediction error vector is as follows: for each sensor node position, at the same time, the measured value in the measured environmental field data sequence of the node is subtracted from the model predicted value in the predicted environmental field sequence to obtain a difference; the differences of all sensor nodes at all effective comparison times are arranged in order into a one-dimensional column vector, which is the prediction error vector. Simultaneously, a parameter sensitivity matrix is ​​maintained, which describes the degree of influence of changes in various correctable physical parameters in the simplified three-dimensional thermodynamic digital model on the predicted values ​​corresponding to each measuring point in the sensor network. The maintenance process of the parameter sensitivity matrix is ​​as follows: First, a set of correctable physical parameters in the simplified three-dimensional thermodynamic digital model is determined, including but not limited to the convective heat transfer coefficient of the inner wall of the equipment, the surface convective heat transfer coefficient of the sample container, and the local resistance coefficient of the air duct; then, by performing adjoint solution or perturbation analysis on the simplified three-dimensional thermodynamic digital model, the change in predicted environmental parameters such as temperature and wind speed at each grid node (corresponding to the sensor measuring point) in the model is calculated when each correctable physical parameter undergoes a unit minute change, and these changes are organized into a matrix according to the relationship between the parameter and the measuring point position, which is the parameter sensitivity matrix; this matrix is ​​usually calculated and stored offline and directly called during online calibration; Based on the prediction error vector and parameter sensitivity matrix, and introducing a weight matrix positively correlated with the storage urgency of samples near the sensor location, a set of corrections for the internally correctable physical parameters of the simplified three-dimensional thermodynamic digital model is calculated using a least squares estimation algorithm with a regularization term. The specific calculation process of the least squares estimation algorithm with a regularization term is as follows: First, a weight matrix is ​​constructed. This matrix is ​​a diagonal matrix, where each element on the diagonal is equal to the weighted average of the storage urgency of all digital twin samples within a certain spatial range around the corresponding sensor node. The weighting method uses inverse distance weighting, which ensures that the sensor data in the high storage urgency sample area... The first step involves multiplying the parameter sensitivity matrix, weight matrix, and prediction error vector together to obtain an intermediate matrix. The second step involves multiplying the intermediate matrix by a regularization coefficient (typically between 0.001 and 0.1) to prevent numerical instability during matrix inversion. The third step involves inverting the matrix obtained in the third step. The fifth step involves multiplying the parameter sensitivity matrix, weight matrix, and prediction error vector together. The sixth step involves multiplying the inverse matrix obtained in the fourth step with the result vector obtained in the fifth step; the final result vector is the desired correction vector for all correctable physical parameters. Finally, the calculated correction values ​​are used to update the corresponding internal parameters of the simplified three-dimensional thermodynamic digital model, and the state of the simplified three-dimensional thermodynamic digital model is reinitialized using the latest measured environmental field data to complete the online calibration. The specific process of model state reinitialization is as follows: First, each component of the calculated correction vector is added to the current value of the corresponding correctable physical parameter of the simplified three-dimensional thermodynamic digital model to complete the parameter update; then, the latest collected and cleaned measured environmental field data (temperature and wind speed distribution) is used as the boundary and initial conditions to drive the updated simplified three-dimensional thermodynamic digital model to perform a fast steady-state or quasi-steady-state calculation without future predictions, so that the temperature and velocity field states inside the model are consistent with the latest measured state of the physical world, thereby providing a calibrated and state-synchronized accurate digital model for the next round of simulation prediction and decision planning.

[0025] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0026] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0027] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0028] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0029] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0030] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0031] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the invention's technical solutions and their equivalents, the invention also intends to include these modifications and variations.

Claims

1. A veterinary blood sample viability maintenance environment control system, characterized in that, Specifically, it includes: Data acquisition and modeling module: Reads sample identifiers and obtains corresponding metadata, creates digital twins of samples; Based on real-time data from a multi-sensor network, initializes a simplified three-dimensional thermodynamic digital model in the storage device, binds the twin to the physical location, and generates a comprehensive data base; Simulation and Situation Analysis Module: Calls the data base to perform multiphysics coupling simulation, predicts the evolution of the internal environmental field, performs demand clustering analysis on the twin set, and generates a suitability map by calculating the temperature matching degree of each point in space; Collaborative planning module: Receives the suitability map and, using a rolling time-domain optimization method, jointly solves for the optimal sample location allocation scheme and equipment control command sequence; Execution and calibration control module: executes the sequence of device control commands to drive the coordinated action of the distributed actuator array, and corrects the internal parameters of the simplified three-dimensional thermodynamic digital model based on feedback data from the multi-sensor network.

