Intelligent control method for environmental parameters in industrial constant temperature workshops

Abstract

This invention relates to the field of physical control and regulation, specifically to an intelligent control method for environmental parameters in industrial constant-temperature workshops. It involves collecting production scheduling data and real-time operating power data from production equipment to construct a dynamic spatial distribution matrix of equipment heat sources. This matrix is ​​used as a feedforward input, and combined with multi-physics coupled partial differential equations of the workshop's temperature, airflow, and humidity fields, a spatiotemporal distribution parameter prediction model is established. Based on the model predictive control mechanism, the objective function is to minimize the weighted integral of the spatial temperature and humidity deviations of the spatial sensing nodes. This is used to solve for the setpoint sequences of wind speed, supply air temperature, and humidity at each air supply terminal, and then execute these sets. This method overcomes localized heat accumulation caused by equipment start-up and shutdown, suppresses dynamic overshoot in temperature and humidity regulation, reduces energy consumption for heat and cold offsetting, and improves the spatial uniformity and temporal stability of the temperature field.

Description

Technical Field

[0001] This invention relates to the field of physical control and regulation, specifically to an intelligent control method for environmental parameters in industrial constant temperature workshops. Background Technology

[0002] Current industrial constant-temperature workshop environmental control typically employs a closed-loop regulation circuit based on feedback from a single temperature sensor. In actual operation, the control system collects temperature data from the return air vent or a few representative locations within the workshop, treating the entire workshop as a single-capacity thermodynamic object with concentrated heat capacity parameters. When the sensor detects a temperature deviation from the set threshold, the controller outputs a control quantity according to proportional-integral-derivative (PID) arithmetic rules, adjusting the opening of the chilled water valve or the power of the heater to maintain a constant temperature. This control method treats the temperature field within the workshop as a uniformly distributed scalar, supplying global heating and cooling based solely on the arithmetic mean temperature deviation at a single point or a few nodes. It fails to consider the spatial differences in the heat sources of equipment within the workshop, the dynamic heat flow changes caused by equipment start-up and shutdown, or the distributed characteristics of the coupling between multiple physical quantities such as temperature, airflow, and humidity.

[0003] Existing technologies treat industrial constant temperature workshops as centralized parameter objects for single-loop closed-loop feedback regulation. When the spatial distribution of heat sources in the equipment changes dynamically, there are problems of regulation lag and local heat accumulation caused by uneven spatial distribution of heat flow disturbance and multi-field coupling of temperature and humidity airflow. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent control method for environmental parameters in industrial constant temperature workshops, which can solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] An intelligent control method for environmental parameters in an industrial constant-temperature workshop includes the following steps: collecting production schedule data and real-time operating power data of production equipment in the industrial constant-temperature workshop; constructing a dynamic spatial distribution matrix of equipment heat sources based on the production schedule data and the real-time operating power data; using the dynamic spatial distribution matrix of equipment heat sources as a feedforward input, and combining it with the multi-physics coupled partial differential equations of the temperature field, airflow field, and humidity field of the industrial constant-temperature workshop, establishing a spatiotemporal distribution parameter prediction model; based on the spatiotemporal distribution parameter prediction model and the model predictive control mechanism, minimizing the weighted integral of the spatial temperature deviation and spatial humidity deviation of multiple spatial sensing nodes in the industrial constant-temperature workshop as the objective function, solving for the wind speed setpoint sequence, air supply temperature setpoint sequence, and humidity setpoint sequence of each air supply terminal; and sending the wind speed setpoint sequence, the air supply temperature setpoint sequence, and the humidity setpoint sequence to the corresponding variable air volume terminal actuator and the surface cooler water valve actuator.

[0007] Preferably, the step of constructing the dynamic spatial distribution matrix of the heat sources of the production equipment includes: extracting the equipment location coordinates and equipment start-up and shutdown sequence from the production scheduling data, and extracting the instantaneous heat power value from the real-time operating power data; performing spatial mapping in a three-dimensional discrete grid of the industrial constant temperature workshop based on the equipment location coordinates to determine the grid node where each heat source is located; calculating the heat flux density vector of each grid node as a function of time based on the instantaneous heat power value and the equipment start-up and shutdown sequence; and diffusing the heat flux density vector to adjacent grid nodes by combining the air medium heat conduction attenuation factor of the industrial constant temperature workshop with the spatial relative position relationship between the grid nodes, thereby generating the dynamic spatial distribution matrix of the heat sources of the equipment that dynamically evolves with time and spatial coordinates.

[0008] Preferably, the step of establishing a spatiotemporal distribution parameter prediction model by combining the multi-physics field coupled partial differential equations of the temperature field, airflow field, and humidity field of the industrial constant temperature workshop includes: constructing a set of field quantity control equations with temperature variable, airflow velocity variable, and humidity variable as state vectors; introducing a buoyancy term driven by temperature gradient as the first coupling term between the airflow field and the temperature field in the set of field quantity control equations, introducing a humidity transport term caused by airflow convection as the second coupling term between the airflow field and the humidity field, and introducing a latent heat exchange term caused by evaporation and condensation as the third coupling term between the humidity field and the temperature field; using the air supply outlet position parameters and return air outlet position parameters of the industrial constant temperature workshop as boundary constraints of the set of field quantity control equations, and discretizing the set of field quantity control equations containing the first coupling term, the second coupling term, and the third coupling term using the finite volume method to generate the spatiotemporal distribution parameter prediction model.

[0009] Preferably, the step of minimizing the weighted integral of the spatial temperature deviation and spatial humidity deviation of multiple spatial sensing nodes in the industrial constant temperature workshop as the objective function includes: obtaining the three-dimensional spatial coordinates of each spatial sensing node in the industrial constant temperature workshop; calculating the spatial distance weighting coefficient between each spatial sensing node and the corresponding air supply terminal; setting the time decay weighting coefficient under different prediction step sizes based on the prediction time domain length of the model predictive control mechanism; multiplying the spatial temperature deviation and the spatial humidity deviation by the corresponding spatial distance weighting coefficient and the time decay weighting coefficient respectively, and then performing time domain integration to construct a comprehensive cost function that includes spatial distribution constraints and time domain change constraints as the objective function.

[0010] Preferably, the step of establishing a spatiotemporal distribution parameter prediction model by using the dynamic spatial distribution matrix of the equipment heat source as a feedforward input and combining it with the multiphysics coupled partial differential equations of the temperature field, airflow field, and humidity field of the industrial constant temperature workshop includes: mapping the heat flux density vector of each grid node in the dynamic spatial distribution matrix of the equipment heat source to a time-varying heat source term in the energy conservation equation of the multiphysics coupled partial differential equation; substituting the time-varying heat source term into the constant term on the right-hand side of the energy conservation equation, so that the time-varying heat source term directly participates in the transient calculation of the temperature field; and dynamically refreshing the spatial distribution value of the time-varying heat source term according to the time update frequency of the dynamic spatial distribution matrix of the equipment heat source, thereby forming the spatiotemporal distribution parameter prediction model triggered by real-time feedforward as the heat source changes.

