Rotary dehumidifier control model correction method, device and equipment and storage medium

Abstract

The invention discloses a rotary dehumidifier model correction method, device and equipment and a storage medium, and is applied to the field of industrial equipment. A mechanism model of a rotary dehumidifier is established, and to-be-identified model parameters in the mechanism model are determined; acquiring equipment data acquired at each data acquisition point of the rotary dehumidifier, and screening steady-state equipment data from the equipment data based on a hierarchical clustering method; constructing an objective function which takes the minimum deviation between the model output value of the mechanism model and the actual measurement value as an optimization objective; and parameter values of the to-be-identified model parameters are updated based on the steady-state equipment data until the target function is smaller than a preset threshold value, and operation of the rotary dehumidifier is controlled based on the optimized mechanism model. According to the method, the mechanism model of the rotary dehumidifier is constructed, the steady-state equipment data is screened from the equipment data through the hierarchical clustering method, parameter optimization of the mechanism model is carried out, the obtained mechanism model is more accurate, energy consumption optimization control of the rotary dehumidifier is carried out through the optimized mechanism model, and the energy consumption optimization effect is improved.

Description

Technical Field

[0001] The present invention relates to the field of industrial equipment, and particularly to a method, device, equipment and computer-readable storage medium for calibrating a control model of a rotary dehumidifier. Background Art

[0002] A rotary dehumidifier is an important device in industrial production. Rotary dehumidification is divided into a dehumidification zone and a regeneration zone. The dehumidification zone is used to adsorb water in the air to make the air meet the usage requirements. In the regeneration zone, the heated dry regeneration air contacts the adsorption material to take away the water in the adsorption material and restore its moisture absorption capacity. To ensure the dehumidification and regeneration effects, a surface cooler is generally installed in front of the dehumidification zone and a reheater is installed in front of the regeneration zone. However, with the change of outdoor air temperature and humidity, and the complication of the rotary dehumidification process (double rotors, return air), it is difficult for users to evaluate whether the system is operating in the interval with the lowest energy consumption. To ensure the end-use effect, the supply cooling capacity of the surface cooler and the supply heat of the reheater are generally over-supplied.

[0003] Many manufacturers have tried to use models to simulate the performance of rotary dehumidifiers, but most of the current ones are black-box data models, which have low extensional accuracy and low interpretability, and are of limited help for the research on the control and energy consumption optimization of rotary dehumidifiers. Summary of the Invention

[0004] The purpose of the present invention is to provide a method, device, equipment and storage medium for calibrating a control model of a rotary dehumidifier, which is applied to the field of industrial equipment. The method constructs a mechanism model of the rotary dehumidifier, and screens steady-state equipment data from the equipment data through hierarchical clustering to optimize the parameters of the mechanism model, so that the obtained mechanism model is more accurate, and the energy consumption optimization control of the rotary dehumidifier is carried out through the optimized mechanism model to improve the energy consumption optimization effect.

[0005] To solve the above technical problems, the present invention provides a method for calibrating a control model of a rotary dehumidifier, including:

[0006] Establish a mechanism model of the rotary dehumidifier and determine the model parameters to be identified in the mechanism model;

[0007] Obtain the equipment data collected at each data acquisition point of the rotary dehumidifier, and screen the steady-state equipment data from the equipment data based on hierarchical clustering;

[0008] Construct an objective function with the minimum deviation between the model output value and the actual measurement value of the mechanism model as the optimization target;

[0009] Update the parameter values of the model to be identified based on the steady-state device data until the objective function is less than a preset threshold, and obtain the mechanism model with corrected parameters. Control the operation of the rotary dehumidifier based on the mechanism model with corrected parameters.

[0010] Optionally, the screening of the steady-state device data from the device data based on the hierarchical clustering method includes:

[0011] Determine the minimum steady-state time period based on the sampling time interval and the preset minimum number of steady-state time points;

[0012] Determine the target device data from the device data, combine the target device data collected at the same time point into a target data group, and determine the Euclidean distance between the target data groups;

[0013] Perform hierarchical clustering on all the target data groups based on the Euclidean distance between the target data groups to obtain a hierarchical clustering tree;

[0014] Determine the target hierarchical clustering threshold, and classify the target data groups based on the target hierarchical clustering threshold and the hierarchical clustering tree to obtain a clustering result;

[0015] Determine the steady-state data group from the target data groups based on the minimum steady-state time period and the clustering result;

[0016] Determine all the device data within the acquisition time period corresponding to the steady-state data group as the steady-state device data.

[0017] Optionally, the determination of the target hierarchical clustering threshold includes:

[0018] Sequentially determine the Euclidean distances between the target data groups during the generation of the hierarchical clustering tree as the hierarchical clustering thresholds in descending order;

[0019] Determine the clustering results under each hierarchical clustering threshold, perform a difference operation on the clustering results to obtain a first difference result;

[0020] Determine the number of 0s in the first difference result, perform a difference operation on the number of 0s in all the first difference results to obtain a second difference result;

[0021] Determine the second difference result with the maximum value as the target second difference result, and determine the two first difference results participating in the calculation of the target second difference result as the target first difference results;

[0022] Determine the clustering result that calculates the target first difference result as the target clustering result;

[0023] Determine the maximum hierarchical clustering threshold among those participating in the calculation of the target clustering result as the target hierarchical clustering threshold.

