Intelligent measuring and controlling system of grain moisture in dryer based on weight change

Abstract

The present application relates to a kind of intelligent measurement and control system of grain moisture based on weight change dryer, specifically to the field of intelligent measurement and control of grain moisture, by real-time weight monitoring and intelligent analysis, the directness and accuracy of grain moisture measurement are significantly improved, based on dynamic prediction and collaborative scheduling, the task allocation and load of multiple dryers are optimized, the overall operation time is shortened, the factory capacity is improved, by real-time monitoring and fine adjustment to global moisture distribution, the final moisture content of each batch of grain is highly consistent, the product quality uniformity is improved, at the same time, accurate process control avoids excessive drying, reduces energy consumption, the whole design can adapt to the change of equipment state and production task, realizes the upgrade from single machine automation to factory level collaborative intelligence.

Description

Technical Field

[0001] This invention relates to the field of intelligent measurement and control of grain moisture, and more specifically, to an intelligent measurement and control system for grain moisture in a dryer based on weight changes. Background Technology

[0002] In the actual production and operation of large-scale grain storage and processing enterprises, distributed drying plants have become the mainstream operating mode. These plants are usually equipped with dozens or even hundreds of drying towers, which simultaneously process grains from different production areas and varieties with significant differences in initial moisture content. The production process needs to meet diversified order requirements, that is, different batches of grain need to be dried to a specified and precise target moisture content to ensure subsequent storage safety and processing quality. Due to differences in model, service life, maintenance status and the characteristics of the grains they process, the drying equipment varies significantly in terms of real-time processing capacity, thermal energy utilization efficiency and operational stability. In addition, the volatility of incoming grains, the urgency of orders and changes in energy supply are intertwined with internal and external factors, forming a highly dynamic and multi-constrained complex production environment. In this scenario, ensuring the maximization of the overall capacity of the plant, the optimization of energy consumption and the high consistency of moisture content of all batches of grain leaving the plant are the key to improving the core competitiveness of enterprises, and also place extremely high demands on the overall collaborative control of the production process.

[0003] Currently, automated control technology for single dryers is relatively mature. A typical solution involves deploying weight sensors at the dryer's base, with signals fed into a programmable logic controller (PLC). The PLC calculates the net grain weight by measuring the equipment's own weight, obtaining the initial moisture content and the total weight after grain feeding. Then, combining this with a preset target moisture content, it uses a formula to calculate the target weight at the end of drying. The PLC controls the start and stop of the drying operation by continuously comparing the real-time weight with the target weight. However, this technical solution only achieves single-machine-level closed-loop control based on the final target, which has significant limitations. Specifically, the independent operation of multiple dryers creates "data silos." Key process information such as real-time weight changes, calculated instantaneous drying rates, and estimated completion times are not effectively collected, aggregated, and comprehensively analyzed. Existing data... While distributed control systems or manufacturing execution systems can schedule production, their scheduling decisions are often based on fixed planned durations or historical experience, lacking precise perception of the real-time processing capacity and dynamic load of each piece of equipment. Because they cannot grasp the true progress and efficiency of each dryer based on weight change data in real time, the upper-level scheduling system struggles to achieve optimal dynamic allocation of tasks, load balancing, and energy coordination. This directly leads to some equipment operating under overload for extended periods while others remain idle, preventing the factory's overall capacity from reaching its optimal level. Furthermore, the inability to finely coordinate and adjust the final moisture content of each machine results in inconsistent final moisture content across different batches of grain, affecting the stability and uniformity of overall quality. Therefore, there is an urgent need for a factory-level intelligent control system that can deeply integrate real-time weight change data, achieve multi-machine collaborative optimization, and precisely control global moisture content. Summary of the Invention

[0004] This invention addresses the technical problems existing in the prior art by providing an intelligent measurement and control system for grain moisture in a dryer based on weight changes, thereby solving the problems mentioned in the background art.