[0011] Preferably, the steps of solving the wind speed setpoint sequence, air supply temperature setpoint sequence, and humidity setpoint sequence for each air supply terminal include: obtaining the extreme values ​​of air volume adjustment of the variable air volume terminal actuator and the extreme values ​​of valve opening of the surface cooler water valve actuator; converting the extreme values ​​of air volume adjustment and valve opening into the state space equation control input constraint boundary of the model predictive control mechanism; during the rolling optimization solution of the objective function, using a quadratic programming numerical solution process to search for the wind speed control vector, air supply temperature control vector, and humidity control vector within the control input constraint boundary that minimize the objective function; and truncating the wind speed control vector, air supply temperature control vector, and humidity control vector according to the control period to generate the wind speed setpoint sequence, air supply temperature setpoint sequence, and humidity setpoint sequence.

[0012] Preferably, the step of diffusing the heat flux density vector to adjacent grid nodes by combining the air medium heat conduction attenuation factor of the industrial constant temperature workshop with the spatial relative position relationship between the grid nodes includes: obtaining the spatial distribution information of physical obstacles in the three-dimensional model of the equipment layout in the industrial constant temperature workshop; calculating the air permeability between any two grid nodes based on the spatial distribution information of physical obstacles, and using the air permeability as the spatial correction coefficient of the air medium heat conduction attenuation factor; for grid node pairs with physical obstacles blocking them, reducing the diffusion intensity of the heat flux density vector using the spatial correction coefficient; for grid node pairs without obstruction, calculating the natural diffusion component of the heat flux density vector according to the distance attenuation function in the spatial relative position relationship, and superimposing the natural diffusion component with the diffusion component attenuated by obstruction to complete the asymmetric diffusion of the heat flux density vector to adjacent grid nodes.

[0013] Preferably, the step of discretizing the field control equations containing the first coupling term, the second coupling term, and the third coupling term using the finite volume method includes: constructing an interface interpolation scheme for the temperature variable, the airflow velocity variable, and the humidity variable at the three-dimensional discrete grid interface of an industrial constant temperature workshop; for the convective transport term in the first coupling term, the second coupling term, and the third coupling term, introducing an upwind difference scheme to calculate the interface physical flux in the interface interpolation scheme to suppress numerical oscillations at high Pecklet numbers; for the diffusion transport term in the field control equations, using a central difference scheme to calculate the gradient flux between adjacent grid nodes, and substituting the interface physical flux and the gradient flux into the integral conservation equation of the finite volume method to generate a fully implicit discretized linear equation system as the spatiotemporal distribution parameter prediction model.

[0014] Preferably, the step of calculating the spatial distance weight coefficient between each spatial sensing node and its corresponding air supply terminal includes: extracting the absolute values ​​of the spatial temperature deviation and the spatial humidity deviation of each spatial sensing node in real time; comparing the absolute values ​​with a preset deviation threshold, and marking the spatial sensing node as a heat accumulation disturbance node when the absolute value is greater than the preset deviation threshold; for the heat accumulation disturbance node, extracting the spatial path length between the heat accumulation disturbance node and its corresponding air supply terminal, mapping the spatial path length using an inverse proportional function to generate a basic spatial weight, and superimposing the absolute value as a correction multiplier on the basic spatial weight to generate an increased spatial distance weight coefficient, thereby increasing the spatial distribution constraint ratio of the heat accumulation disturbance node in the objective function.

[0015] Preferably, the step of dynamically refreshing the spatial distribution values ​​of the time-varying heat source terms based on the time update frequency of the dynamic spatial distribution matrix of the equipment heat sources includes: obtaining a first update cycle of the production schedule data and a second update cycle of the real-time operating power data, wherein the first update cycle is greater than the second update cycle; when the model solution time step of the spatiotemporal distribution parameter prediction model is less than the second update cycle, using a linear interpolation calculation process to perform time-dimensional densification interpolation on the real-time operating power data between two adjacent second update cycles; converting the densified interpolated real-time operating power data into the corresponding interpolated heat flux density, and combining it with the equipment start-stop status in the production schedule data to generate a sub-step level continuous time-varying heat source term; loading the continuous time-varying heat source term into the energy conservation equation step by step for solving.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0017] 1. This invention constructs a dynamic spatial distribution matrix of equipment heat sources by collecting production scheduling data and real-time operating power data. This matrix is ​​then fed forward to a multi-physics coupled partial differential equation to establish a spatiotemporal distribution parameter prediction model. The control sequence is solved by minimizing the weighted integral of spatial deviations across multiple sensor nodes in the objective function. The spatial location and time-varying intensity of thermal disturbances are obtained through the dynamic spatial distribution matrix of equipment heat sources. The spatiotemporal distribution parameter prediction model is used to deduce the coupled propagation process of heat flow in the temperature, airflow, and humidity fields. By feeding heat source changes forward to the control quantity calculation stage, this invention overcomes the adjustment lag caused by sudden changes in heat sources, suppresses local heat accumulation and dynamic overshoot during temperature and humidity regulation, and improves the spatial temperature uniformity and temporal stability of the constant-temperature workshop under dynamic heat source disturbances.

[0018] 2. Based on the equipment location coordinates and instantaneous thermal power values, the heat flux density vector is calculated. Asymmetric diffusion of the heat flux density vector is then performed, considering the shading relationship of physical obstacles. An upwind differential scheme is introduced into the multiphysics coupled partial differential equations to suppress numerical oscillations. Substep-scale densification interpolation with different update cycles is used to refresh the time-varying heat source terms. The spatial distance weighting coefficient is adjusted for heat accumulation disturbance nodes in the objective function. The solution is then obtained by combining the control input constraint boundary of the actuator. These methods recreate the uneven propagation physical process of heat flux within the workshop due to shading, improving the accuracy of the prediction model in extrapolating spatial asymmetric disturbances, avoiding calculation distortion under high convection intensity, improving the time matching degree between heat source interpolation updates and control cycles, reducing the coupling interference of local disturbed areas on the global temperature and humidity field, and ensuring the feasibility of control commands under the physical limit constraints of the actuator. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the overall implementation of the intelligent control of environmental parameters in an industrial constant temperature workshop according to the present invention.

[0020] Figure 2 This is a flowchart illustrating the construction of the dynamic spatial distribution matrix of the device heat source in this invention.

[0021] Figure 3 A flowchart for establishing the spatiotemporal distribution parameter prediction model of the present invention;

[0022] Figure 4 This is a flowchart of the finite volume method for discretizing the multiphysics coupling equations of the present invention.

[0023] Figure 5 This is a flowchart illustrating the objective function construction and control sequence solution of the present invention.