[0024] Optionally, the determining of the Euclidean distances between the target data groups includes:

[0025] Set the weights of the target device data of each type;

[0026] Determine the Euclidean distances between the target device data of the same type among the target data groups;

[0027] Perform weighted summation on the Euclidean distances between the target device data of the same type based on the weights to obtain the Euclidean distances between the target data groups.

[0028] Optionally, the method further includes:

[0029] Taking the current time point as a reference, obtain the historical device data of a preset time period;

[0030] If the historical device data is the steady-state device data, perform averaging processing on the historical device data to obtain the average historical device data;

[0031] Input the average historical device data into the target function to obtain the output target function value;

[0032] If the target function value is greater than the preset threshold, update the parameter values of the model parameters to be identified based on the average historical device data until the target function is less than the preset threshold.

[0033] Optionally, the mechanism model includes: mass conservation model, mass transfer model, energy conservation model, heat transfer model, adsorption heat model, equilibrium adsorption model.

[0034] Optionally, the device data includes: input quantity and output quantity;

[0035] The input quantity includes: treated air inlet temperature, treated air inlet moisture content, regeneration air inlet temperature, and regeneration air inlet moisture content;

[0036] The output quantity includes: treated air outlet temperature, treated air outlet moisture content, regeneration air outlet temperature, and regeneration air outlet moisture content.

[0037] To solve the above technical problems, the present invention provides a method for correcting a control model of a rotary dehumidifier, including:

[0038] A first module for establishing the mechanism model of the rotary dehumidifier and determining the model parameters to be identified in the mechanism model;

[0039] A second module, configured to obtain device data collected at each data acquisition point of the rotary dehumidifier, and screen out steady-state device data from the device data based on the hierarchical clustering method;

[0040] A third module, configured to construct an objective function with the optimization objective of minimizing the deviation between the model output value and the actual measurement value of the mechanism model;

[0041] A fourth module, configured to update the parameter values of the model parameters to be identified based on the steady-state device data until the objective function is less than a preset threshold, obtain the mechanism model with corrected parameters, and control the operation of the rotary dehumidifier based on the mechanism model with corrected parameters.

[0042] To solve the above technical problems, the present invention provides an electronic device, including:

[0043] A memory, configured to store a computer program;

[0044] A processor, configured to implement the above-mentioned method for correcting the control model of the rotary dehumidifier when executing the computer program.

[0045] To solve the above technical problems, the present invention provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, the above-mentioned method for correcting the control model of the rotary dehumidifier is implemented.

[0046] It can be seen that the method of the present invention determines the model parameters to be identified in the mechanism model by establishing a mechanism model of the rotary dehumidifier; obtains device data collected at each data acquisition point of the rotary dehumidifier, and screens out steady-state device data from the device data based on the hierarchical clustering method; constructs an objective function with the optimization objective of minimizing the deviation between the model output value and the actual measurement value of the mechanism model; updates the parameter values of the model parameters to be identified based on the steady-state device data until the objective function is less than a preset threshold, obtains an optimized mechanism model, and controls the operation of the rotary dehumidifier based on the optimized mechanism model.

[0047] The method of the present invention constructs a mechanism model of the rotary dehumidifier, screens out steady-state device data from the device data through the hierarchical clustering method for parameter optimization of the mechanism model, makes the obtained mechanism model more accurate, and optimizes the energy consumption control of the rotary dehumidifier through the optimized mechanism model, thereby improving the energy consumption optimization effect. Description of the Drawings

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the provided drawings.

[0049] Figure 1 It is a flowchart of a method for correcting a control model of a rotary dehumidifier provided by an embodiment of the present invention;

[0050] Figure 2 It is a schematic diagram of a mechanism model architecture provided by an embodiment of the present invention;

[0051] Figure 3 It is a flowchart of a method for identifying model parameters provided by an embodiment of the present invention;

[0052] Figure 4 It is a flowchart of a method for correcting model parameters provided by an embodiment of the present invention;

[0053] Figure 5 It is a flowchart of a device for correcting a control model of a rotary dehumidifier provided by an embodiment of the present invention. Specific embodiments

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0055] A rotary dehumidifier is an important device in industrial production, mainly used in the production workshops of lithium batteries, semiconductors, and pharmaceutical manufacturing. Due to the special nature of the production materials in these workshops, the relative humidity indoors is generally controlled within 1 - 2%. The traditional surface cooling dehumidification method cannot meet this condition. Rotary dehumidification uses the mechanism of adsorption dehumidification to process the air to an ultra-low dew point (-20°C to -60°C), thereby meeting the usage requirements of the end workshops.