2. The veterinary blood sample viability maintenance environment control system according to claim 1, characterized in that: In the data acquisition and modeling module, the process of creating a sample digital twin is as follows: The sample identifier is obtained by reading the RFID tag attached to the sample container. Based on the sample identifier, the corresponding sample metadata is retrieved from the associated laboratory information management system. The sample metadata includes the optimal activity retention temperature, the allowable temperature deviation threshold, the preset total settling time, and the sample priority weight. Based on sample metadata, a structured data object is instantiated in the virtual space as a sample digital twin. The sample digital twin includes sample identifier, optimal activity holding temperature, allowable temperature deviation threshold, preset total settling time, real-time remaining settling time calculated based on the current time and sample collection timestamp, and sample priority weight. It also introduces a derived parameter, namely the urgency of storage demand, which is determined by the sample priority weight and the ratio of the real-time remaining idle time to the preset total idle time.

3. The veterinary blood sample viability maintenance environment control system according to claim 2, characterized in that: The process of initializing a simplified three-dimensional thermodynamic digital model is as follows: It continuously receives real-time monitoring data streams from a pre-deployed multi-sensor network inside the storage device. The multi-sensor network consists of temperature sensors, humidity sensors, and micro-airflow wind speed sensors that are uniformly distributed on three-dimensional spatial grid nodes inside the storage device. Maintain a discretized digital model that includes the internal geometry of the storage device, mesh generation, material thermal properties, and dynamic boundary conditions, i.e., a three-dimensional thermodynamic simplified digital model; During initialization, the latest real-time monitoring data from the multi-sensor network is mapped to the corresponding grid nodes of the simplified three-dimensional thermodynamic digital model through a spatial interpolation algorithm, thereby synchronizing the current state of the simplified three-dimensional thermodynamic digital model with the real environmental state of the physical world.

4. The veterinary blood sample viability maintenance environment control system according to claim 3, characterized in that: The process of generating the integrated data base is as follows: When the sample container is placed in the physical shelf location of the storage device, obtain the coordinate information of that physical shelf location; Perform the binding operation to associate and record the sample digital twin with the coordinate information; Finally, all the created sample digital twins and their associated coordinate information, along with the current state snapshot of the completed state-synchronized three-dimensional thermodynamic simplified digital model, are encapsulated into a unified data structure, namely the integrated data base.

5. The veterinary blood sample viability maintenance environment control system according to claim 4, characterized in that: In the simulation and situation analysis module, the operation process for performing multiphysics coupling simulation and demand clustering analysis is as follows: Extract the current state of the simplified 3D thermodynamic digital model from the integrated data base, as well as the physical coordinates and thermal properties of all sample digital twins; Using the current state as the initial condition, and superimposing a dynamic heat source term formed by the set of physical coordinates associated with each sample digital twin, the coupled fluid motion and energy conservation equations are solved transiently. The three-dimensional temperature field and airflow velocity field inside the storage device are predicted in rolling predictions for multiple consecutive time slices in the future, forming a predicted environmental field sequence containing temperature field data for each future moment. Meanwhile, based on the two attributes of optimal activity retention temperature and storage urgency contained in each sample digital twin in the sample digital twin set, density-based clustering analysis is performed to group samples with similar optimal activity retention temperature and similar storage urgency into the same demand cluster. Subsequently, for each point inside the storage device, based on its distance to the physical coordinates of each sample digital twin, the urgency of storage needs of each sample digital twin, and the optimal activity retention temperature of each sample digital twin, a spatial demand temperature field representing the comprehensive expected temperature value of the location point is calculated.

6. The veterinary blood sample viability maintenance environment control system according to claim 5, characterized in that: The process of generating the suitability map is as follows: For each future moment and each spatial location in the predicted environmental field sequence, the comprehensive expected temperature value of the corresponding location is obtained from the spatial demand temperature field. Based on the urgency of storage demand and the allowable temperature deviation threshold of each sample digital twin around the location, a matching tolerance characterizing the temperature tolerance range of the location is dynamically calculated. The predicted temperature values ​​of the corresponding future time and spatial location points recorded in the predicted environmental field sequence are compared with the comprehensive expected temperature value of the location point. The deviation relative to the normalized result of the matching tolerance is mapped to a spatiotemporal temperature matching degree value between zero and one through a nonlinear decay function. After completing this calculation for all spatiotemporal points, a four-dimensional data field with the same dimension as the predicted environmental field sequence is obtained. This four-dimensional data field is the microenvironment suitability map.