[0024] Figure 6 This is a flowchart of the dynamic refresh and time interpolation process for the time-varying heat source item in this invention. Detailed Implementation

[0025] refer to Figure 1 In one embodiment, the intelligent control method for environmental parameters in an industrial constant-temperature workshop is applied to the constant temperature and humidity control system of an electronic component manufacturing workshop. The workshop is 24 meters long, 18 meters wide, and 4.5 meters high, and houses 12 surface mount soldering machines, 8 chip testing machines, and 6 packaging machines. Sixteen variable air volume (VAV) air supply terminals and four centralized return air vents are evenly distributed on the workshop ceiling. Thirty-two integrated temperature and humidity sensors are installed at different heights and locations within the workshop. The control system establishes communication connections with the production management system, equipment operation monitoring system, and HVAC actuators via an industrial Ethernet network, with a data transmission latency of less than 100 milliseconds.

[0026] The control system first collects production scheduling data from the production equipment in the industrial constant-temperature workshop through the open interface of the production management system. This data includes task allocation information for each piece of equipment within the next 24 hours, equipment start-up and shutdown times, production batch numbers, and single-batch production durations. Simultaneously, the control system collects real-time operating power data from each piece of equipment through the equipment operation monitoring system. This real-time operating power data is updated with a 1-second sampling period and includes the equipment's three-phase active power, reactive power, and power factor. After data cleaning and format conversion, the collected production scheduling data and real-time operating power data are stored in the control system's real-time database. The data cleaning process includes removing outliers, filling in missing data, and aligning timestamps.

[0027] refer to Figure 2 Based on production scheduling data and real-time operating power data, the control system constructs a dynamic spatial distribution matrix of heat sources for production equipment. First, the control system extracts the unique identifier information of each production piece of equipment from the production scheduling data. Based on this unique identifier, it retrieves the corresponding equipment's location coordinates from the workshop equipment layout database. The equipment location coordinates are represented in a three-dimensional Cartesian coordinate system with the southwest corner of the workshop floor as the origin, and the unit is meters. Simultaneously, the control system extracts the equipment start-up and stop sequence from the production scheduling data. This sequence includes the start-up and stop times of each piece of equipment within the predicted future time domain. Finally, the control system extracts the instantaneous heat power value of each piece of equipment from the real-time operating power data. This instantaneous heat power value is calculated by multiplying the equipment's active power by its heat conversion coefficient. The heat conversion coefficient is pre-calibrated according to the equipment type and stored in the equipment parameter database.

[0028] The control system divides the three-dimensional space of the industrial constant-temperature workshop into a uniform discrete grid. The grid has a step size of 1.5 meters in the X, Y, and Z axes, resulting in 576 grid nodes in a 16×12×3 grid. Based on the position coordinates of each production device, the control system performs spatial mapping within the three-dimensional discrete grid to determine the grid nodes where each heat source is located. For larger production devices, multiple grid nodes they occupy are marked as heat source nodes, and the heat power value of each heat source node is proportionally distributed according to the volume distribution of the device. Based on the instantaneous heat power value of each device and the device start-up and shutdown sequence, the control system calculates the heat flux density vector of each grid node over time. The direction of the heat flux density vector is perpendicular to the device surface and points towards the interior of the workshop, and the magnitude of the heat flux density vector is equal to the device's heat power value divided by the device's heat dissipation surface area.

[0029] The control system combines the air medium heat conduction attenuation factor of the industrial constant temperature workshop with the spatial relative position relationship between grid nodes to diffuse the heat flux density vector to adjacent grid nodes. The air medium heat conduction attenuation factor is calculated in real time based on the temperature, pressure, and humidity of the air in the workshop, and its physical meaning is the attenuation ratio of heat flux density per unit distance. For any two adjacent grid nodes, the control system calculates the spatial relative position vector between them, and determines the directional component of heat flux diffusion based on the angle between the spatial relative position vector and the heat flux density vector. The diffusion component of the heat flux density vector in each direction is proportional to the cosine of the angle in that direction. After diffusion calculation, the control system generates a dynamic spatial distribution matrix of equipment heat sources that evolves dynamically with time and spatial coordinates. The dimension of this matrix is ​​N×M×T, where N and M are the number of nodes in the three-dimensional discrete grid in the horizontal and vertical directions, respectively, and T is the number of time steps in the prediction time domain.

[0030] refer to Figure 3 Using the dynamic spatial distribution matrix of the equipment's heat source as the feedforward input, the control system combines the multi-physics coupled partial differential equations of the temperature, airflow, and humidity fields in the industrial constant-temperature workshop to establish a spatiotemporal distribution parameter prediction model. The control system first constructs a set of field control equations with temperature, airflow velocity, and humidity variables as state vectors. The control equations for the temperature field are based on the law of conservation of energy, the control equations for the airflow field are based on the laws of conservation of mass and momentum, and the control equations for the humidity field are based on the law of conservation of mass.

[0031] In the control equations, the control system introduces a buoyancy term driven by the temperature gradient as the first coupling term between the airflow field and the temperature field. The magnitude of the buoyancy term is proportional to the product of the air density difference and the gravitational acceleration, with the air density difference caused by the temperature gradient. The control system also introduces a humidity transport term caused by airflow convection as the second coupling term between the airflow field and the humidity field. The magnitude of the humidity transport term is proportional to the dot product of the airflow velocity and the humidity gradient. Finally, the control system introduces a latent heat exchange term caused by evaporation and condensation as the third coupling term between the humidity field and the temperature field. The magnitude of the latent heat exchange term is proportional to the product of the latent heat of vaporization of water and the rate of change of humidity.

[0032] refer to Figure 4 The control system uses the location parameters of the supply and return air vents in the industrial constant-temperature workshop as boundary constraints for the control equations. The supply air vent boundary conditions include supply air velocity, supply air temperature, and supply air humidity, while the return air vent boundary conditions include return air velocity and return air pressure. For the workshop walls, floor, and ceiling, the control system employs either adiabatic boundary conditions or third-type boundary conditions, with the third type considering heat exchange between the walls and the external environment. For the control equations containing first, second, and third coupling terms, the control system uses the finite volume method for discretization, generating a spatiotemporal parameter prediction model. During the finite volume method discretization process, the control system uses each grid node as the center of the control volume, integrating the control equations within the control volume to obtain a discretized algebraic equation system.

[0033] refer to Figure 5 Based on a spatiotemporal distributed parameter prediction model and a model predictive control mechanism, the control system uses the minimization of the weighted integral of the spatial temperature and humidity deviations of multiple spatial sensing nodes within an industrial constant-temperature workshop as its objective function. This is used to solve for the setpoint sequences of wind speed, supply air temperature, and humidity at each air supply terminal. The prediction time domain of the model predictive control mechanism is set to 10 control cycles, and the control time domain is set to 3 control cycles, with each control cycle lasting 30 seconds. Within each control cycle, the control system first acquires the real-time temperature and humidity measurements of each spatial sensing node, and then calculates the spatial temperature and humidity deviations for each node. The spatial temperature deviation is the difference between the real-time temperature measurement and the setpoint temperature value, and the spatial humidity deviation is the difference between the real-time humidity measurement and the setpoint humidity value.

