Intelligent warehouse temperature and humidity control method and system

Abstract

The present application relates to temperature and humidity control technical field, specifically to a kind of intelligent granary temperature and humidity control method and system, comprising: collection temperature and humidity data and projection to controlled state mapping space generates transient heat and humidity balance term;Extract the cross interference component and associated feedback component between temperature and humidity adjustment branch to construct mutual interference coupling matrix group, perform variable decoupling operation to generate decoupling target control vector.Simultaneously utilize dense space point array temperature and humidity data to calculate point deviation and construct field gradient tensor group, combined with controlled medium physical parameters to generate mass-energy migration trend flow, to decoupling target control vector Space reconstruction is executed, and field collaborative control flow is generated.Finally, according to the operation data of composite environment regulating mechanism and instantaneous contribution rate, operating condition sensitivity is generated, and composite load constraint term is determined accordingly, and collaborative execution response flow is output through weight distribution.The present application eliminates temperature and humidity mutual interference, realizes the accurate suppression of controlled physical quantity field distribution and energy efficiency balance.

Description

Technical Field

[0001] This invention relates to the field of temperature and humidity control technology, specifically to an intelligent method and system for temperature and humidity control in grain warehouses. Background Technology

[0002] In the field of monitoring and regulating industrial controlled environments, achieving coordinated stability of multiple non-electrical variables such as temperature and humidity is the core task of precision control systems.

[0003] Currently, various collaborative control schemes for such controlled physical quantities have emerged in the industry. For example, existing control methods are mainly based on the multi-loop feedback principle, which collects environmental parameters inside and outside the controlled space and uses preset logic thresholds to drive the actuator to perform compensation switching of a single variable; subsequent improved schemes have introduced correlation analysis of the physical properties of the controlled object and use multi-source sensor signals to assist in adjusting the sampling period in order to improve the system's sensitivity to fluctuations in environmental variables.

[0004] However, the above-mentioned control methods still have significant shortcomings in the application of complex physical fields regarding the cross-interference and coordinated mitigation between variables. On the one hand, existing control logics often treat temperature and humidity as independent controlled loops, failing to fully consider the strong coupling interaction between heat and moisture inside the controlled medium at the physical level. This makes it easy to cause fluctuations in one variable when adjusting a single variable, resulting in frequent adjustment oscillations and control overshoot near the stable point of the system. On the other hand, existing systems lack a coordinated compensation mechanism in the scheduling of actuators, failing to effectively integrate the influence of different actuators such as air conditioning, ventilation, and dehumidification on the distribution of the physical field. This makes it difficult to achieve precise mitigation and energy efficiency balance of the distribution of controlled physical quantities under complex meteorological backgrounds.

[0005] In summary, existing control systems require a precise control method that can address cross-interference of non-electrical variables, coordinated suppression of physical quantities, and joint control of actuators.

[0006] Therefore, this invention proposes an intelligent method and system for controlling temperature and humidity in grain warehouses. Summary of the Invention

[0007] The purpose of this invention is to provide an intelligent method and system for controlling temperature and humidity in grain warehouses, so as to achieve precise suppression and energy efficiency balance of the distribution of controlled physical quantities.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] An intelligent method for controlling temperature and humidity in grain warehouses includes:

[0010] Temperature and humidity data within the controlled field are collected and mapped to a controlled state mapping space with moisture content as the horizontal axis and enthalpy as the vertical axis, generating a transient heat and humidity balance term. Based on the transient heat and humidity balance term, the cross-interference component of the temperature regulation branch on the humidity variable and the correlation feedback component of the humidity regulation branch on the temperature variable are extracted, generating a mutual interference coupling matrix group. The transient heat and humidity balance term and the mutual interference coupling matrix group are subjected to variable decoupling operation to generate a decoupled target control vector.

[0011] Acquire dense spatial lattice temperature and humidity data, calculate the parameter deviation between points based on the decoupled target control vector, and construct a field gradient tensor set; based on the field gradient tensor set and the physical parameters of the controlled medium, determine the spatial diffusion path and generate a mass-energy migration trend flow; use the mass-energy migration trend flow to perform spatial reconstruction on the decoupled target control vector and generate a field cooperative control flow.

[0012] The system acquires the operating data of the composite environmental control mechanism and calculates the instantaneous contribution rate of the controlled physical quantity based on the field-domain collaborative control flow to generate the operating condition sensitivity. It then determines the composite load constraint terms based on the operating condition sensitivity and uses the composite load constraint terms to perform weight allocation on the field-domain collaborative control flow, outputting the collaborative execution response flow.

[0013] Preferably, the specific process of generating the transient thermal and humidity balance term includes: acquiring temperature and humidity data of points within the controlled field; mapping the temperature and humidity data to a controlled state mapping space with moisture content as the horizontal axis and enthalpy as the vertical axis; locating the spatial transient state point generated by mapping the temperature and humidity data in the controlled state mapping space; forming an offset vector based on the spatial transient state point and the corresponding equilibrium steady-state point, and calculating the instantaneous thermodynamic gradient to generate the transient thermal and humidity balance term.

[0014] Preferably, the specific process of generating the mutual interference coupling matrix group includes: extracting the influence weight of the temperature regulation branch on the humidity variable based on the transient thermal humidity balance term, and establishing the cross-interference component; extracting the feedback gain of the humidity regulation branch on the temperature variable, and establishing the associated feedback component; filling the cross-interference component and the associated feedback component as matrix elements into an array structure of a preset dimension, performing matrix structure assembly, and generating the mutual interference coupling matrix group.

[0015] Preferably, the specific process of generating the decoupled target control vector includes: using the mutual interference coupling matrix group as an analytical operator to perform matrix inverse analytical operation on the transient thermal and humidity balance term; extracting mutually independent temperature control components and humidity control components from the data stream after performing the matrix inverse analytical operation; and performing vectorization synthesis on the temperature control components and the humidity control components to generate the decoupled target control vector.

[0016] Preferably, the specific process of constructing the field gradient tensor set and generating the mass-energy migration trend flow includes: acquiring dense spatial lattice temperature and humidity data, using the decoupled target control vector as the point control reference, and calculating the spatial deviation value of the dense spatial lattice temperature and humidity data relative to the point control reference; performing gradient analysis on the spatial deviation value along the three-dimensional coordinate axis to construct the field gradient tensor set; extracting the potential energy driving direction of the controlled physical quantity in the controlled field based on the field gradient tensor set, wherein the controlled physical quantity includes the field temperature component and the field humidity component; acquiring physical parameters of the controlled medium, including the medium thermal conductivity, medium porosity, and water permeability coefficient, analyzing the flow direction of the controlled physical quantity according to the physical parameters and the potential energy driving direction, and establishing the spatial diffusion path; and generating the mass-energy migration trend flow according to the time evolution trend of the controlled physical quantity in the spatial diffusion path.

[0017] Preferably, the specific process of generating the field cooperative control flow includes: acquiring field geometric dimension data containing length, width, and height components within the controlled field; using the field geometric dimension data to establish the physical boundary coordinates of the controlled field and construct a controlled three-dimensional physical space; projecting the decoupled target control vector onto the controlled three-dimensional physical space to construct a field target distribution benchmark; using the mass-energy migration trend flow to analyze the instantaneous diffusion deviation of controlled physical quantities within the controlled three-dimensional physical space and generate a spatial regulation compensation term; and using the spatial regulation compensation term to perform superposition and reconstruction on the field target distribution benchmark to generate the field cooperative control flow.

[0018] Preferably, the specific process of generating the operating condition sensitivity includes: acquiring operating data of the composite environmental control mechanism, including frequency and power consumption; analyzing the dynamic intervention ratio of the operating data on the controlled physical quantity during changes, and establishing the instantaneous contribution rate of the controlled physical quantity, based on the field collaborative control flow; using the instantaneous contribution rate, establishing the equipment action response intensity of the composite environmental control mechanism to the field collaborative control flow, and generating the operating condition sensitivity; the composite environmental control mechanism includes a temperature control mechanism for performing heat exchange, a ventilation mechanism for performing gas migration within the controlled field, and a dehumidification mechanism for performing moisture content regulation.

[0019] Preferably, the specific process of determining the composite load constraint term and outputting the collaborative execution response flow includes: taking the time for the controlled physical quantity to return to steady state and the overall operating power consumption as the balance target, and defining the control execution capability boundary based on the sensitivity of operating conditions to establish the composite load constraint term; using the composite load constraint term to allocate the execution weight of the field collaborative control flow, and outputting the collaborative execution response flow.

[0020] An intelligent grain warehouse temperature and humidity control system includes:

[0021] The thermal and humidity decoupling analysis module acquires temperature and humidity data of the controlled field, temperature and humidity data of the dense spatial lattice, and operating data of the composite environmental regulation mechanism; maps the temperature and humidity data to the controlled state mapping space to generate transient thermal and humidity balance terms; establishes a mutual interference coupling matrix group through the transient thermal and humidity balance terms; performs variable decoupling operations; and outputs the decoupling target control vector.