[0056] Rotary dehumidification is divided into a dehumidification area and a regeneration area. The dehumidification area is used to adsorb water in the air to make the air meet the usage requirements; in the regeneration area, the heated dry regeneration air contacts the adsorption material to take away the water in the adsorption material and restore its moisture absorption capacity. The adsorption material is driven by a motor to rotate continuously, and the regeneration area and the dehumidification area alternate continuously to achieve the effect of continuous dehumidification.

[0057] The rotary wheel dehumidifier is the main energy-consuming equipment in the air conditioning system. The lower the temperature of the processed air before entering the rotary wheel, the more conducive it is to air dehumidification. The higher the temperature of the regeneration air, the more conducive it is to the regeneration of the adsorption material. To ensure the dehumidification and regeneration effects, a surface cooler is generally installed before the dehumidification area, and a reheater is installed before the regeneration area. However, with the changes in the outdoor air temperature and humidity, and the complexity of the rotary wheel dehumidification process (double rotary wheel, return air), it is difficult for users to evaluate whether the system is operating in the interval with the lowest energy consumption. To ensure the end-use effect, the supply cooling capacity of the surface cooler and the supply heat of the reheater are generally over-supplied. To solve this problem, many manufacturers have tried to use models to simulate the performance of the rotary wheel dehumidifier, but most of them are currently black-box data models, which have low extensional accuracy and low interpretability, and are of limited help for the operation and energy consumption optimization research of the rotary wheel dehumidifier.

[0058] The following combines Figure 1 , Figure 1 is a flowchart of a method for calibrating a control model of a rotary wheel dehumidifier provided by an embodiment of the present invention. The method may include:

[0059] S101: Establish a mechanism model of the rotary wheel dehumidifier and determine the model parameters to be identified in the mechanism model.

[0060] In this embodiment, a mechanism model of the rotary wheel dehumidifier can be established, and the model parameters to be identified in the mechanism model can be determined. Parameter identification is to determine a set of parameter values of the model according to experimental data and the established model, so that the numerical results calculated by the model can best fit the test data.

[0061] This embodiment does not limit the construction method of the mechanism model. Generally, the mechanism model includes: mass conservation model, mass transfer model, energy conservation model, heat transfer model, adsorption heat model, equilibrium adsorption model, etc.

[0062] The expression of the mass conservation model is:

[0063] ;

[0064] The expression of the mass transfer model is:

[0065] ;

[0066] The expression of the energy conservation model is:

[0067] ;

[0068] The expression of the heat transfer model is:

[0069] ;

[0070] The expression of the adsorption heat model is:

[0071] ;

[0072] The expression of the equilibrium adsorption model is:

[0073] ;

[0074] The definitions of the parameters in the above formula are: Y is the water content in air (g / g), W is the water content in the rotor (g / g), v is the air velocity (m / s), t is the time (s), z is the axial direction of the rotor (m), is the equilibrium water content in air (g / g), T is the air temperature (K), T* is the rotor temperature (K), Q is the adsorption heat of the rotor (kJ), is the latent heat of water evaporation (kJ / kg), is the atmospheric pressure (kPa), is the saturation vapor pressure (kPa), is a parameter related to the rotor material, is a parameter related to the mass transfer coefficient of the rotor, is a parameter related to the rotor material, is a parameter related to the mass transfer coefficient of the rotor and the adsorption heat of the rotor, is a parameter related to the heat transfer coefficient of the rotor and the rotor material, is a parameter related to the mass transfer coefficient of the rotor, the rotor material and the adsorption heat of the rotor, is a parameter related to the mass transfer coefficient of the rotor and the rotor material, is the rotor equilibrium adsorption parameter 1, is the rotor equilibrium adsorption parameter 2, is the rotor equilibrium adsorption parameter 3, is the rotor equilibrium adsorption parameter 4, is the rotor equilibrium adsorption parameter 5.

[0075] In this embodiment, the model architecture of the mechanism model can be as Figure 2 shown. The process parameters of the mechanism model can include: the inlet temperature of the treated air, the moisture content of the inlet treated air, the inlet temperature of the regeneration air, the moisture content of the inlet regeneration air and the air flow rate; the rotor design parameters can include: the rotor speed, the rotor treatment area and the rotor regeneration area; the model parameters to be identified can include: the rotor material model parameters and the equilibrium adsorption parameters; the output of the mechanism model can include: the outlet temperature of the treated air, the moisture content of the outlet treated air, the outlet temperature of the regeneration air and the moisture content of the outlet regeneration air.

[0076] In this embodiment, the model parameters to be identified can specifically be: - and - .

[0077] S102: Obtain the device data collected at each data acquisition point of the rotary dehumidifier, and screen the steady-state device data from the device data based on the hierarchical clustering method.

[0078] In this embodiment, the device data collected at each data acquisition point of the rotary dehumidifier can be obtained. Further, the original device data collected can be preprocessed, and the data with missing values and abnormal mutation values can be deleted to ensure the data quality.