[0034] The control system acquires the three-dimensional spatial coordinates of each spatial sensor node within the industrial constant-temperature workshop and calculates the spatial distance weighting coefficient between each spatial sensor node and its corresponding air supply terminal. The corresponding air supply terminal is the one that has the greatest impact on the area where the sensor node is located. The correspondence is predetermined based on the workshop airflow organization simulation results and stored in the control system. The spatial distance weighting coefficient is inversely proportional to the spatial distance between the spatial sensor node and its corresponding air supply terminal. Simultaneously, based on the prediction time domain length of the model predictive control mechanism, the control system sets a time decay weighting coefficient for different prediction step sizes. The time decay weighting coefficient decays exponentially with the increase of the prediction step size.

[0035] The control system multiplies the spatial temperature deviation and spatial humidity deviation by their corresponding spatial distance weighting coefficients and time decay weighting coefficients, respectively, and then integrates them in the time domain to construct a comprehensive cost function that includes spatial distribution constraints and time-domain variation constraints as the objective function. The objective function also includes a penalty term for the rate of change of the control variable to avoid frequent adjustments by the actuators. The control system obtains the extreme values ​​of airflow regulation of the variable air volume terminal actuator and the extreme values ​​of valve opening of the surface cooler water valve actuator, and transforms these extreme values ​​into the control input constraint boundaries of the model predictive control mechanism's state-space equations. During the rolling optimization solution of the objective function, the control system uses a quadratic programming numerical solution process to search for the wind speed control vector, supply air temperature control vector, and humidity control vector within the control input constraint boundaries that minimize the objective function. The wind speed control vector, supply air temperature control vector, and humidity control vector are then truncated according to the control cycle to generate wind speed setpoint sequences, supply air temperature setpoint sequences, and humidity setpoint sequences.

[0036] The control system sends the setpoint sequences for wind speed, supply air temperature, and humidity to the corresponding variable air volume (VAV) terminal actuators and surface cooler water valve actuators. The VAV terminal actuators adjust the valve opening based on the wind speed setpoint to control the supply air volume; the surface cooler water valve actuators adjust the chilled water valve opening based on the supply air temperature setpoint to control the supply air temperature; and the humidifiers adjust the humidification amount based on the humidity setpoint to control the supply air humidity. At the start of the next control cycle, the control system repeats the above process, achieving closed-loop intelligent control of the environmental parameters in the industrial constant temperature workshop.

[0037] The table below shows the basic parameters and control cycle settings for an industrial constant temperature workshop. This table clarifies the basic parameter conditions for the operation of the control system and provides a basis for the parameter configuration of each module.

[0038] In this embodiment, the basic parameters and control cycle settings of the industrial constant temperature workshop are shown in Table 1.

[0039] Table 1 Basic Parameters and Control Cycle Settings for Industrial Constant Temperature Workshops

[0040] Parameter categories Parameter name Parameter value unit Workshop Dimensions length 24 rice Workshop Dimensions width 18 rice Workshop Dimensions high 4.5 rice Grid generation X-axis step size 1.5 rice Grid generation Y-axis step size 1.5 rice Grid generation Z-axis step size 1.5 rice Number of devices Patch soldering equipment 12 tower Number of devices Chip testing equipment 8 tower Number of devices Packaging equipment 6 tower Executive agency Variable air volume (VAV) air supply terminal 16 indivual Executive agency Centralized return air vent 4 indivual Sensor Node Temperature and humidity integrated node 32 indivual Control parameters Control cycle 30 Second Control parameters Prediction Time Domain 10 One cycle Control parameters Control Time Domain 3 One cycle Data collection Power sampling period 1 Second Data collection Production schedule update cycle 3600 Second

[0041] In this embodiment, a dynamic spatial distribution matrix of equipment heat sources is constructed by collecting production scheduling data and real-time operating power data. This matrix is ​​then fed forward to a multiphysics coupled partial differential equation to establish a spatiotemporal distribution parameter prediction model. The control sequence is solved by minimizing the weighted integral of the spatial deviation of multiple sensor nodes in the objective function. This process achieves distributed prediction and control of environmental parameters in an industrial constant-temperature workshop, considering the spatial distribution and time-varying characteristics of equipment heat sources, as well as the multiphysics coupling relationship between temperature, airflow, and humidity.

[0042] In a preferred embodiment, the control system further considers the impact of physical obstacles within the workshop on heat flow diffusion when constructing the dynamic spatial distribution matrix of equipment heat sources. The control system first obtains a three-dimensional model of the equipment layout within the industrial constant-temperature workshop from the workshop equipment layout database, and extracts the spatial distribution information of physical obstacles from the three-dimensional equipment layout model. Physical obstacles include the production equipment itself, material shelves, partition walls, and other objects that obstruct airflow. The control system maps the spatial distribution information of physical obstacles onto a three-dimensional discrete mesh, marking mesh nodes occupied by physical obstacles as impenetrable nodes.

[0043] Based on the spatial distribution information of physical obstacles, the control system calculates the air permeability between any two grid nodes. For any two adjacent grid nodes, if the line connecting them does not pass through any physical obstacle, the air permeability is 1; if the line connecting them passes through a physical obstacle, the air permeability is calculated based on the material properties and thickness of the obstacle. The air permeability ranges from 0 to 1, where 0 represents completely impenetrable and 1 represents completely permeable. The control system uses the air permeability as a spatial correction coefficient for the air medium thermal conduction attenuation factor, performing spatial correction on the air medium thermal conduction attenuation factor.

[0044] For mesh node pairs obstructed by physical obstacles, the control system uses a spatial correction factor to reduce the diffusion intensity of the heat flux density vector. The diffusion component of the heat flux density vector in the obstruction direction is equal to the diffusion component without obstruction multiplied by the spatial correction factor. For unobstructed mesh node pairs, the control system calculates the natural diffusion component of the heat flux density vector based on the distance attenuation function in the spatial relative position relationship. The distance attenuation function adopts a Gaussian function form, and its expression is:

[0045]

[0046] in, This is the distance decay function value. The spatial distance between two grid nodes. The characteristic length of heat diffusion. The value is determined in advance based on the airflow characteristics within the workshop.

[0047] The control system superimposes the natural diffusion component with the diffusion component attenuated by shading to complete the asymmetric diffusion of the heat flux density vector to adjacent grid nodes. The asymmetric diffusion process considers the obstruction of heat flux propagation by physical barriers, resulting in varying diffusion intensities of heat flux density in different directions, which better reflects the actual physical process of heat flux propagation within the workshop. After asymmetric diffusion calculation, the control system generates a more accurate dynamic spatial distribution matrix of equipment heat sources.

[0048] The control system maps the heat flux density vector of each grid node in the dynamic spatial distribution matrix of the equipment's heat sources to the time-varying heat source term of the energy conservation equation in the multiphysics coupled partial differential equations. The general form of the energy conservation equation is:

[0049]

[0050] in, air density, The specific heat capacity of air at constant pressure. For air temperature, For time, The airflow velocity vector The thermal conductivity of air. This represents the heat source intensity per unit volume.