[0022] The field gradient characterization module constructs a field gradient tensor group using the decoupled target control vector and the dense spatial lattice temperature and humidity data, establishes the mass-energy migration trend flow based on the physical parameters of the controlled medium, performs spatial reconstruction on the decoupled target control vector, and outputs the field cooperative control flow.

[0023] The collaborative execution decision module uses the operating data of the composite environmental regulation mechanism and the field collaborative control flow to establish the sensitivity of the operating conditions, determine the composite load constraint terms, perform weight allocation on the field collaborative control flow through the composite load constraint terms, and output the collaborative execution response flow.

[0024] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0025] 1. This invention performs temperature, humidity, and thermo-mass coupling collaborative analysis to map sampled data to the controlled state mapping space, generating transient thermo-humidity balance terms. It then utilizes the mutual interference coupling matrix group to perform a substantial coordination design for variable decoupling. This correlation method achieves precise isolation of cross-interference between controlled physical quantities. In the strongly coupled physical field scenario of a grain pile, it solves the problem of global physical field instability caused by the adjustment of a single variable, eliminates adjustment oscillations and control overshoot during the control process, and improves the control accuracy of independent controlled variables when the physical state deviates.

[0026] 2. This invention combines a decoupled control vector group with a distributed field gradient tensor group by performing a field distribution gradient equilibrium representation, and introduces physical parameters including the thermal conductivity, porosity, and permeability of the medium to construct a mass-energy transfer characterization flow. The spatial dimension of the decoupled control vector group is reconstructed using the coherent coordination between the aforementioned physical quantities. In grain storage scenarios with large spatial spans and microporous media characteristics, this effectively solves the problems of nonlinear distortion and control lag in the spatial distribution of physical quantities, achieving precise smoothing from local sampling point control to global physical field equilibrium distribution.

[0027] 3. This invention deeply integrates the field-wide collaborative control flow with the operational condition sensitivity of the composite environmental regulation mechanism through energy-efficiency coordinated steady-state closed-loop control, and constructs a parallel strategy scheduling for the execution of composite load constraints. This cross-dimensional collaborative processing method achieves a dynamic balance between execution resources and environmental demands. In application scenarios where the performance of the actuators is limited and external weather disturbances are frequent, it ensures a nonlinear balance between the response rate of the controlled physical quantity returning to steady state and the overall operating power consumption, avoids regulation conflicts between heterogeneous actuators, and improves the overall energy efficiency level. Attached Figure Description

[0028] Figure 1 This is a flowchart of an intelligent grain warehouse temperature and humidity control method according to the present invention;

[0029] Figure 2 This is a flowchart illustrating the field collaborative control reconfiguration process according to an embodiment of the present invention.

[0030] Figure 3 This is a schematic diagram of the structure of an intelligent grain warehouse temperature and humidity control system according to the present invention. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Please see Figures 1 to 3 This invention provides an intelligent method and system for controlling temperature and humidity in grain warehouses, the technical solution of which is as follows:

[0033] Example 1

[0034] Reference Figure 1 This is a flowchart of an intelligent grain warehouse temperature and humidity control method according to the present invention. This embodiment provides a specific application scenario for the intelligent grain warehouse temperature and humidity control method. The method is applied in a controlled field of grain storage. It eliminates cross-interference of temperature and humidity through thermal and humidity decoupling, uses gradient tensor sets to smooth spatial distribution deviations, and combines operational condition sensitivity to achieve energy efficiency synergy of the regulating mechanism. This solves the problems of control lag and energy consumption imbalance in the controlled field, and realizes precise steady-state control of the controlled physical quantities. The specific steps are as follows:

[0035] Temperature and humidity data within the controlled field are collected and mapped to a controlled state mapping space with moisture content as the horizontal axis and enthalpy as the vertical axis, generating a transient heat and humidity balance term. Based on the transient heat and humidity balance term, the cross-interference component of the temperature regulation branch on the humidity variable and the correlation feedback component of the humidity regulation branch on the temperature variable are extracted, generating a mutual interference coupling matrix group. The transient heat and humidity balance term and the mutual interference coupling matrix group are subjected to variable decoupling operation to generate a decoupled target control vector.

[0036] Acquire dense spatial lattice temperature and humidity data, calculate the parameter deviation between points based on the decoupled target control vector, and construct a field gradient tensor set; based on the field gradient tensor set and the physical parameters of the controlled medium, determine the spatial diffusion path and generate a mass-energy migration trend flow; use the mass-energy migration trend flow to perform spatial reconstruction on the decoupled target control vector and generate a field cooperative control flow.

[0037] The system acquires the operating data of the composite environmental control mechanism and calculates the instantaneous contribution rate of the controlled physical quantity based on the field-domain collaborative control flow to generate the operating condition sensitivity. It then determines the composite load constraint terms based on the operating condition sensitivity and uses the composite load constraint terms to perform weight allocation on the field-domain collaborative control flow, outputting the collaborative execution response flow.

[0038] Furthermore, the specific process of generating the transient thermal and humidity balance term includes: acquiring temperature and humidity data of points within the controlled field; mapping the temperature and humidity data to a controlled state mapping space with moisture content as the horizontal axis and enthalpy as the vertical axis; locating the spatial transient state point generated by mapping the temperature and humidity data in the controlled state mapping space; forming an offset vector based on the spatial transient state point and the corresponding equilibrium steady-state point, and calculating the instantaneous thermodynamic gradient to generate the transient thermal and humidity balance term.

[0039] Specifically, temperature and humidity data are acquired from a lattice of sensors within the controlled area of ​​the grain storage facility. This temperature and humidity data are then mapped to a two-dimensional thermodynamic coordinate space defined by moisture content on the horizontal axis and enthalpy on the vertical axis; this two-dimensional thermodynamic coordinate space is the controlled state mapping space.

[0040] The mapping process is as follows: using temperature and humidity data acquired by sensors, combined with the current atmospheric pressure, the moisture content and enthalpy of the corresponding point are calculated using aerodynamic formulas. Specifically, the process includes: first, based on the instantaneous temperature of the point, determining or calculating the saturation pressure baseline for the air's capacity to hold water vapor at that temperature. This baseline represents the moisture carrying capacity limit of the controlled point at the current temperature. Real-time atmospheric pressure is introduced as an environmental background parameter, and the partial pressure of water vapor is established using the product of relative humidity and the saturation pressure baseline. Subsequently, based on the ratio of the partial pressure of water vapor divided by the real-time atmospheric pressure, the mass of water vapor contained in a unit mass of dry air is calculated, thus obtaining the moisture content of the corresponding point. In this step, the introduction of real-time atmospheric pressure is used to correct for the influence of different altitudes or meteorological fluctuations on the moisture molar ratio.

[0041] The increase in sensible heat of dry air with increasing temperature is summed with the latent heat of vaporization of water vapor corresponding to the moisture content and the heat of superheated steam with increasing temperature to obtain the total heat content, i.e., enthalpy value, at the corresponding point.

[0042] The data is mapped to a controlled state mapping space with moisture content as the horizontal axis and enthalpy as the vertical axis.

[0043] In the controlled state mapping space, the coordinate projection points of temperature and humidity data are used as spatial transient state points. A preset equilibrium steady-state point for safe grain storage is calibrated. This equilibrium steady-state point is based on the saturated adsorption equilibrium state of the corresponding grain at safe storage moisture and target standard low temperature (e.g., 15°C-20°C). Its coordinate range within the controlled state mapping space is typically limited to a physical range with a moisture content of 0.008 kg / kg to 0.012 kg / kg and an enthalpy of 35 kJ / kg to 45 kJ / kg, representing the physical state coordinates of the controlled field reaching the steady-state target. The coordinates of the spatial transient state points are calculated using analytical geometry methods. Pointing to the coordinates of the equilibrium steady-state point Construct a spatial displacement trajectory (directed line segment) to create an offset vector with physical directionality. .

[0044] The slope of the energy level difference change corresponding to the offset vector within the controlled state mapping space is calculated to establish the instantaneous thermodynamic gradient. This is achieved by calculating the ratio of the vector's x and y coordinates. (in, Determine the slope of the energy level difference change. The slope, in physical terms, is defined as the instantaneous thermodynamic gradient, which reflects the intensity of the energy (enthalpy) change caused by a unit change in moisture during the transition of a controlled field from a transient to a steady state. A pre-defined extreme range of the slope in the thermodynamic space is set, ranging from 500 kJ / kg to 10000 kJ / kg; the lower limit corresponds to the extreme condition of isenthalpy change with fluctuating high moisture content, while the upper limit corresponds to the extreme condition of drastic sensible heat change under dry air conditions.