[0079] This embodiment does not limit the specific type of the collected device data, which mainly includes the input and output quantities of the model; the input quantities can include: the temperature of the processed air inlet, the moisture content of the processed air inlet, the temperature of the regeneration air inlet, and the moisture content of the regeneration air inlet. The output quantities can include: the temperature of the processed air outlet, the moisture content of the processed air outlet, the temperature of the regeneration air outlet, and the moisture content of the regeneration air outlet.

[0080] Since the external meteorological parameters change continuously, the operation of the rotary dehumidification is not always in the "steady state". When establishing the system model, to ensure the accuracy and rationality of the model, it is necessary to avoid using the "transient state" data as much as possible. This requires identifying and extracting the "steady state" data from the historical production data.

[0081] This embodiment does not limit the specific method of presenting the steady-state device data. Generally, the steady-state device data can be screened from the device data based on the hierarchical clustering method.

[0082] First, this embodiment can set the minimum steady-state time period for screening the steady-state data. This embodiment does not limit the specific method of setting the minimum steady-state time period. It can be set manually based on the actual application scenario, or the minimum steady-state time period can be determined based on the sampling time interval and the preset minimum number of steady-state time points. Specifically, the minimum steady-state time period can be the product of the sampling time interval and the preset minimum number of steady-state time points.

[0083] Further, this embodiment can screen the device data and assign weights. For the collected original data, according to the user's understanding of the device and the process knowledge, the parameters related to the stable operation of the device are screened, and the screened data are assigned weights according to the importance of the variables.

[0084] Specifically, in this embodiment, the target device data can be determined from the device data, the target device data collected at the same time point can be combined into a target data group, and the Euclidean distance between each target data group can be determined. In the process of calculating the Euclidean distance, the weights of each type of target device data are set; the Euclidean distance between the target device data of the same type between the target data groups is determined; the Euclidean distance between the target device data of the same type is weighted and summed based on the weights to obtain the Euclidean distance between each target data group.

[0085] Taking the m*n data set as an example (m is the number of data samples under each data dimension, and n is the data dimension, that is, the number of data types), the calculation formula of the Euclidean distance is shown in the following formula:

[0086] ;

[0087] In the formula, is the Euclidean distance between the u-th target data group and the v-th target data group, is the weight under the i-th data dimension, is the target device data of the i-th data dimension in the u-th target data group, is the target device data of the i-th data dimension in the v-th target data group.

[0088] This embodiment can calculate the Euclidean distance between different target data groups , and obtain a distance matrix of size m*m:

[0089] ;

[0090] This embodiment can perform hierarchical clustering on all target data groups based on the Euclidean distance between each target data group to obtain a hierarchical clustering tree. In the process of generating the hierarchical clustering tree, the two target data groups with the smallest Euclidean distance are classified into one target data group, and the Euclidean distance between the target data groups is recalculated, and this process is repeated until all target data groups are classified into one category, forming a complete hierarchical clustering tree.

[0091] Furthermore, this embodiment can determine the target hierarchical clustering threshold, and classify the target data groups based on the target hierarchical clustering threshold and the hierarchical clustering tree to obtain the clustering result. Different hierarchical clustering thresholds will affect the clustering effect, and thus affect the final steady-state data quality.

[0092] This embodiment does not limit the specific method for determining the target hierarchical clustering threshold. Generally, the Euclidean distances between each target data group during the generation of the hierarchical clustering tree can be sequentially determined as the hierarchical clustering thresholds in descending order; determine the clustering results under each hierarchical clustering threshold, perform a difference operation on each clustering result to obtain a first difference result; determine the number of 0s in the first difference result, and perform a difference operation on the number of 0s in all the first difference results to obtain a second difference result; determine the second difference result with the maximum value as the target second difference result, and determine the two first difference results participating in the calculation of the target second difference result as the target first difference results; determine the clustering result of the calculated target first difference result as the target clustering result; determine the largest hierarchical clustering threshold participating in the calculation of the target clustering result as the target hierarchical clustering threshold.

[0093] Specifically, this embodiment can start from the maximum Euclidean distance to calculate the clustering result. The clustering result is an m*1 array, and the classification results of each target data group are stored in the array, such as 1, 2, 3, etc., representing the categories to which the target data groups belong. This embodiment can perform clustering in descending order of the hierarchical clustering threshold to obtain the clustering result. The order of the clustering results is consistent with the order of clustering based on the hierarchical clustering threshold.

[0094] This embodiment can perform a difference operation on the clustering results obtained by clustering at each Euclidean distance, that is, perform a difference operation on the values stored in the clustering results to obtain the first difference results of each clustering result.

[0095] The first difference result represents the number of target data groups classified into the same type under different hierarchical clustering thresholds. That is, the number of 0s in the first difference result represents the number of occurrences where adjacent target data groups are classified into the same category. The more 0s in the first difference result, the more occurrences where adjacent target data groups are classified into the same category; the fewer 0s in the first difference result, the fewer occurrences where adjacent target data groups are classified into the same category.