[0051] refer to Figure 6 The control system substitutes the time-varying heat source term into the constant term on the right-hand side of the energy conservation equation, allowing the time-varying heat source term to directly participate in the transient calculation of the temperature field. During the iterative solution of the multiphysics coupled partial differential equations, the control system dynamically updates the spatial distribution values ​​of the time-varying heat source term based on the time update frequency of the equipment's dynamic spatial distribution matrix. The control system acquires a first update cycle for production scheduling data and a second update cycle for real-time operating power data, where the first update cycle is longer than the second update cycle. In this embodiment, the first update cycle is 3600 seconds, and the second update cycle is 1 second.

[0052] When the time step of the spatiotemporal distribution parameter prediction model is less than the second update cycle, the control system uses a linear interpolation process to perform time-dimensional densification interpolation on the real-time operating power data between two adjacent second update cycles. The formula for linear interpolation is:

[0053]

[0054] in, Interpolation time The device power value, and These are the time points of two adjacent second update cycles. and They are respectively and The device power measurement value at any given time.

[0055] The control system converts the densified interpolated real-time operating power data into corresponding interpolated heat flux density, and combines this with the equipment start / stop status in the production scheduling data to generate a sub-step-level continuous time-varying heat source term. For equipment marked as stopped in the production scheduling, its interpolated heat flux density is forcibly set to 0. The control system loads the continuous time-varying heat source term into the energy conservation equation step by step for solution, forming a spatiotemporal distribution parameter prediction model that is triggered by real-time feedforward as the heat source changes.

[0056] The table below shows examples of the thermal power characteristics and start-up / shutdown sequences of different types of production equipment. This table illustrates the changes in thermal power of different equipment at different production stages, providing specific data references for constructing the dynamic spatial distribution matrix of equipment heat sources.

[0057] In this embodiment, examples of the thermal power characteristics and start-up / shutdown sequences of different types of production equipment are shown in Table 2.

[0058] Table 2 Examples of Thermal Power Characteristics and Start-up / Shutdown Sequences for Different Types of Production Equipment

[0059] Equipment type Equipment Number Position coordinates (X, Y, Z) Standby thermal power Operating thermal power Startup time Stop time Single batch duration Patch soldering equipment SMT-01 (3,3,0.8) 0.5 3.2 08:00:00 12:00:00 30 Patch soldering equipment SMT-02 (3,6,0.8) 0.5 3.2 08:00:00 12:00:00 30 Patch soldering equipment SMT-03 (3,9,0.8) 0.5 3.2 09:00:00 13:00:00 30 Chip testing equipment TEST-01 (9,3,0.6) 0.3 1.8 08:30:00 17:30:00 60 Chip testing equipment TEST-02 (9,6,0.6) 0.3 1.8 08:30:00 17:30:00 60 Chip testing equipment TEST-03 (9,9,0.6) 0.3 1.8 10:00:00 18:00:00 60 Packaging equipment PACK-01 (15,3,0.7) 0.4 2.5 09:00:00 17:00:00 45 Packaging equipment PACK-02 (15,6,0.7) 0.4 2.5 09:00:00 17:00:00 45 Packaging equipment PACK-03 (15,9,0.7) 0.4 2.5 10:30:00 18:30:00 45

[0060] In this embodiment, by combining the physical obstacle shading relationship to perform asymmetric diffusion of the heat flux density vector, the physical process of uneven heat flow propagation under the influence of shading within the workshop is restored, improving the accuracy of the dynamic spatial distribution matrix of the equipment heat source. By employing a linear interpolation calculation process to densify the real-time operating power data in the time dimension, a sub-step-level continuous time-varying heat source term is generated, improving the time matching degree between heat source interpolation updates and control cycles, enabling the spatiotemporal distribution parameter prediction model to more accurately deduce the dynamic changes in heat flux.

[0061] In a preferred embodiment, the control system further optimizes the discretization scheme of the finite volume method when establishing the spatiotemporal distribution parameter prediction model to improve the stability and accuracy of numerical calculations. The control system constructs an interface interpolation scheme for temperature, airflow velocity, and humidity variables at the three-dimensional discrete grid interface of the industrial constant-temperature workshop. The interface interpolation scheme adopts a second-order upwind scheme, which calculates the physical quantity values ​​at the interface by utilizing the physical quantity values ​​of the two upstream grid nodes, thus exhibiting high computational accuracy.

[0062] For the convective transport terms in the first, second, and third coupling terms, the control system introduces an upwind differential scheme to calculate the interface physical flux within the interface interpolation lattice, in order to suppress numerical oscillations at high Peklay numbers. The Peklay number is defined as:

[0063]

[0064] in, For Pecklace numbers, The magnitude of the airflow velocity, The grid feature length is given. When the Peklayt number is greater than 2, the central difference scheme exhibits numerical oscillations, while the upwind difference scheme maintains the stability of numerical computation.

[0065] For the diffusion and transport terms in the field control equations, the control system employs a central difference scheme to calculate the gradient flux between adjacent grid nodes. The central difference scheme has second-order accuracy, enabling accurate calculation of the diffusion and transport process. The control system substitutes the interface physical flux and gradient flux into the integral conservation equations of the finite volume method, generating a fully implicit discretized linear equation system as a prediction model for spatiotemporal distribution parameters. The fully implicit discretized scheme exhibits unconditional stability, allowing for larger time steps and improving computational efficiency.

[0066] When solving the objective function, the control system further optimized the calculation method of the spatial distance weighting coefficient to improve the control effect on local heat accumulation areas. The control system extracts the absolute values ​​of the spatial temperature deviation and spatial humidity deviation of each spatial sensing node in real time and compares these absolute values ​​with preset deviation thresholds. These preset deviation thresholds are pre-set according to the workshop's process requirements; in this embodiment, the temperature deviation threshold is 0.5℃ and the humidity deviation threshold is 3%RH. When the absolute value of the spatial temperature deviation or spatial humidity deviation exceeds the preset deviation threshold, the control system marks the spatial sensing node as a heat accumulation disturbance node.

[0067] For nodes experiencing thermal accumulation disturbances, the control system extracts the spatial path length between the disturbed node and its corresponding air supply terminal. The spatial path length is the shortest airflow path from the air supply terminal to the sensing node, not the straight-line distance. The spatial path length is pre-calculated using workshop airflow organization simulation results and stored in the control system. The control system uses an inverse proportional function to map the spatial path length, generating basic spatial weights. The expression for the inverse proportional function is:

[0068]

[0069] in, Based on the weights of the space, The length of the grid feature.

[0070] The control system uses the absolute value of the spatial temperature deviation or spatial humidity deviation as a correction multiplier, superimposed on the base spatial weights to generate an increased spatial distance weight coefficient. The formula for calculating the spatial distance weight coefficient is:

[0071]

[0072] in, Spatial distance weighting coefficient, For correction factor, This refers to the spatial temperature deviation or spatial humidity deviation. Correction factor. The value of is predetermined based on the workshop's control requirements. In this embodiment, The value is 2.0.