[0045] When the absolute value of the instantaneous thermodynamic gradient is greater than the preset first energy slope threshold, the offset vector approaches the vertical axis (enthalpy axis), indicating that the environmental deviation is mainly caused by sensible heat (temperature). It is determined that the current environmental deviation is primarily due to sensible heat exchange, and the temperature control actuator is adjusted first. When the absolute value of the instantaneous thermodynamic gradient is less than the preset second energy slope threshold, the offset vector approaches the horizontal axis (humidity axis), indicating that the deviation is mainly caused by latent heat (moisture). It is determined that the current environmental deviation is primarily due to latent heat exchange, and the dehumidification or humidification actuator is adjusted first. When the second energy slope threshold is less than or equal to the absolute value of the instantaneous thermodynamic gradient and less than or equal to the first energy slope threshold, it is determined to be a thermo-humidity coupled state, requiring the temperature control and dehumidification mechanisms to be activated simultaneously and power allocated proportionally. The first and second energy slope thresholds are preset based on the thermo-humidity balance characteristic curve of grain storage. In this embodiment, the range of the first energy slope threshold is... The range of the second energy slope threshold is: .

[0046] By integrating the above computational parameters, a transient thermal and moisture balance term is generated to characterize the degree to which the controlled environment deviates from steady state.

[0047] The specific method for integrating the above-mentioned calculation parameters is as follows: the magnitude of the offset vector is... As intensity information, the slope The normalized value is used as the gain coefficient. The specific normalization process is as follows: The extreme range of the slope in the thermodynamic space is pre-defined, and the calculated absolute value of the slope is mapped to a dimensionless interval between 0 and 1. When the slope approaches the maximum value on the vertical axis, the normalized gain coefficient approaches 1, representing an extreme bias towards sensible heat; when the slope approaches the minimum value on the horizontal axis, the gain coefficient approaches 0, representing a bias towards latent heat. The data structure of the transient heat and humidity balance term is defined as follows: ,in The direction angle of the offset vector. This balance term, as a composite control vector, is directly mapped to the power output function of the grain storage air conditioning system and the dehumidification system. The output ratio under the total control intensity is dynamically determined by the magnitude of the gain coefficient, thereby achieving decoupled control of the heat and humidity load.

[0048] This invention provides a method to transform temperature and humidity scalars into offset vectors with spatial orientation properties using thermodynamic spatial coordinate projection technology. This enables the quantitative characterization of the deviation state of the controlled environment and provides a physically directional decision-making basis for subsequent precise decoupling control.

[0049] Furthermore, the specific process of generating the mutual interference coupling matrix group includes: extracting the influence weight of the temperature regulation branch on the humidity variable based on the transient thermal humidity balance term, and establishing the cross-interference component; extracting the feedback gain of the humidity regulation branch on the temperature variable, and establishing the associated feedback component; filling the cross-interference component and the associated feedback component as matrix elements into an array structure of a preset dimension, performing matrix structure assembly, and generating the mutual interference coupling matrix group.

[0050] Specifically, the transient thermal and moisture balance terms generated within the controlled field are obtained.

[0051] The rate of change of the moisture content component over time in the transient heat and humidity balance term is used to analyze the humidity fluctuations caused by the temperature regulation branch when performing regulation actions (such as changes in air conditioning cooling power). Specifically, a sliding window period (e.g., 5 to 10 minutes) is established as the sampling window, and the infinitesimal increment ΔP of the temperature regulation branch power within this window is simultaneously acquired. T The infinitesimal increment ΔP of the humidity regulation branch power H And the instantaneous change in moisture content in the controlled state mapping space (the algebraic difference between the current value and the previous sampling value of the physical quantity in the controlled field within the sliding window time period). In order to isolate the independent contribution of a single branch under the condition of simultaneous closed loop of dual loops, a partial gain identification operator is introduced.

[0052] The partial gain identification operator is defined as a dynamic stripping operator based on time-domain correlation identification, and its calculation formula is: Partial gain identification operator = (Instantaneous change - Humidity branch self-gain component coefficient × Incremental ΔP) H Temperature increment ΔP T In order to eliminate the identification error caused by the large lag characteristic of the grain warehouse, the incremental micro-element in the formula is a historical power value after physical response delay correction from the corresponding actuator to the sensing point, so as to ensure that the data is physically aligned on the time axis.

[0053] The humidity branch self-gain component coefficient represents the nominal intervention intensity of a unit change in dehumidification power on the humidity content of the field. Its initial value is established by dividing the rated dehumidification rate of the dehumidification mechanism under standard operating conditions by the volume of the controlled space, or by measuring it through a single-variable step response experiment performed under the silent state of the temperature control loop. To address the coefficient deviation caused by changes in operating conditions during actual dynamic operation, a dynamic calibration operator (a mathematical operator characterizing the efficiency attenuation / enhancement ratio of the actuator under actual operating conditions relative to standard operating conditions) is introduced and product-fitted with the initial self-gain component coefficient to obtain the instantaneous self-gain coefficient. This instantaneous self-gain coefficient is used to calibrate the theoretical adjustment increment under non-standard operating conditions, thereby eliminating the stripping error caused by equipment performance degradation or changes in environmental resistance in the partial gain identification calculation.

[0054] By subtracting the nominal contribution of the humidity branch itself from the total change, the proportion of humidity deviation caused purely by changes in temperature regulation power is isolated. The nominal contribution is defined as the theoretical regulation of the actuator under uninterrupted operation, and is calculated by multiplying the increment of the instantaneous operating frequency or power consumption by a preset self-gain component coefficient. This self-gain component coefficient is a nominal constant obtained by shutting down other coupling branches, operating only a single actuator, and recording the rate of change of environmental physical quantities during the commissioning phase.

[0055] Based on this, the humidity deviation after separation is divided by the corresponding change in power of the temperature regulation branch. The resulting ratio is the influence weight of the humidity variable, which is used to establish the cross-interference component. The larger the value of this component, the stronger the coupling interference of temperature on humidity, thus providing quantitative support for the dynamic reconstruction of the cross-interference coupling matrix group.

[0056] The energy state information reflected by the transient thermal-humidity balance term is extracted synchronously, and the associated interference on the energy state of the controlled field when the humidity regulation branch performs regulation actions (such as changing the operating frequency of the dehumidifier) ​​is analyzed. By calculating the ratio of the change in enthalpy in the controlled field to the change in power of the humidity regulation branch, the energy feedback ratio caused by a unit humidity regulation action is extracted and identified as the feedback gain of the temperature variable, thereby establishing the associated feedback component.

[0057] The cross-interference components and the associated feedback components are used as matrix elements and filled into an array structure of a preset dimension. A matrix-based structure is then constructed to generate a mutual interference coupling matrix group. Specifically, the array structure in the form of a second-order square matrix of the preset dimension is as follows: the first row of the array structure corresponds to the temperature control dimension, and the second row corresponds to the humidity control dimension.

[0058] The mutual interference coupling matrix group is a second-order square matrix with the temperature control loop and the humidity control loop as its dimensions. Its specific structure is defined as follows: the first row corresponds to the change response of the temperature controlled variable, the second row corresponds to the change response of the humidity controlled variable; the first column corresponds to the control input of the temperature regulation branch, and the second column corresponds to the control input of the humidity regulation branch.

[0059] The specific assembly and filling logic is as follows: First, determine the main loop gain component at the diagonal position of the array, write the preset temperature branch self-gain component coefficient into the index bit [1,1], write the humidity branch self-gain component coefficient into the index bit [2,2], and construct the basic diagonal support of the matrix.

[0060] Subsequently, the mapping and filling of cross-interference elements are performed: the established cross-interference components (characterizing the influence of the temperature branch on the humidity variable) are written into index [2,1], representing the interference of the first column input on the second row output; the established correlation feedback components (characterizing the influence of the humidity branch on the temperature variable) are written into index [1,2], representing the interference of the second column input on the first row output.

[0061] During the writing process, normalization scaling is performed to eliminate dimensional differences. Specifically, the full-scale power references for the temperature and humidity actuators are retrieved, and the calculated component coefficients are compared with the corresponding branch's full-scale reference to convert them into dimensionless percentage gain weights. This process ensures that all elements in the matrix are of the same order of magnitude, ultimately forming a set of mutual interference coupling matrices capable of quantifying the thermal-humidity coupling relationship within the controlled field.

[0062] This invention transforms the physical correlation between temperature and humidity into explicit matrix parameters, thereby achieving structured quantification of the cross-interference relationship of physical quantities. It solves the problem of inaccurate control caused by cross-loop interference and provides quantitative support for the precise decoupling of controlled variables.

[0063] Furthermore, the specific process of generating the decoupled target control vector includes: using the mutual interference coupling matrix group as an analytical operator to perform matrix inverse analytical operation on the transient thermal and humidity balance term; extracting mutually independent temperature control components and humidity control components from the data stream after performing the matrix inverse analytical operation; and performing vectorization synthesis on the temperature control components and the humidity control components to generate the decoupled target control vector.