[0096] Therefore, this embodiment can count the number of 0s in each first difference result and perform a difference operation on the number of 0s in the first difference result to obtain a second difference result.

[0097] The second difference result is used to represent the change situation of the number of target data groups classified into the same type under different hierarchical clustering thresholds, that is, the change value of the number of 0s in the first difference result. When the second difference result is the largest, it means that the hierarchical clustering threshold is set too small at this time, resulting in steady-state data being classified into different categories. Therefore, the previous (sorted in descending order of Euclidean distance) hierarchical clustering threshold can be selected as the target hierarchical threshold.

[0098] That is, the second difference result with the maximum value is determined as the target second difference result, and the two first difference results participating in the calculation of the target second difference result are determined as the target first difference results; the clustering result for calculating the target first difference result is determined as the target clustering result, which can be the first target clustering result and the second target clustering result respectively; the maximum hierarchical clustering threshold participating in the calculation of the target clustering result is determined as the target hierarchical clustering threshold. For example, the first target clustering result is obtained by clustering under the first hierarchical clustering threshold, and the second target clustering result is obtained by clustering under the second hierarchical clustering threshold, and the first hierarchical clustering threshold is greater than the second hierarchical clustering threshold, then the first hierarchical clustering threshold is determined as the target hierarchical clustering threshold.

[0099] Illustrating with an example, if there are four clustering results A, B, C, and D in the clustering results in the order from the largest to the smallest hierarchical clustering threshold, let the first difference result between A and B be a, and the first difference result between B and C be b. The second difference result c can be obtained between a and b. If c is the maximum value among all the second difference results, the maximum value of the hierarchical clustering thresholds used for clustering to obtain A, B, and C is selected as the target hierarchical clustering threshold. Since in this embodiment, the clustering is performed in the order from the largest to the smallest, that is, the hierarchical clustering threshold corresponding to the clustering result A is determined as the target hierarchical clustering threshold.

[0100] Since in this embodiment, the Euclidean distances between the target data groups in the process of generating the hierarchical clustering tree are sequentially determined as the hierarchical clustering thresholds in the order from the largest to the smallest, the first hierarchical clustering threshold is greater than the second hierarchical clustering threshold, that is, the first hierarchical clustering threshold is the previous hierarchical clustering threshold of the second hierarchical clustering threshold.

[0101] This embodiment can determine the steady-state data group from the target data groups based on the minimum steady-state time period and the clustering result. This embodiment does not limit the specific manner of determining the steady-state data group. Generally, all consecutive and same-category target data groups can be extracted from the target data groups as the target data group set; when the acquisition time period corresponding to the target data group set is greater than the minimum steady-state time period, the target data groups in the target data group set can be determined as the steady-state data groups. This embodiment can determine all the device data within the acquisition time period corresponding to the steady-state data group as the steady-state device data. The steady-state device data acquired at the same time point can be used as a steady-state device data group.

[0102] S103: Construct an objective function with the optimization objective of minimizing the deviation between the model output value and the actual measurement value of the mechanism model.

[0103] In this embodiment, an objective function can be constructed with the optimization goal of minimizing the deviation between the output value of the incentive model and the actual measurement value. The actual measurement value in this embodiment is the output quantity in the collected device data, and the model output value is the output quantity output by the model.

[0104] In this embodiment, the expression of the objective function f can be:

[0105] ;

[0106] In the formula, q is the steady-state operating condition number, is the weight of the treated air outlet temperature in the objective function, is the model output value of the treated air outlet temperature under the i-th steady-state operating condition, is the actual measurement value of the treated air outlet temperature under the i-th steady-state operating condition, is the weight of the regeneration air outlet temperature in the objective function, is the model output value of the regeneration air outlet temperature under the i-th steady-state operating condition, is the actual measurement value of the regeneration air outlet temperature under the i-th steady-state operating condition, is the weight of the moisture content of the treated air outlet in the objective function, is the model output value of the moisture content of the treated air outlet under the i-th steady-state operating condition, is the actual measurement value of the moisture content of the treated air outlet under the i-th steady-state operating condition, is the weight of the moisture content of the regeneration air outlet in the objective function, is the model output value of the moisture content of the regeneration air outlet under the i-th steady-state operating condition, is the actual measurement value of the moisture content of the regeneration air outlet under the i-th steady-state operating condition.

[0107] In this embodiment, a steady-state data segment can be determined. Each steady-state data segment contains continuous groups of steady-state device data. In this embodiment, the steady-state device data in each steady-state data segment can be averaged to obtain a steady-state operating condition.

[0108] In this embodiment, the steady-state operating conditions can be divided into a reference operating condition and a test operating condition. The reference operating condition is used to obtain the parameters of the model to be identified, and the test operating condition is used to test the parameters of the model to be identified.