[0073] By using the above methods, the control system increases the spatial distribution constraint weight of the thermal accumulation disturbed nodes in the objective function, so that when solving the control sequence, the control system can give priority to the temperature and humidity regulation needs of the thermal accumulation disturbed nodes and reduce the coupling interference of the local disturbed area on the global temperature and humidity field.

[0074] The control system acquires the extreme values ​​of airflow regulation of the variable air volume (VAV) terminal actuator and the extreme values ​​of valve opening of the cooler water valve actuator. The extreme values ​​of airflow regulation of the VAV terminal actuator include the minimum and maximum airflow, and the extreme values ​​of valve opening of the cooler water valve actuator include the minimum and maximum opening. The control system transforms these extreme values ​​of airflow regulation and valve opening into the state-space equations of the model predictive control mechanism, controlling the input constraints and boundaries. The general form of the state-space equations is:

[0075]

[0076]

[0077] in, For the first The state vector at time t, For the first The control input vector at time t, For the first The output vector at time step 1 Here is the state transition matrix. For the input matrix, This is the output matrix.

[0078] The expression for the control input constraint boundary is:

[0079]

[0080] in, To control the lower bound of the input vector, This is to control the upper limit of the input vector.

[0081] During the rolling optimization solution of the objective function, the control system utilizes a quadratic programming numerical solution process to search for wind speed control vectors, supply air temperature control vectors, and humidity control vectors within the control input constraint boundaries that minimize the objective function. The expression for the objective function is:

[0082]

[0083] in, The objective function value, To predict the length of the time domain, The number of spatial sensing nodes. For the first Spatial distance weighting coefficient of each sensor node For the first The time decay weighting coefficient for each prediction step size For the first The sensing node at the _th ... Temperature deviation per prediction step For the first The sensing node at the _th ... Humidity deviation per prediction step, This is the weighting factor for humidity deviation. To control the length of the time domain, The penalty coefficient is the rate of change of the control quantity. For the first The change in control quantity for each control step.

[0084] The quadratic programming problem is solved using the interior-point method, which converges to the optimal solution within a finite number of iterations. The control system extracts the obtained wind speed control vector, supply air temperature control vector, and humidity control vector according to the control period, generating wind speed setpoint sequences, supply air temperature setpoint sequences, and humidity setpoint sequences. The control system then sends these setpoint sequences to the corresponding variable air volume (VAV) terminal actuators and surface cooler water valve actuators, which adjust accordingly.

[0085] The table below shows an example of the location distribution and weight coefficient calculation of spatial sensing nodes. The table illustrates the calculation process of the spatial distance weight coefficient of sensing nodes at different locations, as well as the correction results of the weight coefficient of nodes affected by thermal accumulation.

[0086] In this embodiment, the spatial sensing node location distribution and weight coefficient calculation example are shown in Table 3.

[0087] Table 3. Example of spatial sensing node location distribution and weighting coefficient calculation

[0088] Node number Position coordinates (X, Y, Z) Corresponding air supply terminal Spatial path length Basic spatial weights Temperature deviation Deviation threshold Are you disturbed? Corrected weights S01 (1.5,1.5,1.5) VAV-01 2.1 0.227 0.2 0.5 no 0.227 S02 (4.5,1.5,1.5) VAV-02 2.3 0.189 0.3 0.5 no 0.189 S03 (7.5,1.5,1.5) VAV-03 2.5 0.160 0.7 0.5 yes 0.384 S04 (10.5,1.5,1.5) VAV-04 2.2 0.207 0.4 0.5 no 0.207 S05 (13.5,1.5,1.5) VAV-05 2.4 0.174 0.2 0.5 no 0.174 S06 (16.5,1.5,1.5) VAV-06 2.6 0.148 0.1 0.5 no 0.148 S07 (19.5,1.5,1.5) VAV-07 2.3 0.189 0.3 0.5 no 0.189 S08 (22.5,1.5,1.5) VAV-08 2.1 0.227 0.2 0.5 no 0.227

[0089] In this embodiment, by introducing an upwind difference scheme into the multiphysics coupled partial differential equations to suppress numerical oscillations, computational distortion under high convection intensity is avoided, thus improving the numerical stability and computational accuracy of the prediction model. By correcting the spatial distance weighting coefficients for the thermal accumulation disturbance nodes, the response speed and adjustment accuracy of the control system to local thermal accumulation regions are improved, and the coupling interference of local disturbance regions on the global temperature and humidity field is reduced. By combining the control input constraint boundary of the actuator for solution, the feasibility of control commands under the physical limit constraints of the actuator is ensured.

[0090] Furthermore, during operation, the control system continuously evaluates the prediction accuracy of the spatiotemporal distribution parameter prediction model. The control system compares the actual measured values ​​of each spatial sensor node with the model's predicted values ​​and calculates the prediction error. When the prediction error exceeds a preset threshold, the control system initiates an adaptive correction process for the model parameters. This adaptive correction process employs the recursive least squares method, using historical prediction error data to online correct the model's state transition matrix, input matrix, and output matrix to adapt to changes in workshop environmental conditions.

[0091] The formula for calculating recursive least squares is:

[0092]

[0093]

[0094]

[0095] in, For the first The estimated values ​​of the model parameters at time 10:00. For the first Gain matrix at time step For the first The actual measured value at time, For the first The regression vector at time step, For the first The covariance matrix at time t, Forgetting factor, The identity matrix. Forgetting factor. The value range is from 0.9 to 1.0. In this embodiment, The value is 0.98.

[0096] Through the adaptive correction process of model parameters, the control system can continuously maintain the prediction accuracy of the spatiotemporal distribution parameter prediction model and adapt to the influence of factors such as changes in equipment layout, personnel flow, and external environmental conditions within the workshop.

[0097] In a preferred embodiment, the control system also incorporates dynamic distribution prediction of personnel heat sources. The control system acquires real-time personnel location information through a personnel positioning system within the workshop and calculates the heat flux density vector of personnel heat sources by combining this information with the thermal radiation characteristics of the personnel. The heat flux density vector of personnel heat sources is proportional to the number of personnel and the intensity of their activity. The control system superimposes the heat flux density vector of personnel heat sources with the heat flux density vector of equipment heat sources to generate a comprehensive dynamic spatial distribution matrix of heat sources including personnel heat sources. This comprehensive dynamic spatial distribution matrix of heat sources is used as a feedforward input to the spatiotemporal distribution parameter prediction model, further improving the model's accuracy in predicting thermal disturbances within the workshop.

[0098] The formula for calculating the heat power of a heat source from personnel is:

[0099]

[0100] in, Total thermal power of personnel For the number of personnel, Basal metabolic heat power of personnel, This is the activity intensity correction factor. Activity intensity level. Basal metabolic rate of energy. The value is 100W, which is the activity intensity correction factor. The value is 0.5, representing the activity intensity level. The personnel are determined in advance based on the nature of their work.