[0064] Specifically, the mutual interference coupling matrix set and the transient thermal-humidity balance term are obtained. Using the mutual interference coupling matrix set as an analytical operator, a matrix inverse analytical operation is performed on the transient thermal-humidity balance term.

[0065] The specific process of performing matrix inverse analytical operations includes: calculating the inverse matrix of the mutual perturbation coupling matrix, and using the algebraic cofactor method or Gaussian elimination method to calculate the inverse matrix of the mutual perturbation coupling matrix. This is used as an analytical operator. First, the determinant of the mutual interference coupling matrix is ​​calculated. If the absolute value of the determinant is within a preset threshold range (10 in this embodiment), then... -5 Up to 10 -8 If the matrix is ​​close to a singular state, there is a risk that the inversion will not converge or the numerical solution will oscillate significantly.

[0066] If the determinant of the matrix is ​​at a preset singular value (e.g., 10) −5 Within the range of ), a perturbation factor is introduced to the diagonal components to ensure the existence of the inverse matrix. The perturbation factor is a compensation term, which is usually a very small positive value added to the diagonal of the matrix, ranging from 0.01% to 0.1% of the original value of the diagonal elements. Its specific value is: the absolute value of the original value of the current diagonal element multiplied by a preset perturbation coefficient (0.05% in this embodiment); if the product result is less than 10 -6 The forced correction is 10. -6 This ensures that the matrix determinant deviates significantly from the singular interval.

[0067] Analytical operators are used to perform matrix multiplication on the transient thermo-humidity balance terms, resulting in a two-dimensional column vector, defined as the decoupling control vector. Its data structure is [G T G H ] T The superscript T is a matrix operator used to convert row vectors [G] into matrix vectors. T G H Convert to a column vector, where the subscript T represents the temperature dimension, and the first element G... T The temperature control gain signal after decoupling correction, the second element G H This is the humidity control gain signal after decoupling correction. This process aims to eliminate crosstalk between temperature and humidity loops, transforming coupled temperature and humidity deviations into independent control gain requirements.

[0068] Subsequently, from the calculated two-dimensional decoupled control vector, the first row of elements is extracted as an independent temperature control component, and the second row of elements is extracted as an independent humidity control component. The temperature control component and the humidity control component are then vectorized and synthesized to generate a decoupled target control vector with directional and intensity attributes.

[0069] The specific synthesis process is as follows: First, an orthogonal control coordinate system is established with humidity control as the horizontal axis and temperature control as the vertical axis. Then, the temperature control component and humidity control component are used as coordinate values ​​to establish the endpoint coordinates, and a directed line segment from the origin to the endpoint is constructed, thus completing the vectorization synthesis. Each component in the decoupled target control vector corresponds to the power setpoint of the grain silo temperature control actuator and the dehumidification actuator, respectively. The structure of the decoupled target control vector includes:

[0070] Intensity component: This is the Euclidean norm of the decoupled target control vector, used to quantify the total output energy level of the actuator in the controlled field. The higher this component is, the greater the total load power of the air conditioning and dehumidification system.

[0071] Directional component: The angle between the decoupled target control vector and the horizontal axis, used to define the proportional characteristics of the thermal and humidity loads. A directional component biased towards the vertical axis indicates that the current task is dominated by temperature control, while a biased towards the horizontal axis indicates that the task is dominated by humidity control.

[0072] This invention precisely isolates the dynamic coupling interference between temperature and humidity through matrix inverse analysis and vectorization synthesis, transforming the thermal and humidity deviation into a decoupled control vector with intensity and direction attributes. This enables the dynamic allocation of the actuator's output power on demand, effectively solving the control oscillation caused by cross-loop interference and significantly improving the response accuracy and operational stability of grain warehouse environmental regulation.

[0073] Furthermore, the specific process of constructing the field gradient tensor set and generating the mass-energy migration trend flow includes: acquiring dense spatial lattice temperature and humidity data; using the decoupled target control vector as the point control reference; calculating the spatial deviation value of the dense spatial lattice temperature and humidity data relative to the point control reference; performing gradient analysis on the spatial deviation value along the three-dimensional coordinate axes to construct the field gradient tensor set; extracting the potential energy driving direction of the controlled physical quantity in the controlled field based on the field gradient tensor set, wherein the controlled physical quantity includes the field temperature component and the field humidity component; acquiring physical parameters of the controlled medium, including the medium thermal conductivity, medium porosity, and water permeability coefficient; analyzing the flow direction of the controlled physical quantity according to the physical parameters and the potential energy driving direction to establish the spatial diffusion path; and generating the mass-energy migration trend flow according to the time evolution trend of the controlled physical quantity in the spatial diffusion path.

[0074] Specifically, dense spatial lattice temperature and humidity data are acquired, and the intensity (total power) and direction (heat and moisture distribution ratio) of the decoupled target control vector are projected onto each sensing point as a point-level control reference. This reference is a target state vector composed of the expected equilibrium moisture content and equilibrium enthalpy. Using empirical thermodynamic formulas, the raw temperature and humidity data at each point are converted into real-time moisture content and real-time enthalpy, establishing a dimension consistent with the control reference. The spatial deviation of the real-time temperature and humidity data at each point relative to this reference is calculated in the energy (enthalpy) and mass (moisture content) dimensions. The calculation process includes subtracting the control reference vector from the real-time state vector at each point to obtain the spatial deviation value. This spatial deviation value is a binary deviation vector including both mass and energy components.

[0075] A three-dimensional physical space rectangular coordinate system is established with the geometric center of the grain silo's bottom surface as the origin and the sides and height of the silo as the axes. The spatial deviation values ​​are mapped and aligned with the spatial physical coordinates (x, y, z) of each point. Since the sensor points are discrete and cannot be directly differentiated, the following process is performed to convert the scattered deviation values ​​into a continuous spatial deviation function:

[0076] Obtain the coordinates of each sensor location and its corresponding spatial deviation value. Dynamically select the interpolation algorithm based on the sensor deployment density: when the sensor locations are distributed in a regular grid, a spline interpolation algorithm (e.g., cubic spline interpolation) is used. A third-order polynomial is established for each grid node, and a second-order derivative continuity constraint is applied at the junctions to ensure smooth boundary deviation changes. When the sensor locations are irregularly or sparsely distributed, a kriging interpolation algorithm (e.g., ordinary kriging) is used. By analyzing spatial correlation, a continuous spatial deviation distribution function within the controlled field is generated. The spatial deviation distribution function f(x,y,z) uses an adaptive resolution grid. The grid cell size is taken as 1 / 5 to 1 / 10 of the average sensor spacing to ensure that spurious numerical oscillations are not generated while meeting the accuracy requirements of numerical differential calculations.

[0077] The local partial derivatives of the continuous spatial deviation distribution function are calculated using the central difference operator (a numerical method that approximates the derivative of a point by using the function values ​​of a symmetrical neighborhood around the target point). Specifically, the process is as follows: at the grid node coordinates (x, y, z) to be calculated, adjacent nodes with a step size of are taken along each coordinate axis. The instantaneous rate of change at the point is approximated by calculating the ratio of the difference in deviation values ​​between adjacent nodes to twice the step size. The partial derivatives at this point in the x, y, and z axes are obtained to construct a temperature gradient vector representing the spatial variation trend of energy and a humidity gradient vector representing the spatial variation trend of moisture.

[0078] The obtained temperature gradient vector and humidity gradient vector are subjected to tensor product operation to construct a 3×3 field gradient tensor set. This tensor set uses diagonal elements to represent the diffusion intensity of each physical quantity and off-diagonal elements to represent the spatial coupling characteristics of the heat and moisture components, thereby characterizing the physical distribution of heat and moisture non-uniformity within the grain storage field.

[0079] Based on the field gradient tensor set, the transfer vectors of controlled physical quantities (field temperature component and field humidity component) between high and low potential energy regions are identified. The potential energy driving direction is defined as the direction of the physical driving force for the spontaneous migration of heat from high temperature region to low temperature region and moisture from high humidity region to low humidity region.

[0080] The physical parameters of the controlled medium (grain pile) are retrieved. These physical parameters include the thermal conductivity of the medium, which reflects the heat conduction capacity; the porosity of the medium, which reflects the air flow space; and the moisture permeability coefficient, which reflects the resistance to moisture movement. A three-dimensional medium resistance distribution array is constructed based on the physical parameters to characterize the degree of obstruction to mass-energy migration at different spatial locations of the grain pile.