7. The veterinary blood sample viability maintenance environment control system according to claim 6, characterized in that: In the collaborative planning module, the process of constructing the optimization problem and decision variables is as follows: Using the received microenvironment suitability map as input, a rolling optimization time domain with a fixed future step size is defined; The first type of decision variable is defined as a binary decision matrix that encodes the sample location allocation scheme, representing the allocation relationship between each sample digital twin and each physical shelf location; The second type of decision variable is defined as a continuous decision vector sequence of all distributed actuators constituting the device control command sequence at each time step within the rolling optimization time domain; The overall optimization objective function is the sum of two parts. The first part is the future adaptive benefit, which is obtained by weighted integration of the microenvironment suitability map in the spatial and temporal dimensions within the rolling optimization time domain. The weight of each spatial location point is proportional to the urgency of the storage demand of the sample digital twin located at that spatial location point. The second part is the control energy consumption cost, which is obtained by multiplying the square norm of the continuous decision vector sequence that constitutes the equipment control command sequence by a preset positive real number penalty coefficient, aiming to minimize the energy consumption of control actions.

8. The veterinary blood sample viability maintenance environment control system according to claim 7, characterized in that: The process of finding the optimal solution is as follows: A decomposition-coordination iterative optimization algorithm is used to solve the overall optimization objective function. Each iteration contains two alternating sub-optimization processes. The first sub-optimization process is location planning. Under the condition of fixing the current continuous decision vector sequence that constitutes the equipment control command sequence, the binary decision matrix used to encode the sample location allocation scheme is updated by solving an integer programming sub-problem with the binary decision matrix used to encode the sample location allocation scheme as the decision variable. The second sub-optimization process is control optimization. Under the condition of fixing the current sample position allocation scheme, a convex quadratic programming subproblem is solved, with the continuous decision vector sequence constituting the equipment control command sequence as the decision variable. The continuous decision vector sequence constituting the equipment control command sequence is updated. The two sub-optimization processes are executed alternately until the change in the overall optimization objective function value is less than a preset threshold. The final jointly optimized optimal sample position allocation scheme and optimal equipment control command sequence are output, and the equipment control command corresponding to the first moment in the rolling optimization time domain of the optimal equipment control command sequence is issued for execution.

9. The veterinary blood sample viability maintenance environment control system according to claim 8, characterized in that: The execution and calibration control module performs the following operations: executing control commands and collecting feedback data. Receive the device control command sequence and immediately parse the control command vector of the first moment in the device control command sequence, convert it into a specific control signal to drive each independent execution unit in the distributed actuator array. The control signal is converted into a physical quantity by the power driver, driving each actuator to work together to change the local microenvironment inside the storage device. Within a preset time period after the control command takes effect, real-time monitoring data from all sensors in the multi-sensor network are collected synchronously at a frequency higher than the control cycle. Timestamp alignment and outlier cleaning are performed on these feedback data to generate a measured environmental field data sequence that corresponds to the output dimension of the three-dimensional thermodynamic simplified digital model and is time-synchronized.

10. The veterinary blood sample viability maintenance environment control system according to claim 9, characterized in that: The process of correcting the internal parameters of the simplified three-dimensional thermodynamic digital model is as follows: Based on the measured environmental field data sequence, the predicted environmental field sequence obtained by simulation prediction using the current three-dimensional thermodynamic simplified digital model under the same control command input is compared, and the prediction error vector between the two is calculated. At the same time, a parameter sensitivity matrix is ​​maintained, which describes the degree of influence of the changes of each correctable physical parameter in the three-dimensional thermodynamic simplified digital model on the predicted values ​​corresponding to each measuring point of the sensor network. Based on the prediction error vector and parameter sensitivity matrix, and by introducing a weight matrix that is positively correlated with the urgency of storage requirements for samples near the sensor location, a set of correction amounts for the internal correctable physical parameters of the simplified three-dimensional thermodynamic digital model is calculated using a least squares estimation algorithm with regularization terms. Finally, the calculated correction values ​​are used to update the corresponding internal parameters of the simplified three-dimensional thermodynamic digital model, and the state of the simplified three-dimensional thermodynamic digital model is reinitialized using the latest measured environmental field data, thus completing the online calibration.

Images