[0109] This embodiment does not limit the specific method for determining the reference operating condition. Generally, the Euclidean distance between steady-state operating conditions can be obtained based on the target device data and their weights in each steady-state operating condition, and the steady-state operating conditions corresponding to the preset number of maximum Euclidean distances can be determined as the reference operating condition. This embodiment does not limit the size of the preset number, which can be set based on actual applications.

[0110] S104: Update the parameter values of the model to be identified based on the steady-state device data until the objective function is less than the preset threshold, obtain the mechanism model with corrected parameters, and control the operation of the rotary dehumidifier based on the mechanism model with corrected parameters.

[0111] In this embodiment, the parameter values of the model to be identified can be updated based on the steady-state device data until the objective function is less than the preset threshold, obtain the optimized mechanism model, and control the operation of the rotary dehumidifier based on the optimized mechanism model.

[0112] This embodiment does not limit the specific method for updating the parameter values of the model to be identified. Due to the strong non-linear characteristics of the model, using the optimization method based on gradient descent is extremely likely to fall into the local optimal solution. For solving the global optimal solution, in this embodiment, a heuristic algorithm can be selected for identifying the model parameters, such as genetic algorithm, particle swarm optimization algorithm, Nelder-Mead, etc. The Nelder-Mead algorithm is an algorithm for finding the local minimum of a multivariate function, and its advantage is that it does not require the function to be differentiable and can converge to the local minimum relatively quickly.

[0113] Since the identification of model parameters requires a long calculation time, its application scenarios are mainly for scenarios such as initial model establishment, equipment model change, equipment maintenance, etc., and are not suitable for application in the real-time energy consumption optimization model of the rotary dehumidifier (the optimization period is 0.5h - 1h). To ensure the accuracy of the online operation of the mechanism model, this embodiment can use the identified model parameters as the initial values and further correct the parameters.

[0114] Specifically, based on the current time point, obtain the historical device data of the preset time period; this embodiment does not limit the size of the preset time period, which can be set based on the actual application, and generally can be 10 minutes.

[0115] If the historical device data is steady-state device data, the historical device data can be averaged to obtain the average historical device data; input the average historical device data into the objective function to obtain the output objective function value.

[0116] If the objective function value is less than the preset threshold, there is no need to update the model parameters of the mechanism model.

[0117] If the objective function value is greater than the preset threshold, update the parameter values of the model to be identified based on the average historical device data until the objective function is less than the preset threshold.

[0118] Based on the above embodiments, the method of the present invention constructs the mechanism model of the rotary dehumidifier, screens the steady-state device data from the device data through hierarchical clustering method for parameter optimization of the mechanism model, makes the obtained mechanism model more accurate, and optimizes the energy consumption control of the rotary dehumidifier through the optimized mechanism model, thereby improving the energy consumption optimization effect.

[0119] The following, in combination with Figure 3 , Figure 3 is a flow chart for identifying model parameters provided by an embodiment of the present invention. In the parameter identification process, steady-state condition data can be screened from historical data collected during the operation of the rotary dehumidifier, and the steady-state condition data is used as the reference condition data to update the rotary model. An optimization algorithm is executed through the parameter identification model, and optimization parameters of the parameter identification model such as parameter initial values, convergence conditions, and optimization conditions are set. When the algorithm converges, the identified model parameters can be verified using the steady-state condition data. If the accuracy meets the standard, the identified model parameters can be output; when the algorithm does not converge or the accuracy does not meet the standard, the optimization parameters can be modified again for identifying the model parameters.

[0120] The following, in combination with Figure 4 , Figure 4 is a flow chart for correcting model parameters provided by an embodiment of the present invention. In the parameter correction process, an optimization algorithm can be executed through the parameter identification model, and optimization parameters of the parameter identification model such as parameter initial values, convergence conditions, and optimization conditions are set; the model parameters identified in the parameter identification process are updated to the parameter initial values; taking the current time point as the reference, equipment data during the real-time operation of the rotary dehumidifier is obtained; steady-state screening is performed on the collected real-time data to obtain steady-state data; the rotary model is verified through the steady-state data to obtain the error value of the model; when the error value is less than the preset threshold, the parameters of the model can be not updated, and then enter the next correction cycle; when the error value is greater than the preset threshold, the model parameters can be corrected through the parameter correction model; if the algorithm converges, the corrected model parameters can be updated to the parameter initial values, and then enter the next correction cycle; if the algorithm does not converge, the optimization parameters can be modified again and continuously optimized until the algorithm converges.

[0121] The following, in combination with Figure 5 , Figure 5 is a flow chart of a device for correcting a control model of a rotary dehumidifier provided by an embodiment of the present invention. The device may include:

[0122] The first module 100 is used to establish a mechanism model of the rotary dehumidifier and determine the model parameters to be identified in the mechanism model;

[0123] The second module 200 is used to obtain equipment data collected at each data acquisition point of the rotary dehumidifier and screen steady-state equipment data from the equipment data based on the hierarchical clustering method;

[0124] The third module 300 is used to construct an objective function with the minimum deviation between the model output value and the actual measurement value of the mechanism model as the optimization target;

[0125] The fourth module 400 is configured to update the parameter values of the model to be identified based on the steady-state device data until the objective function is less than a preset threshold, so as to obtain the mechanism model with corrected parameters, and control the operation of the rotary dehumidifier based on the mechanism model with corrected parameters.