[0081] Using potential energy as the driving force, the effective migration velocity vector of controlled physical quantities under the action of a three-dimensional medium resistance distribution array is calculated using the heat and mass transfer equation of porous media. The effective migration velocity vector includes: an energy dimension (the directional intensity of the temperature gradient corrected based on thermal conductivity) and a mass dimension (the flow angle of the moisture gradient corrected based on permeability and porosity). Taking each densely packed spatial point as the starting point and the effective migration velocity vector as the tangent direction, spatial integration is performed in the three-dimensional physical space coordinate system.

[0082] The specific process of the spatial integration operation is as follows: using dense spatial points as the starting coordinates, along the direction of the instantaneous migration velocity vector at the current position, the position increment is calculated according to a preset spatial step size to obtain the spatial coordinate point at the next moment; by iteratively cyclically iterating this stepping process, a series of discrete spatial coordinate points are sequentially connected to generate a continuous curve representing the mass-energy migration trajectory, which is the spatial diffusion path. This path has a streamlined structure, and its actual direction is not only driven by the difference in high and low potential energy, but also limited by the physical structural characteristics inside the grain pile (such as path curvature caused by local density).

[0083] The system tracks the evolution of controlled physical quantities along the spatial diffusion path in real time, calculating the heat exchange and moisture migration per unit time. This dynamically changing vector information is then aggregated into a mass-energy migration trend flow with spatiotemporal attributes. This trend flow is used to predict the evolution of temperature and humidity distribution within the grain pile over a predetermined time period.

[0084] The specific process of generating the mass-energy migration trend flow includes: extracting the energy gradient component and mass gradient component defined by the field gradient tensor on the spatial diffusion path, and introducing the Fourier thermal conductivity operator and the Fick diffusion operator respectively, performing instantaneous mass-energy flux projection operation, and obtaining the energy migration flux and mass migration flux of each node on the path at the instantaneous moment.

[0085] Specifically, after establishing the spatial diffusion path, the path is discretized into a series of infinitesimal nodes. At each node, the diagonal and off-diagonal elements of the field gradient tensor are extracted.

[0086] Execution law operator fusion method: using Fourier's heat conduction law operator (in For heat flux, Thermal conductivity, The energy migration flux due to heat conduction is calculated (based on the temperature gradient), while the additional heat flux caused by moisture migration is superimposed; the Fick diffusion law operator is used. (in For quality flux, The diffusion coefficient is... (c represents the concentration gradient) The mass migration flux caused by the concentration gradient is calculated and superimposed with a thermal migration correction term driven by the temperature gradient. This fusion method obtains the total energy migration flux and total mass migration flux at each node along the path at any instant, encompassing the effects of thermal and moisture interference. For example, when a significant increase in the mass gradient due to moisture stratification is detected in a local grain pile layer, the weight of the Fick operator is automatically increased, and the velocity and flux of moisture moving to the low-potential region at that point are identified through instantaneous flux projection calculations. This method allows for quantitative analysis of the thermal and moisture exchange within millisecond steps for each micro-element along the path, providing a high-precision data foundation for mass-energy migration trend flow. This scheme corrects the physical bias of single-dimensional predictions through tensor operations involving multi-law fusion, improving the system's ability to predict moisture condensation and local heating trends within the grain pile.

[0087] Introducing time step Based on the minimum size of the spatial grid and the maximum equivalent velocity of mass-energy transfer at the current node, the time step is set to a value that satisfies the condition that the time step is less than or equal to the ratio of the minimum size of the spatial grid to the maximum equivalent velocity. This ensures the convergence and physical realism of the numerical calculation during the time-domain evolution process and prevents energy overflow or computational crashes due to excessively large step sizes. Based on the flux value at the current moment, the changes in physical quantities at each spatial point are predicted for the next moment. The specific logic is: Next moment state = Current state + (Input flux - Output flux) × By fitting the results of simulations over multiple consecutive time steps, the evolution slope and decay characteristics of controlled physical quantities are extracted. For example, it is identified that moisture in a high-humidity area is expected to diffuse along a path to a dry-temperature area after 2 hours. The four-dimensional dataset, containing spatial coordinates, migration direction, flux intensity, and temporal evolution parameters, is encapsulated as a mass-energy migration trend flow. This trend flow is presented as a dynamic vector field, providing a "panoramic view" of the mass-energy flow within the grain pile over a future period.

[0088] This step, by constructing a field gradient tensor and integrating the physical parameters of the grain pile medium, achieves accurate mapping and evolution prediction of the heat and moisture migration path inside the grain warehouse, solving problems such as lagging temperature and humidity control and monitoring blind spots in deep grain piles, and providing a field physical basis for realizing the intelligent transformation from "passive feedback" to "active prediction".

[0089] Furthermore, the specific process of generating the field cooperative control flow includes: acquiring field geometric dimension data containing length, width, and height components within the controlled field; using the field geometric dimension data to establish the physical boundary coordinates of the controlled field and construct a controlled three-dimensional physical space; projecting the decoupled target control vector onto the controlled three-dimensional physical space to construct a field target distribution benchmark; using the mass-energy migration trend flow to analyze the instantaneous diffusion deviation of controlled physical quantities within the controlled three-dimensional physical space and generate a spatial regulation compensation term; and using the spatial regulation compensation term to perform superposition and reconstruction on the field target distribution benchmark to generate the field cooperative control flow.

[0090] Specifically, the geometric dimensions of the grain storage area, including its length, width, and height components, are obtained. Using this data, physical boundary coordinates (e.g., 30m × 20m × 10m) are established in a three-dimensional Cartesian coordinate system, thus constructing a controlled three-dimensional physical space.

[0091] The decoupled target control vector is used as a global driving source and projected onto the controlled three-dimensional physical space through a spatial weight mapping function, thus dividing the controlled three-dimensional physical space into... Each grid cell is a spatial grid unit. Based on the physical installation coordinates of the actuators (such as air conditioning vents and dehumidifier outlets), a spatial weight function is established for each grid cell relative to the actuator. ,in This is the Euclidean distance from the center point of the grid to the actuator. The side length and step size of the spatial grid are uniformly set to 1 / 5 to 1 / 10 of the average spacing between the sensors.

[0092] Subsequently, spatial mapping projection is performed: the magnitude of the decoupled target control vector represents the total control intensity, and its vector direction represents the allocation ratio of thermal and humidity control; a decay diffusion function based on physical distance is introduced (e.g., using an exponential decay model). , It is a natural constant. For spatial distance, The attenuation coefficient, with a value range of 0.1 to 0.5, is used to characterize the physical attenuation rate of control energy as spatial distance increases. The total control intensity is weighted according to the distance between the spatial grid and the actuator outlet. Specifically, grid points closer to the actuator are assigned higher target gain weights, while grid points farther away are assigned lower target gain weights, thereby calculating the ideal temperature target value and ideal moisture content target value corresponding to each grid coordinate point. Finally, the ideal target values ​​of all grid points are encapsulated and established as the field target distribution benchmark. This benchmark is no longer a single setpoint, but a "target cloud map" that matches the actual physical contour of the grain silo. It is used as a static reference field for subsequent spatial compensation and coordinated control.

[0093] Using the mass-energy migration trend flow, the evolution difference of controlled physical quantities within a preset period is extracted at each grid point in the controlled three-dimensional physical space. The state value predicted by the trend flow at the current moment plus the preset period is compared with the target distribution benchmark of the field at the current moment (the state value of the point minus the corresponding target benchmark value). This identifies the instantaneous diffusion deviation that will occur at each grid point due to internal mass-energy migration (such as cold air sinking or local moisture accumulation). The instantaneous diffusion deviation includes polarity (positive or negative) and amplitude (magnitude). Positive deviation (predicted value greater than the benchmark value): indicates that the area will tend to be overheated or overhumidified in the future, requiring a reduction in compensation intensity or an increase in reverse adjustment. Negative deviation (predicted value less than the benchmark value): indicates that the area will tend to be undercooled or underdry in the future, requiring an increase in compensation intensity.

[0094] The specific process of generating the field collaborative control flow includes: constructing a spatial regulation compensation term with reverse hedging polarity using the instantaneous diffusion deviation; and recombining the spatial regulation compensation term into the field target distribution benchmark in a weighted superposition manner to generate the field collaborative control flow.

[0095] First, the instantaneous diffusion deviation of all grid points is spatially topologically encapsulated. Scalar decomposition is then performed on the instantaneous diffusion deviation at each grid point, extracting its deviation modulus in the temperature and humidity dimensions. Subsequently, a mirror inversion operation is performed: a negative feedback hedging operator (with a value of -1) is introduced to invert the deviation modulus, generating a spatial control compensation term with reverse polarity. That is, the spatial control compensation term = instantaneous diffusion deviation × negative feedback gain coefficient × (-1). Here, the feedback gain coefficient represents the control sensitivity of the actuator (air conditioner or fan). A base gain value is preset based on the thermal conductivity and moisture permeability of the grain medium within the controlled field. The current load rate of the actuator is monitored in real time, and the base gain value is dynamically corrected to obtain the final feedback gain coefficient. Each operator in the spatial control compensation term corresponds to a specific coordinate index in three-dimensional space, used to accurately correct the target benchmark at the corresponding location during data reconstruction, achieving early hedging against non-uniform migration trends within the field.