[0005] The technical solution of this invention to solve the above-mentioned technical problems is as follows: specifically including: an edge computing module deployed in each dryer, a central processing module deployed in the factory control center, a collaborative scheduling module, and a control optimization module, wherein; Edge computing module: In response to the start of grain loading in the dryer, it obtains the dryer's own weight, the total initial weight after loading, and the initial moisture content of the grain, calculates the net weight of the loaded grain, and calculates the target total weight when drying is completed based on the preset target moisture content; during the drying process, the real-time total weight is continuously collected by the weight sensor installed in the dryer, and the remaining dewatering equivalent representing the current remaining dewatering workload and the instantaneous dewatering power representing the instantaneous drying efficiency are dynamically calculated, and both are output to the central processing module; The central processing module receives and integrates the remaining dehydration equivalent and instantaneous dehydration power data uploaded by all edge computing modules, establishes and maintains a dynamic digital twin for each dryer, and updates the state of the digital twin driven by the real-time received remaining dehydration equivalent and instantaneous dehydration power. Based on the updated state of the digital twin, a dynamic prediction algorithm is applied to estimate the expected completion time of each dryer, and further, based on the received remaining dehydration equivalent and instantaneous dehydration power, the predicted final moisture content of each batch of grain is predicted. Collaborative scheduling module: Based on the estimated completion time and instantaneous dewatering power of each dryer provided by the central processing module, it performs online multi-objective optimization decision-making, allocates suitable dryers to newly arrived drying tasks, and generates scheduling adjustment instructions for drying tasks that are already being processed; Control optimization module: Based on the predicted final moisture content distribution of all batches of grain in processing provided by the central processing module, it compares it with the preset global moisture control target range and calculates the adjustment amount used to fine-tune the target total weight of each dryer; the adjustment amount is sent to the edge computing module of the corresponding dryer, so that the edge computing module controls the drying process according to the adjusted target total weight. In a preferred embodiment, the specific process of calculating the net weight of the grain and the target total weight upon completion of drying in the edge computing module is as follows: First, subtract the dryer's own weight from the total initial weight after loading the grain to obtain the net weight of the loaded grain. Then, multiply the net weight of the loaded grain by the difference between the initial moisture content of the grain and the net weight of the grain to obtain an intermediate value. Next, divide this intermediate value by the difference between the preset target moisture content and the net weight of the grain to obtain a calculation result. Finally, add this calculation result to the dryer's own weight, and the sum obtained is the target total weight when drying is completed.

[0006] In a preferred embodiment, the process of dynamically calculating the remaining dewatering equivalent representing the current remaining dewatering workload and the instantaneous dewatering power representing the instantaneous drying efficiency specifically includes: During the drying process, the real-time total weight is continuously collected at preset fixed time intervals. The remaining water removal equivalent is obtained by subtracting the target total weight when drying is completed from the latest collected real-time total weight. Instantaneous water removal power is calculated by subtracting the total real-time weight collected at the previous sampling time from the total real-time weight collected at the current sampling time to obtain a weight difference. The weight difference is then divided by a fixed time interval to obtain an original rate of change value. A dynamic smoothing coefficient, determined by a basic smoothing factor and a sensitivity coefficient, is introduced. The original rate of change value at the current sampling time is then weighted and mixed with the smoothed rate of change value calculated at the previous sampling time. The weight of the original rate of change value at the current sampling time is the dynamic smoothing coefficient, and the weight of the smoothed rate of change value at the previous sampling time is one minus the dynamic smoothing coefficient. The smoothed rate of change value at the current sampling time is obtained by mixing and calculating as the instantaneous water removal power. The edge computing module packages the remaining water removal equivalent and instantaneous water removal power, along with the device identifier and time information, into a unified data frame and uploads it.

[0007] In a preferred embodiment, the process of establishing and maintaining a dynamic digital twin for each dryer in the central processing module, and applying a dynamic prediction algorithm based on the updated digital twin status to estimate the expected completion time, specifically includes: The central processing module constructs a corresponding digital twin for each dryer. Each digital twin continuously receives and stores pairing information consisting of the remaining water removal equivalent and instantaneous water removal power uploaded from the corresponding edge computing module. This drives the digital twin's status to synchronize with the dryer in real time and maintains a dynamic data set that stores multiple sets of such pairing information recently. Based on the recent historical pairing information stored in the dynamic dataset, a nonlinear dynamic model is constructed for each dryer to describe the change in its dehydration efficiency as the drying process progresses. A recursive parameter estimation algorithm with a forgetting factor is adopted. By utilizing continuously updated historical pairing information, the goal is to make the model's dewatering power approximate the actual received instantaneous dewatering power. The initial dewatering power amplitude parameter, efficiency decay rate parameter, and steady-state dewatering power basis parameter in the nonlinear dynamic model are identified and corrected online in real time. After obtaining the nonlinear dynamic model at the current moment, and after the parameters have been identified and corrected online, the remaining water removal equivalent at the current moment is used as the starting value for integration, and zero is used as the ending value for integration. The inverse of the water removal power of the model is integrated to obtain the expected completion time considering the nonlinear decay characteristics of future efficiency. This integration operation is the dynamic prediction algorithm used.

[0008] In a preferred embodiment, the nonlinear dynamic model defines the relationship between the model dewatering power and the remaining dewatering equivalent, indicating that the model dewatering power is equal to an initial dewatering power amplitude parameter multiplied by a power of the natural constant base, the exponent of which is the product of the negative efficiency decay rate parameter and the total amount of moisture removed, plus a steady-state dewatering power base parameter, wherein the total amount of moisture removed is obtained by subtracting the remaining dewatering equivalent at the beginning of drying from the remaining dewatering equivalent at the current moment, and the remaining dewatering equivalent at the beginning of drying is calculated by subtracting the target total weight at the end of drying from the total initial weight after loading.