[0126] Based on the above embodiments, the method of the present invention constructs a mechanism model of a rotary dehumidifier, screens steady-state device data from device data through hierarchical clustering for parameter optimization of the mechanism model, makes the obtained mechanism model more accurate, and performs energy consumption optimization control of the rotary dehumidifier through the optimized mechanism model to improve the energy consumption optimization effect.

[0127] Based on the above embodiments, the second module 200 may include:

[0128] A first unit configured to determine a minimum steady-state time period based on a sampling time interval and a preset minimum number of steady-state time points;

[0129] A second unit configured to determine target device data from the device data, combine the target device data collected at the same time point into a target data group, and determine the Euclidean distance between the target data groups;

[0130] A third unit configured to perform hierarchical clustering on all the target data groups based on the Euclidean distance between the target data groups to obtain a hierarchical clustering tree;

[0131] A fourth unit configured to determine a target hierarchical clustering threshold, and classify the target data groups based on the target hierarchical clustering threshold and the hierarchical clustering tree to obtain a clustering result;

[0132] A fifth unit configured to determine a steady-state data group from the target data groups based on the minimum steady-state time period and the clustering result;

[0133] A sixth unit configured to determine all the device data within the acquisition time period corresponding to the steady-state data group as the steady-state device data.

[0134] Based on the above embodiments, the fourth unit may include:

[0135] A first sub-unit configured to sequentially determine the Euclidean distance between the target data groups during the generation of the hierarchical clustering tree as the hierarchical clustering threshold in descending order;

[0136] A second sub-unit configured to determine the clustering results under the hierarchical clustering thresholds, perform a difference operation on the clustering results to obtain a first difference result;

[0137] A third sub-unit, configured to determine the number of 0s in the first difference result, perform a difference operation on the number of 0s in all the first difference results, and obtain a second difference result;

[0138] A fourth sub-unit, configured to determine the second difference result with the maximum value as the target second difference result, and determine the two first difference results participating in the calculation of the target second difference result as the target first difference results;

[0139] A fifth sub-unit, configured to determine the clustering result that calculates the target first difference result as the target clustering result;

[0140] A sixth sub-unit, configured to determine the maximum hierarchical clustering threshold participating in the calculation of the target clustering result as the target hierarchical clustering threshold.

[0141] Based on the above embodiments, the second unit includes:

[0142] A seventh sub-unit, configured to set the weights of the target device data of each type;

[0143] An eighth sub-unit, configured to determine the Euclidean distance between the target device data of the same type among the target data groups;

[0144] A ninth sub-unit, configured to perform a weighted sum on the Euclidean distance between the target device data of the same type based on the weights, and obtain the Euclidean distance between each target data group.

[0145] Based on the above embodiments, the apparatus may further include:

[0146] A fifth module, configured to obtain historical device data of a preset time period with the current time point as a reference;

[0147] A sixth module, configured to, if the historical device data is the steady-state device data, perform an averaging process on the historical device data to obtain average historical device data;

[0148] A seventh module, configured to input the average historical device data into the target function to obtain an output target function value;

[0149] An eighth module, configured to, if the target function value is greater than the preset threshold, update the parameter value of the model parameter to be identified based on the average historical device data until the target function is less than the preset threshold.

[0150] Based on the above embodiments, the mechanism model includes: a mass conservation model, a mass transfer model, an energy conservation model, a heat transfer model, an adsorption heat model, and an equilibrium adsorption model.

[0151] Based on the above embodiments, the device data includes: input quantity and output quantity;

[0152] The input quantity includes: the temperature of the inlet air to be processed, the moisture content of the inlet air to be processed, the temperature of the inlet air for regeneration, and the moisture content of the inlet air for regeneration;

[0153] The output quantity includes: the temperature of the outlet air to be processed, the moisture content of the outlet air to be processed, the temperature of the outlet air for regeneration, and the moisture content of the outlet air for regeneration.

[0154] Based on the above embodiments, the present invention further provides an electronic device, which may include a memory and a processor. Among them, the memory stores a computer program, and when the processor calls the computer program in the memory, the steps provided by the above embodiments can be implemented. Of course, the device may further include various necessary network interfaces, power supplies, and other components, etc.

[0155] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a terminal or a processor, the method provided by the embodiments of the present invention can be implemented; the storage medium may include: various media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, etc., which can store program codes.