[0096] Subsequently, using the three-dimensional grid index as a reference, the reverse offset components of each point in the spatial control compensation term are extracted, and differentiated weighting coefficients are assigned according to the temperature and humidity sensitivity of different points within the controlled field. Then, the weighted compensation components are successively superimposed onto the corresponding coordinate points of the field target distribution benchmark, achieving dynamic correction of the original ideal target value.

[0097] During the superposition process, boundary constraint calculations are performed simultaneously: the rated hardware parameters of the air conditioning unit and dehumidification system within the controlled field are obtained, the linear adjustment saturation thresholds corresponding to the controlled physical quantities are defined, and the weighted superimposed collaborative control data streams are subtracted from the saturation thresholds one by one for numerical comparison. If the calculated control gain exceeds the hardware execution limit, a saturated nonlinear operator is introduced to perform smoothing and amplitude limiting processing on the value at that point, constraining the output command within the safe operating range.

[0098] By limiting the amplitude of the reconstructed numerical values, it is ensured that the generated commands do not exceed the linear adjustment range of the air conditioning and dehumidification systems. Finally, the reconstructed global grid point data is temporally encapsulated to generate a field-coordinated control flow with feedforward hedging properties. This control flow not only includes the current control objective but also pre-compensation for future mass-energy migration trends, thus achieving precise "zero temperature difference, zero fluctuation" control of the field environment at the physical level. This scheme, by integrating dynamic evolution trends into the static control baseline in advance, gives the control commands a forward-looking capability that surpasses real-time feedback, completely eliminating the spatiotemporal lag phenomenon in large-scale field control.

[0099] This invention solves the problems of uneven distribution of controlled physical quantities and control lag in large fields by using spatial modeling and feedforward compensation. It realizes the dimensional reconstruction from point commands to spatial field commands, and improves the spatial matching accuracy of control commands in controlled three-dimensional physical space.

[0100] Furthermore, the specific process for generating the operating condition sensitivity includes: acquiring operating data of the composite environmental control mechanism, including frequency and power consumption; analyzing the dynamic intervention ratio of the operating data on the controlled physical quantity during changes, and establishing the instantaneous contribution rate of the controlled physical quantity, based on the field collaborative control flow; using the instantaneous contribution rate, establishing the equipment action response intensity of the composite environmental control mechanism to the field collaborative control flow, and generating the operating condition sensitivity; the composite environmental control mechanism includes a temperature control mechanism for performing heat exchange, a ventilation mechanism for performing gas migration within the controlled field, and a dehumidification mechanism for performing moisture content regulation.

[0101] Specifically, operational data of the composite environmental control mechanism within the controlled area of ​​the grain storage facility is acquired. This composite environmental control mechanism includes a temperature control mechanism for heat exchange, a ventilation mechanism for gas migration within the controlled area, and a dehumidification mechanism for moisture content regulation. The temperature control mechanism (e.g., a variable frequency air conditioner) provides the instantaneous frequency and power load, the ventilation mechanism (e.g., an axial flow fan) provides the motor frequency, and the dehumidification mechanism provides the real-time power consumption.

[0102] By combining field-based collaborative control flow and monitoring environmental feedback within a certain step size, the dynamic intervention ratio of each mechanism's actions on the evolution of the physical field is analyzed. Parameter identification based on the least squares method is used to calculate the mapping relationship between the mechanism's frequency increment and the gradient of changes in field physical quantities. Specifically, this includes constructing a linear observation relationship between the mechanism's frequency increment and the gradient of changes in physical quantities, i.e., gradient of changes in physical quantities = mapping relationship coefficient × mechanism frequency increment + constant intercept; the constant intercept reflects the spontaneous trend of changes in field physical quantities when there are no mechanism actions (i.e., frequency increment is 0). The least squares method is used to minimize the sum of squared errors between measured and predicted values, thereby calculating the mapping relationship coefficient reflecting the intensity of the mechanism's frequency change and the environmental response. For example, after increasing the frequency of the temperature control mechanism by 10Hz, the natural convection loss caused by ventilation is eliminated, and the temperature drop slope caused purely by the temperature control action is calculated, thus establishing the instantaneous contribution rate of the temperature control mechanism under the current operating condition. Using this instantaneous contribution rate, the response intensity of the hardware following instructions is further quantified. This intensity reflects the hardware's "execution efficiency". If the instantaneous contribution rate is lower than the preset standard, the mechanism is determined to be in an inefficient operating condition (possibly due to filter clogging or frost).

[0103] As a preferred implementation, the process of generating operating condition sensitivity includes: comparing the instantaneous contribution rate with the preset hardware rated adjustment characteristics, and multidimensionally fusing the response intensity and response delay to generate operating condition sensitivity.

[0104] Specifically, the hardware rated regulation characteristics refer to the nominal functional relationship between the output power and physical influence quantities of the equipment under standard laboratory conditions (such as standard temperature and humidity, no filter blockage, and rated voltage). For example, the theoretical rate of temperature drop per unit volume that should occur for every 1Hz increase in air conditioner frequency. ).

[0105] Dividing the measured instantaneous contribution rate (i.e., actual gain) by the hardware's rated adjustment characteristics yields the performance attenuation coefficient (i.e., the response intensity component). For example, a performance attenuation coefficient of 0.8 indicates that the equipment is currently only operating at 80% of its nominal performance, potentially indicating issues such as insufficient refrigerant, dust accumulation in the heat exchanger, or excessive wind resistance in the grain pile. Simultaneously, the time span from the emission of the coordinated control flow to the extreme value change of the physical quantity at the grid point is measured in real time and established as the response delay component.

[0106] Finally, perform multi-dimensional data fusion: Operating condition sensitivity = intensity weight × response intensity component + latency weight × -（时延敏感因子×响应时延分量） The time delay sensitivity factor is an adjustable parameter used to control the sensitivity's "tolerance" to time lag. In grain storage warehouses with thicker grain piles, this factor is typically set relatively low to accommodate the inherent high thermal inertia of porous media. The response intensity component is nonlinearly fitted to the normalized response time delay component to quantify and generate the operating condition sensitivity. This sensitivity serves as a dynamic calibration operator, describing the actual execution performance of the hardware under the current physical environment, ensuring accurate allocation of control loads based on the "physical condition" of each device. This sensitivity also serves as a correction parameter for subsequent scheduling strategies, indicating which type of mechanism should be prioritized to achieve the objective at the current moment. This scheme provides a "benchmark" for real-time verification of execution performance by establishing an instantaneous differential correlation between hardware actions and physical effects, improving the control robustness of large grain storage warehouses under conditions of aging control equipment or drastic changes in operating conditions.

[0107] This process establishes the real-time control efficiency boundary of the hardware by identifying the dynamic gain between the hardware's execution actions and the physical changes in the field in real time. It solves the problem of instruction deviation caused by hardware performance degradation or changes in environmental conditions in traditional control, and provides a data closed loop for realizing fine-grained scheduling of energy consumption on demand.

[0108] Furthermore, the specific process of determining the composite load constraint term and outputting the collaborative execution response flow includes: taking the time for the controlled physical quantity to return to steady state and the overall operating power consumption as the balance target, and defining the control execution capability boundary based on the sensitivity of operating conditions to establish the composite load constraint term; using the composite load constraint term to allocate the execution weight of the field collaborative control flow, and outputting the collaborative execution response flow.

[0109] Specifically, the system acquires a preset time (e.g., 30 minutes) for the controlled physical quantity to return to a preset steady-state range, and obtains real-time power consumption data for each mechanism. A balance is established between the time duration and overall operating power consumption. The system also establishes the boundary of control execution capability: combining the generated operating condition sensitivity, it analyzes the effective adjustment bandwidth of the temperature control, ventilation, and dehumidification mechanisms under the current operating conditions.

[0110] The boundary of regulation execution capability is not the rated power of the hardware, but rather "rated power operation × operating condition sensitivity". If the sensitivity of an air conditioner is only 0.5, its effective regulation capability in the boundary calculation is halved, thus establishing a composite load constraint term composed of spatiotemporal constraints, energy consumption limits, and equipment effectiveness.

[0111] Subsequently, control weight allocation is performed, and the field collaborative control flow is scheduled using the composite load constraint term. For the temperature and humidity deviation at each grid point within the field, the output ratio of each device is determined through a weight allocation operator.