[0101] In this embodiment, by introducing dynamic distribution prediction of personnel heat sources, the impact of personnel activities on the temperature and humidity field of the workshop is considered, further improving the accuracy and stability of environmental parameter control in industrial constant temperature workshops.

[0102] It should be noted that the above embodiments are merely illustrative examples. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical principles of the present invention, and these modifications and variations should also be considered within the scope of protection of the present invention. For example, the step size of the three-dimensional discrete mesh can be adjusted according to the size of the workshop and the control accuracy requirements; the prediction time domain length and control time domain length of the model predictive control can be adjusted according to the system response speed requirements; the weight coefficients in the objective function can be adjusted according to the process requirements of the workshop.

[0103] In a preferred embodiment, the control system employs a distributed computing architecture, distributing the solution process of the spatiotemporal distributed parameter prediction model and the optimization solution process of model predictive control to multiple computing nodes for parallel execution. The distributed computing architecture adopts a master-slave structure, with the master node responsible for data acquisition, task allocation, and result aggregation, while the slave nodes are responsible for specific numerical calculation tasks. Through this distributed computing architecture, the control system can significantly improve computing speed, meeting the real-time computing requirements of large-scale workshops and high-density grid partitioning.

[0104] In the distributed computing process, the master node divides the 3D discrete mesh into multiple sub-regions, and each sub-region is assigned to a slave node for computation. After completing the computation task in its respective sub-region, the slave node sends the computation results to the master node. The master node summarizes the computation results from all slave nodes and performs boundary processing to obtain the predicted temperature and humidity field for the entire workshop space. During boundary processing, the master node interpolates and coordinates the boundary data of adjacent sub-regions to ensure the continuity and consistency of the computation results.

[0105] In this embodiment, by adopting a distributed computing architecture, the computing power and real-time performance of the control system are improved, enabling the method to be applied to larger-scale and more complex industrial constant temperature workshop environmental control.

[0106] Furthermore, the control system also possesses fault diagnosis and fault-tolerant control functions. The control system monitors the operating status of each spatial sensor node, variable air volume (VAV) terminal actuator, and surface cooler water valve actuator in real time. When a fault is detected in a sensor node, the control system automatically removes the measurement data from that faulty node and estimates the temperature and humidity values ​​at the faulty node's location using spatial interpolation methods based on the measurement data from adjacent normal nodes. When a fault is detected in an actuator, the control system automatically adjusts the control inputs of other normal actuators to compensate for the impact of the faulty actuator, ensuring the stability of the temperature and humidity field in the workshop.

[0107] The spatial interpolation method employed is Kriging interpolation, which can provide optimal unbiased estimation based on the correlation of spatial data. The formula for Kriging interpolation is as follows:

[0108]

[0109] in, Points to be interpolated The estimated value, For the first Measurement values ​​of known points, For the first The weight coefficients of the known points The number of known points. Weighting coefficients. By solving the Kriging equations, we can see that the Kriging equations take into account the semivariance function characteristics of spatial data.

[0110] In this embodiment, the reliability and robustness of the control system are improved through fault diagnosis and fault-tolerant control functions, ensuring that the industrial constant temperature workshop can still operate normally in the event of partial equipment failure.

[0111] In a preferred embodiment, the control system is also integrated with the workshop's energy management system to achieve synergistic optimization of environmental control and energy consumption. When solving the control sequence, in addition to minimizing the weighted integral of temperature and humidity deviations, the control system also incorporates the energy consumption of the HVAC system as an additional term into the objective function. By adjusting the weighting coefficients of the energy consumption term in the objective function, the control system can optimize energy consumption while meeting the workshop's temperature and humidity process requirements.

[0112] The objective function expression that includes the energy consumption term is:

[0113]

[0114] in, The weighting coefficient for energy consumption. For the first Energy consumption value for each control step. It is calculated based on the air volume, air temperature and humidity of each air supply terminal, as well as the operating characteristics of the chiller, water pump and fan.

[0115] In this embodiment, by integrating with an energy management system, the environmental control and energy consumption of the industrial constant temperature workshop are optimized in a coordinated manner, thereby reducing the operating energy consumption of the HVAC system while ensuring production process requirements are met.

Claims

1. A method for intelligent control of environmental parameters in an industrial constant-temperature workshop, characterized in that, Includes the following steps: Collect production schedule data and real-time operating power data of production equipment in industrial constant temperature workshops; Based on the production schedule data and the real-time operating power data, a dynamic spatial distribution matrix of the heat sources of the production equipment is constructed. Using the dynamic spatial distribution matrix of the heat source of the equipment as the feedforward input, and combining the multi-physics field coupled partial differential equations of the temperature field, airflow field and humidity field of the industrial constant temperature workshop, a spatiotemporal distribution parameter prediction model is established. Based on the spatiotemporal distribution parameter prediction model and model prediction control mechanism, the objective function is to minimize the weighted integral of the spatial temperature deviation and spatial humidity deviation of multiple spatial sensing nodes in the industrial constant temperature workshop, and to solve the wind speed setpoint sequence, air supply temperature setpoint sequence and humidity setpoint sequence of each air supply terminal. The wind speed setpoint sequence, the air supply temperature setpoint sequence, and the humidity setpoint sequence are sent to the corresponding variable air volume terminal actuators and the surface cooler water valve actuators.

2. The intelligent control method for environmental parameters in an industrial constant-temperature workshop according to claim 1, characterized in that, The steps for constructing the dynamic spatial distribution matrix of the heat sources of the production equipment include: Extract the equipment location coordinates and equipment start-up and shutdown sequence from the production scheduling data, and extract the instantaneous thermal power value from the real-time operating power data; Based on the equipment location coordinates, spatial mapping is performed in the three-dimensional discrete grid of the industrial constant temperature workshop to determine the grid nodes where each heat source is located. Based on the instantaneous thermal power value and the device start-up and shutdown sequence, calculate the heat flux density vector of each grid node as a function of time; By combining the thermal conductivity attenuation factor of the air medium in the industrial constant temperature workshop with the spatial relative position relationship between the grid nodes, the heat flux density vector is diffused to adjacent grid nodes to generate the dynamic spatial distribution matrix of the equipment heat source that evolves dynamically with time and spatial coordinates.

3. The intelligent control method for environmental parameters in an industrial constant-temperature workshop according to claim 1, characterized in that, The steps for establishing a spatiotemporal distribution parameter prediction model by combining the multiphysics coupled partial differential equations of the temperature field, airflow field, and humidity field in an industrial constant-temperature workshop include: Construct a set of field control equations with temperature, airflow velocity, and humidity variables as state vectors; In the field control equation set, a buoyancy term driven by temperature gradient is introduced as the first coupling term between the airflow field and the temperature field, a humidity transport term caused by airflow convection is introduced as the second coupling term between the airflow field and the humidity field, and a latent heat exchange term caused by evaporation and condensation is introduced as the third coupling term between the humidity field and the temperature field. The location parameters of the air supply outlet and return air outlet in the industrial constant temperature workshop are used as the boundary constraints of the field quantity control equation set. The field quantity control equation set, which includes the first coupling term, the second coupling term and the third coupling term, is discretized by the finite volume method to generate the spatiotemporal distribution parameter prediction model.