[0156] In the present specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0157] Finally, it should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprises", "comprising" or any other variation thereof is intended to cover a non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the phrase "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0158] The above has introduced in detail a method, device, equipment and storage medium provided by the present invention. Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for calibrating a control model of a rotary dehumidifier, characterized in that, Including: Establish a mechanism model of the rotary dehumidifier and determine the model parameters to be identified in the mechanism model; Obtain the equipment data collected at each data acquisition point of the rotary dehumidifier, and screen the steady-state equipment data from the equipment data based on hierarchical clustering; Construct an objective function with the minimum deviation between the model output value and the actual measurement value of the mechanism model as the optimization objective; Update the parameter values of the model parameters to be identified based on the steady-state equipment data until the objective function is less than a preset threshold, obtain the mechanism model with corrected parameters, and control the operation of the rotary dehumidifier based on the mechanism model with corrected parameters.

2. The method for calibrating the control model of the rotary dehumidifier according to claim 1, wherein, The screening of the steady-state equipment data from the equipment data based on hierarchical clustering includes: Determine the minimum steady-state time period based on the sampling time interval and the preset minimum number of steady-state time points; Determine the target equipment data from the equipment data, combine the target equipment data collected at the same time point into a target data group, and determine the Euclidean distance between each target data group; Perform hierarchical clustering on all the target data groups based on the Euclidean distance between each target data group to obtain a hierarchical clustering tree; Determine the target hierarchical clustering threshold, classify the target data groups based on the target hierarchical clustering threshold and the hierarchical clustering tree to obtain a clustering result; Determine the steady-state data group from the target data groups based on the minimum steady-state time period and the clustering result; Determine all the equipment data within the acquisition time period corresponding to the steady-state data group as the steady-state equipment data.

3. The method for correcting the control model of the rotary dehumidifier according to claim 2, wherein, The determination of the target hierarchical clustering threshold includes: Sequentially determine the Euclidean distances between each target data group during the generation of the hierarchical clustering tree as the hierarchical clustering threshold in descending order; Determine the clustering results under each hierarchical clustering threshold, perform a difference operation on each clustering result to obtain a first difference result; Determine the number of 0s in the first difference result, perform a difference operation on the number of 0s in all the first difference results to obtain a second difference result; Determine the second difference result with the maximum value as the target second difference result, and determine the two first difference results participating in the calculation of the target second difference result as the target first difference results; Determine the clustering result that calculates the target first difference result as the target clustering result; Determine the largest hierarchical clustering threshold participating in the calculation of the target clustering result as the target hierarchical clustering threshold.

4. The method for correcting the control model of the rotary dehumidifier according to claim 2, characterized in that, The determination of the Euclidean distance between each target data group includes: Set the weights of each type of target equipment data; Determine the Euclidean distance between the target equipment data of the same type between the target data groups; Perform a weighted sum on the Euclidean distances between the target equipment data of the same type based on the weights to obtain the Euclidean distance between each target data group.

5. The method for correcting the control model of the rotary dehumidifier according to claim 1, characterized in that It also includes: Taking the current time point as a reference, obtain the historical equipment data of a preset time period; If the historical equipment data is the steady-state equipment data, perform an averaging process on the historical equipment data to obtain the average historical equipment data; Input the average historical device data into the objective function to obtain the output objective function value. If the objective function value is greater than the preset threshold, update the parameter values of the model parameters to be identified based on the average historical device data until the objective function is less than the preset threshold.

6. The method for correcting the control model of the rotary dehumidifier according to claim 1, wherein, The mechanism model includes: a mass conservation model, a mass transfer model, an energy conservation model, a heat transfer model, an adsorption heat model, and an equilibrium adsorption model.

7. The method for correcting the control model of the rotary dehumidifier according to claim 6, wherein, The device data includes: input quantities and output quantities. The input quantities include: the temperature of the processed air at the inlet, the moisture content of the processed air at the inlet, the temperature of the regeneration air at the inlet, and the moisture content of the regeneration air at the inlet. The output quantities include: the temperature of the processed air at the outlet, the moisture content of the processed air at the outlet, the temperature of the regeneration air at the outlet, and the moisture content of the regeneration air at the outlet.

8. A correction device for the control model of a rotary dehumidifier, characterized in that, It includes: A first module for establishing a mechanism model of the rotary dehumidifier and determining the model parameters to be identified in the mechanism model. A second module for obtaining the device data collected at each data collection point of the rotary dehumidifier and screening the steady-state device data from the device data based on the hierarchical clustering method. A third module for constructing an objective function with the minimum deviation between the model output value and the actual measurement value of the mechanism model as the optimization objective. A fourth module for updating the parameter values of the model parameters to be identified based on the steady-state device data until the objective function is less than the preset threshold, obtaining the mechanism model with corrected parameters, and controlling the operation of the rotary dehumidifier based on the mechanism model with corrected parameters.

9. An electronic device, characterized in that, It includes: A memory for storing computer programs. A processor for implementing the method for correcting the control model of the rotary dehumidifier according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, Computer-executable instructions are stored in the computer-readable storage medium, and when the computer-executable instructions are executed by the processor, the method for correcting the control model of the rotary dehumidifier according to any one of claims 1 to 7 is implemented.

Images