[0112] The specific process of weight allocation is as follows: For each grid point within the field, based on the polarity of the temperature and humidity deviation (i.e., the ratio of temperature deviation to humidity deviation), the energy efficiency gain ratio (operating condition sensitivity × theoretical energy efficiency ratio) of each regulating mechanism is obtained in real time. This energy efficiency gain ratio reflects the actual regulating efficiency of the device for the target physical quantity per unit power consumption.

[0113] The specific method for allocating the output ratio of each device is as follows: First, define the set of candidate mechanisms (such as air conditioners, fans, and dehumidifiers) to participate in the regulation, and extract the energy efficiency gain ratio corresponding to each mechanism. Normalize the energy efficiency gain ratio of each candidate mechanism, calculate the sum of the energy efficiency gain ratios of all available mechanisms in the controlled field, and divide the energy efficiency gain ratio of a single mechanism by the sum of the energy efficiency gain ratios. The quotient obtained is the initial weight of that mechanism.

[0114] After obtaining the initial weights, a composite load constraint term is introduced for secondary calibration. If the calculated output (initial weight × total deviation load) of a certain mechanism exceeds its control execution capability boundary (rated power × operating condition sensitivity), its weight is forcibly corrected to the upper limit of the boundary, and the excess load residual is redistributed to the remaining unsaturated mechanisms according to the aforementioned proportion. For example, in a grid area near a warehouse window, if humidity deviation is detected as the main factor and the sensitivity of the ventilation mechanism is higher than that of the dehumidification mechanism, then according to the energy efficiency feedback of the composite load constraint term, the ventilation mechanism is assigned an execution weight of 0.7, and the dehumidification mechanism is assigned an execution weight of 0.3.

[0115] Finally, the allocated weights are converted into physical control parameters. The specific conversion process is as follows: obtain the rated adjustment boundaries of each mechanism (upper and lower frequency limits, or opening range 0%-100%), and map the allocated abstract weights into real-time execution parameters: instantaneous operating frequency (or power / opening) = lower frequency limit + abstract weight × (upper frequency limit - lower frequency limit); if abstract weight = 0, output a shutdown command; if abstract weight > 0, calculate the instantaneous operating frequency (or power / opening) according to the formula, and combine it with hardware response delay alignment to generate a coordinated execution response flow containing the start / stop status, operating frequency, and valve opening of each mechanism. In this way, it is ensured that the field environment is brought back to steady state as quickly as possible without exceeding the power load.

[0116] This invention solves the energy efficiency imbalance problem caused by blind actions of the actuator by establishing a balance mechanism between control time and overall power consumption, combined with precise scheduling based on the hardware execution capability boundary. It achieves a high degree of adaptation between control commands and hardware performance, ensuring the stability of control in controlled grain storage areas under complex operating conditions.

[0117] This invention utilizes thermal and humidity decoupling analysis to map sampled data to the controlled state mapping space and construct a set of mutually disturbing coupling matrices, achieving precise separation of cross-interference between controlled physical quantities. By combining a distributed field gradient tensor set with mass-energy migration trend flow execution space reconstruction, it effectively solves the problems of uneven distribution and control lag of controlled physical quantities in three-dimensional physical space. Furthermore, by coordinating the parallel scheduling of composite load constraints and actuator actions, a synergistic balance is achieved between control accuracy, response rate, and overall operating power consumption.

[0118] Example 2

[0119] This embodiment provides a specific application scenario for an intelligent grain warehouse temperature and humidity control system. The scenario is set as a flat-roofed grain warehouse with a large span physical space, where the controlled storage medium is a pile of wheat grain, and the controlled medium has specific porous physical characteristics. (Refer to...) Figure 2 This is a flowchart of the field collaborative control reconstruction process according to an embodiment of the present invention.

[0120] Reference Figure 3 This invention presents a schematic diagram of an intelligent grain warehouse temperature and humidity control system. The system's acquisition unit acquires real-time temperature and humidity scalars transmitted from temperature and humidity sensors distributed within the controlled area.

[0121] The thermal-humidity decoupling analysis module receives the temperature and humidity scalars and projects them onto an enthalpy-humidity thermodynamic mapping space defined with the moisture content component as the horizontal axis and the enthalpy component as the vertical axis. By calculating the physical trajectory of the current controlled physical quantity deviating from the controlled target point, the transient thermal-humidity balance term is obtained. The module further extracts the influence coefficient of the temperature regulation loop on moisture migration as a cross-interference component and the contribution coefficient of the humidity regulation loop to energy change as a correlation feedback component. The module uses these components to construct a mutual interference coupling matrix and uses the analytical operator generated by this matrix to perform inverse analysis on the transient thermal-humidity balance term, eliminating nonlinear mutual interference and outputting the decoupling target control vector.

[0122] The field gradient characterization module performs field gradient tensor modeling. Specifically, it maps the decoupled target control vector to the three-dimensional coordinate system of the grain pile as a point control reference, and calculates the spatial deviation of the temperature and humidity data provided by the dense array of sensors relative to this reference. Subsequently, it performs gradient analysis on the spatial deviation value through numerical differential operators to construct a set of field gradient tensors. Combining the preset thermal conductivity, porosity, and moisture permeability of the medium, it analyzes the flow direction of the controlled physical quantity in the voids of the controlled medium and establishes the mass-energy migration trend flow. The module further uses the mass-energy migration trend flow to analyze the instantaneous diffusion deviation in the controlled three-dimensional physical space, generates a spatial control compensation term, and superimposes it onto the field target distribution reference generated by the decoupled target control vector, finally generating a field cooperative control flow containing the adjustment weights of each grid coordinate point.

[0123] The collaborative execution decision module acquires real-time data from the composite environmental control mechanisms, including the operating frequency of the temperature control mechanism, the motor frequency of the ventilation mechanism, and the power consumption of the dehumidification mechanism. The module uses this real-time data to analyze the dynamic intervention ratio of each mechanism's actions on the evolution of the physical field, and performs multi-dimensional fusion using a weighted exponential function to generate operational condition sensitivity. Subsequently, the system uses the time for the controlled physical quantity to return to steady state and the overall operating power consumption as a balance objective, and delineates the control execution capability boundary based on the operational condition sensitivity, establishing a composite load constraint term. The module uses this constraint term to execute control strategy scheduling for the field collaborative control flow, assigning action weights to each control mechanism, and outputting a collaborative execution response flow.

[0124] Ultimately, the composite environmental control mechanism executes actions according to the response flow command: it uses the temperature control mechanism to adjust the field enthalpy, uses the ventilation mechanism to perform gas migration to smooth the spatial distribution gradient, and cooperates with the dehumidification mechanism to adjust the humidity, thereby achieving precise and energy-saving control of the controlled physical quantities.

[0125] The thermal-humidity decoupling analysis module decouples the temperature and humidity components, eliminating physical interference between controlled physical quantities and ensuring the independence of the control loop. The field gradient characterization module uses mass-energy migration trend flow to dynamically smooth the field distribution gradient, avoiding control lag and uneven field distribution over large spans. The collaborative execution decision module performs weight allocation of the regulating mechanism through composite load constraints, achieving a balance between steady-state regression time and overall operating power consumption. The system ensures control accuracy and energy efficiency stability in the controlled field under complex physical conditions.

[0126] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent method for controlling temperature and humidity in grain warehouses, characterized in that, include: Temperature and humidity data within the controlled field are collected and mapped to a controlled state mapping space with humidity content as the horizontal axis and enthalpy as the vertical axis, generating a transient heat and humidity balance term. Based on the transient heat and humidity balance term, the humidity fluctuations caused by the temperature regulation branch during regulation are analyzed. A sliding window time period is established as the sampling window, and the infinitesimal increments of the power of the temperature regulation branch (ΔPT), the power of the humidity regulation branch (ΔPH), and the instantaneous changes in humidity content in the controlled state mapping space are simultaneously acquired within this window. A partial gain identification operator is introduced to isolate the independent contribution of a single branch under the condition of simultaneous closed-loop operation of two loops. The partial gain identification operator is defined as a time-domain correlation-based identification... The dynamic stripping operator for partial gain identification is calculated as follows: Partial gain identification operator = (instantaneous change - humidity branch self-gain component coefficient × infinitesimal increment ΔPH) / temperature infinitesimal increment ΔPT. A dynamic calibration operator is introduced and product-fitted with the initial self-gain component coefficient to obtain the instantaneous self-gain coefficient. This eliminates the stripping error caused by equipment performance degradation and / or changes in environmental resistance during the partial gain identification operation. The cross-interference component of the temperature regulation branch on the humidity variable, and the associated feedback component of the humidity regulation branch on the temperature variable are extracted to generate a mutual interference coupling matrix group. The transient thermal-humidity balance term and the mutual interference coupling matrix group are then subjected to variable decoupling operations to generate the decoupled target control vector. Acquire dense spatial lattice temperature and humidity data, calculate parameter deviations between points based on the decoupled target control vector, and construct a temperature gradient vector representing the spatial variation trend of energy and a humidity gradient vector representing the spatial variation trend of moisture. Perform tensor product operation on the obtained temperature gradient vector and humidity gradient vector to construct a field gradient tensor set. Based on the field gradient tensor set and the physical parameters of the controlled medium, determine the spatial diffusion path. The physical parameters include the thermal conductivity, porosity, and moisture permeability of the medium. On the spatial diffusion path, extract the energy gradient component and mass gradient component defined by the field gradient tensor, and introduce the Fourier thermal conductivity operator and the Fick diffusion law operator respectively to perform instantaneous mass-energy flux projection operation to obtain the energy migration flux and mass migration flux of each node on the path at the instantaneous moment, generating a mass-energy migration trend flow. Use the mass-energy migration trend flow to perform spatial reconstruction on the decoupled target control vector to generate a field cooperative control flow. The system acquires the operating data of the composite environmental control mechanism and calculates the instantaneous contribution rate of the controlled physical quantity based on the field collaborative control flow to generate the operating condition sensitivity. It then determines the composite load constraint term based on the operating condition sensitivity and uses the composite load constraint term to perform weight allocation on the field collaborative control flow. If the calculated output of a certain mechanism exceeds its control execution capability boundary, its weight is forcibly corrected to the upper limit of the boundary, and the excess load residual is redistributed to the remaining unsaturated mechanisms, outputting the collaborative execution response flow.