4. The intelligent control method for environmental parameters in an industrial constant-temperature workshop according to claim 1, characterized in that, The steps, with the objective function being the minimization of the weighted integral of the spatial temperature and humidity deviations at multiple spatial sensing nodes within an industrial constant-temperature workshop, include: Obtain the three-dimensional spatial coordinates of each of the spatial sensing nodes in the industrial constant temperature workshop, and calculate the spatial distance weighting coefficient between each of the spatial sensing nodes and the corresponding air supply terminal. Based on the prediction time domain length of the model prediction control mechanism, time decay weight coefficients are set for different prediction step sizes; The spatial temperature deviation and the spatial humidity deviation are multiplied by the corresponding spatial distance weighting coefficient and the time decay weighting coefficient, respectively, and then integrated in the time domain to construct a comprehensive cost function that includes spatial distribution constraints and time domain change constraints as the objective function.

5. The intelligent control method for environmental parameters in an industrial constant-temperature workshop according to claim 1, characterized in that, The steps for establishing a spatiotemporal distribution parameter prediction model by using the dynamic spatial distribution matrix of the equipment heat source as a feedforward input and combining it with the multi-physics coupled partial differential equations of the temperature field, airflow field, and humidity field of the industrial constant temperature workshop include: The heat flux density vector of each grid node in the dynamic spatial distribution matrix of the device heat source is mapped to the time-varying heat source term of the energy conservation equation in the multiphysics coupled partial differential equation. Substituting the time-varying heat source term into the constant term on the right-hand side of the energy conservation equation allows the time-varying heat source term to directly participate in the transient calculation of the temperature field. During the iterative solution of the multiphysics coupled partial differential equations, the spatial distribution values ​​of the time-varying heat source terms are dynamically refreshed based on the time update frequency of the dynamic spatial distribution matrix of the heat source of the equipment, forming the spatiotemporal distribution parameter prediction model that is triggered by real-time feedforward as the heat source changes.

6. The intelligent control method for environmental parameters in an industrial constant-temperature workshop according to claim 1, characterized in that, The steps for solving the sequence of wind speed setpoints, supply air temperature setpoints, and humidity setpoints for each air supply terminal include: Obtain the extreme value of air volume regulation of the variable air volume terminal actuator and the extreme value of valve opening of the surface cooler water valve actuator; The extreme values ​​of air volume regulation and valve opening are transformed into the state space equations of the model predictive control mechanism to control the input constraint boundaries. In the rolling optimization solution of the objective function, the wind speed control vector, supply air temperature control vector, and humidity control vector are searched within the control input constraint boundary using a quadratic programming numerical solution process to minimize the objective function. The wind speed control vector, the supply air temperature control vector, and the humidity control vector are truncated according to the control period to generate the wind speed setpoint sequence, the supply air temperature setpoint sequence, and the humidity setpoint sequence.

7. The intelligent control method for environmental parameters in an industrial constant-temperature workshop according to claim 2, characterized in that, The step of diffusing the heat flux density vector to adjacent grid nodes, based on the air medium heat conduction attenuation factor in an industrial constant-temperature workshop and the spatial relative positional relationship between the grid nodes, includes: Obtain spatial distribution information of physical obstacles in a 3D model of equipment layout within an industrial constant temperature workshop; Based on the spatial distribution information of the physical obstacles, the air permeability between any two grid nodes is calculated, and the air permeability is used as the spatial correction coefficient of the air medium heat conduction attenuation factor. For mesh node pairs that are blocked by the physical obstacles, the diffusion intensity of the heat flux density vector is reduced using the spatial correction coefficient. For unobstructed grid node pairs, the natural diffusion component of the heat flux density vector is calculated based on the distance attenuation function in the spatial relative position relationship. The natural diffusion component is then superimposed with the diffusion component attenuated by shading to complete the asymmetric diffusion of the heat flux density vector to adjacent grid nodes.

8. The intelligent control method for environmental parameters in an industrial constant-temperature workshop according to claim 3, characterized in that, The steps of discretizing the system of field control equations containing the first coupling term, the second coupling term, and the third coupling term using the finite volume method include: At the three-dimensional discrete mesh interface of the industrial constant temperature workshop, an interface interpolation format for the temperature variable, the airflow velocity variable and the humidity variable is constructed. For the convective transport term in the first coupling term, the second coupling term, and the third coupling term, an upwind differential scheme is introduced to calculate the interface physical flux in the interface interpolation lattice, so as to suppress numerical oscillations at high Peklay numbers. For the diffusion transport term in the field control equations, the gradient flux between adjacent grid nodes is calculated using the central difference scheme. The interface physical flux and the gradient flux are substituted into the integral conservation equation of the finite volume method to generate a fully implicit discretized linear equation system as the spatiotemporal distribution parameter prediction model.

9. The intelligent control method for environmental parameters in an industrial constant-temperature workshop according to claim 4, characterized in that, The step of calculating the spatial distance weighting coefficient between each of the aforementioned spatial sensing nodes and the corresponding air supply terminal includes: The absolute values ​​of the spatial temperature deviation and the spatial humidity deviation of each of the spatial sensing nodes are extracted in real time. The absolute value is compared with a preset deviation threshold. When the absolute value is greater than the preset deviation threshold, the spatial sensing node is marked as a thermal accumulation disturbance node. For the disturbed node with thermal accumulation, the spatial path length between the disturbed node with thermal accumulation and the corresponding air supply terminal is extracted. The spatial path length is mapped using an inverse proportional function to generate a basic spatial weight. The absolute value is then superimposed on the basic spatial weight as a correction multiplier to generate an increased spatial distance weight coefficient, thereby increasing the spatial distribution constraint ratio of the disturbed node with thermal accumulation in the objective function.

10. The intelligent control method for environmental parameters in an industrial constant-temperature workshop according to claim 5, characterized in that, The step of dynamically updating the spatial distribution values ​​of the time-varying heat source items based on the time update frequency of the device heat source dynamic spatial distribution matrix includes: The first update cycle for obtaining the production schedule data and the second update cycle for obtaining the real-time operating power data are wherein the first update cycle is longer than the second update cycle. When the model solution time step of the spatiotemporal distribution parameter prediction model is less than the second update cycle, a linear interpolation calculation process is used to perform time-dimensional densification interpolation on the real-time operating power data between two adjacent second update cycles. The real-time operating power data after densification and interpolation is converted into the corresponding interpolated heat flux density, and combined with the equipment start-up and shutdown status in the production scheduling data, a sub-step level continuous time-varying heat source item is generated. The continuous time-varying heat source term is applied to the energy conservation equation step by step for solution.

Images