2. The intelligent grain warehouse temperature and humidity control method according to claim 1, characterized in that, The specific process for generating the transient thermal and humidity balance term includes: acquiring temperature and humidity data at points within the controlled field; mapping the temperature and humidity data to a controlled state mapping space with moisture content as the horizontal axis and enthalpy as the vertical axis; locating the spatial transient state point generated by mapping the temperature and humidity data in the controlled state mapping space; forming an offset vector based on the spatial transient state point and the corresponding equilibrium steady-state point, and calculating the instantaneous thermodynamic gradient to generate the transient thermal and humidity balance term.

3. The intelligent grain warehouse temperature and humidity control method according to claim 1, characterized in that, The specific process of generating the mutual interference coupling matrix group includes: extracting the influence weight of the temperature regulation branch on the humidity variable based on the transient thermal humidity balance term, and establishing the cross-interference component; extracting the feedback gain of the humidity regulation branch on the temperature variable, and establishing the associated feedback component; filling the cross-interference component and the associated feedback component as matrix elements into an array structure of a preset dimension, performing matrix structure assembly, and generating the mutual interference coupling matrix group.

4. The intelligent grain warehouse temperature and humidity control method according to claim 1, characterized in that, The specific process of generating the decoupled target control vector includes: using the mutual interference coupling matrix group as an analytical operator to perform matrix inverse analytical operation on the transient thermal and humidity balance term; extracting mutually independent temperature control components and humidity control components from the data stream after performing the matrix inverse analytical operation; and performing vectorization synthesis on the temperature control components and the humidity control components to generate the decoupled target control vector.

5. The intelligent grain warehouse temperature and humidity control method according to claim 1, characterized in that, The specific process of constructing the field gradient tensor set and generating the mass-energy migration trend flow includes: acquiring dense spatial lattice temperature and humidity data, using the decoupled target control vector as the point control reference, and calculating the spatial deviation value of the dense spatial lattice temperature and humidity data relative to the point control reference; performing gradient analysis on the spatial deviation value along the three-dimensional coordinate axes to construct the field gradient tensor set; extracting the potential energy driving direction of the controlled physical quantity in the controlled field based on the field gradient tensor set, wherein the controlled physical quantity includes the field temperature component and the field humidity component; analyzing the flow direction of the controlled physical quantity according to the physical parameters and the potential energy driving direction to establish the spatial diffusion path; and generating the mass-energy migration trend flow according to the time evolution trend of the controlled physical quantity in the spatial diffusion path.

6. The intelligent grain warehouse temperature and humidity control method according to claim 1, characterized in that, The specific process of generating the field cooperative control flow includes: acquiring field geometric dimension data containing length, width, and height components within the controlled field; using the field geometric dimension data to establish the physical boundary coordinates of the controlled field and construct a controlled three-dimensional physical space; projecting the decoupled target control vector onto the controlled three-dimensional physical space to construct a field target distribution benchmark; using the mass-energy migration trend flow to analyze the instantaneous diffusion deviation of controlled physical quantities within the controlled three-dimensional physical space and generate a spatial regulation compensation term; and using the spatial regulation compensation term to perform superposition and reconstruction on the field target distribution benchmark to generate the field cooperative control flow.

7. The intelligent grain warehouse temperature and humidity control method according to claim 1, characterized in that, The specific process for generating operational condition sensitivity includes: acquiring operational data of the composite environmental control mechanism, including frequency and power consumption; analyzing the dynamic intervention ratio of the operational data on the controlled physical quantity during changes, and establishing the instantaneous contribution rate of the controlled physical quantity, based on the field collaborative control flow; using the instantaneous contribution rate, establishing the equipment action response intensity of the composite environmental control mechanism to the field collaborative control flow, and generating operational condition sensitivity; the composite environmental control mechanism includes a temperature control mechanism for performing heat exchange, a ventilation mechanism for performing gas migration within the controlled field, and a dehumidification mechanism for performing moisture content regulation.

8. The intelligent grain warehouse temperature and humidity control method according to claim 1, characterized in that, The specific process of determining the composite load constraint term and outputting the collaborative execution response flow includes: taking the time for the controlled physical quantity to return to steady state and the overall operating power consumption as the balance target, and defining the control execution capability boundary based on the sensitivity of the operating conditions to establish the composite load constraint term; using the composite load constraint term to allocate the execution weight of the field collaborative control flow, and outputting the collaborative execution response flow.

9. An intelligent grain warehouse temperature and humidity control system, characterized in that, include: The thermal and humidity decoupling analysis module collects temperature and humidity data within the controlled field and maps it to a controlled state mapping space with humidity content as the horizontal axis and enthalpy as the vertical axis, generating a transient thermal and humidity balance term. Based on the transient thermal and humidity balance term, it analyzes the humidity environment fluctuations caused by the temperature regulation branch when performing regulation actions. A sliding window time period is established as the sampling window, and the infinitesimal increments of the power of the temperature regulation branch (ΔPT), the power of the humidity regulation branch (ΔPH), and the instantaneous changes in humidity content in the controlled state mapping space are simultaneously acquired within this window. A partial gain identification operator is introduced to isolate the independent contribution of a single branch under the condition of simultaneous closed-loop operation of two loops. The partial gain identification operator is defined as a time-domain... The dynamic stripping operator for correlation identification is calculated as follows: Partial gain identification operator = (instantaneous change - humidity branch self-gain component coefficient × infinitesimal increment ΔPH) / temperature infinitesimal increment ΔPT. A dynamic calibration operator is introduced and product-fitted with the initial self-gain component coefficient to obtain the instantaneous self-gain coefficient. This eliminates the stripping error caused by equipment performance degradation and / or changes in environmental resistance during the partial gain identification operation. The cross-interference component of the temperature regulation branch on the humidity variable, and the correlation feedback component of the humidity regulation branch on the temperature variable are extracted to generate a mutual interference coupling matrix set. The transient thermal-humidity balance term and the mutual interference coupling matrix set are then subjected to variable decoupling operations to generate the decoupled target control vector. The field gradient characterization module calculates the parameter deviations between points based on the decoupled target control vector, constructing a temperature gradient vector representing the spatial variation trend of energy and a humidity gradient vector representing the spatial variation trend of moisture. The obtained temperature and humidity gradient vectors are then subjected to tensor product operations to construct a field gradient tensor set. Based on the field gradient tensor set and the physical parameters of the controlled medium, the spatial diffusion path is determined. These physical parameters include the medium's thermal conductivity, porosity, and water permeability. Along the spatial diffusion path, the energy gradient and mass gradient components defined by the field gradient tensor are extracted. Fourier's thermal conductivity operator and Fick's diffusion operator are introduced respectively, and instantaneous mass-energy flux projection operations are performed to obtain the energy migration flux and mass migration flux at each node on the path at any instant, generating a mass-energy migration trend flow. This mass-energy migration trend flow is then used to spatially reconstruct the decoupled target control vector, generating a field cooperative control flow. The collaborative execution decision module utilizes the operational data of the composite environmental control mechanism and calculates the instantaneous contribution rate of the controlled physical quantity based on the field collaborative control flow to generate operational condition sensitivity. It then determines composite load constraint terms based on the operational condition sensitivity and uses these composite load constraint terms to perform weight allocation on the field collaborative control flow. If the calculated output of a certain mechanism exceeds its control execution capability boundary, its weight is forcibly corrected to the upper limit of the boundary, and the excess load residual is redistributed to the remaining unsaturated mechanisms, outputting the collaborative execution response flow.

Images