[0009] In a preferred embodiment, the process of predicting the final moisture content of each batch of grain specifically involves: First, based on the target total weight of the dryer, the machine body weight received from the corresponding edge computing module, and the preset target moisture content obtained from the production order, a dry matter mass estimation operation is performed. This dry matter mass estimation operation subtracts the machine body weight from the target total weight to obtain the target net weight, and then multiplies the target net weight by the difference between the target net weight and the preset target moisture content. The calculation result is used as the dry matter mass estimation value of the current batch of grain. Secondly, using the estimated dry matter mass, the real-time total weight received from the corresponding edge computing module, and the machine's own weight, the current instantaneous moisture content estimation operation is performed. This current instantaneous moisture content estimation operation subtracts the machine's own weight from the real-time total weight to obtain the current total grain mass. Then, the difference between the estimated dry matter mass and the quotient of the current total grain mass is used as the estimated instantaneous moisture content of the grain at the current moment. Next, the remaining water removal equivalent at the current moment is obtained from the digital twin of the corresponding dryer. This remaining water removal equivalent is identified as the total amount of water that is expected to be removed from the current moment until the drying is completed. Finally, the final moisture content prediction calculation is performed. This final moisture content prediction calculation first multiplies the current total mass of grain by the current instantaneous moisture content estimate to obtain an intermediate product. Then, the total amount of moisture expected to be removed is subtracted from this intermediate product to obtain the numerator of the prediction formula. At the same time, the total mass of grain is subtracted from the total amount of moisture expected to be removed to obtain the denominator of the prediction formula. Finally, the ratio obtained by dividing the numerator by the denominator is used as the predicted final moisture content of this batch of grain.

[0010] In a preferred embodiment, the process of performing online multi-objective optimization decision-making and allocating a suitable dryer to a newly arrived drying task in the collaborative scheduling module specifically includes: Upon receiving a new drying task, the collaborative scheduling module immediately performs data acquisition and organization operations. These operations include: obtaining the latest status data stored in the digital twins of each dryer in the entire plant from the central processing module, including the estimated completion time and instantaneous dehydration power; extracting the initial moisture content of each batch of grain recently processed by the dryer from the historical task records maintained by the central processing module and associated with each dryer, and calculating a mean and a standard deviation based on these initial moisture contents, which together serve as the historical initial moisture content statistical features of the dryer; obtaining the initial moisture content, target moisture content, and net weight information of the grain for the new task from the production order corresponding to the new task; and organizing the estimated completion time, instantaneous dehydration power, historical initial moisture content statistical features of each dryer, and the initial moisture content, target moisture content, and net weight information of the grain for the new task into a plant-wide decision status set. For each dryer marked as available in the new task and decision state set, a comprehensive suitability score is calculated. The calculation process is as follows: First, a base score component is calculated, which is equal to the weighted sum of a time advantage component and an efficiency advantage component. The time advantage component is obtained by calculating the exponent of the negative estimated completion time divided by a preset time normalization constant, using the natural constant as the base. The efficiency advantage component is obtained by dividing the instantaneous dewatering power of the dryer by the average instantaneous dewatering power of all available dryers in the decision state set. The time advantage component is multiplied by a first time weighting coefficient between zero and one, and the efficiency advantage component is multiplied by a second time weighting coefficient between zero and one. The two products are then added together, and the result is the base score component. Then, the calculation... A characteristic matching degree component is calculated as follows: The absolute value of the difference between the initial moisture content of the grain for the new task and the mean value in the historical initial moisture content statistical characteristics of the dryer is calculated; this absolute value is then divided by the product of a preset adjustment coefficient and the standard deviation in the historical initial moisture content statistical characteristics of the dryer, yielding a quotient; finally, using the natural constant as the base and the negative of the quotient as the exponent, the value of the exponent is calculated, which is the characteristic matching degree component; next, a load balancing component is calculated, which is obtained by adding a preset inhibition intensity coefficient to the product of the number one and the total number of tasks currently assigned to the dryer, obtaining a sum, and then taking the reciprocal of the sum; finally, the basic score component, the characteristic matching degree component, and the load balancing component are multiplied consecutively, and the resulting product is the overall suitability score of the dryer for the new task; After calculating the overall suitability score of all available dryers for the new task, the dryer with the highest overall suitability score is selected as the assignment target for the new task.

[0011] In a preferred embodiment, the process of generating scheduling adjustment instructions for a drying task already in progress specifically includes: The collaborative scheduling module periodically or in response to preset events triggers a plant-wide rescheduling calculation. The rescheduling calculation first obtains the decision status set of the entire plant and includes all currently processed tasks as objects to be reassigned. Based on this set of decision states, calculate the corresponding comprehensive suitability score for each combination of task to be assigned and each available dryer; To formally describe the allocation relationship between tasks and dryers, a binary decision variable with a value of zero or one is defined for each pair of tasks and available dryers. When the binary decision variable has a value of one, it means that the task is allocated to the dryer, and when it has a value of zero, it means that the task is not allocated. The optimization objective is to maximize the sum of the overall suitability scores of all successfully assigned task-dryer combinations in the entire plant. The sum of the overall suitability scores is obtained by multiplying the overall suitability scores of all tasks and all available dryer combinations by the corresponding binary decision variables, and then adding all the products together. Construct an optimization model whose optimization objective is to maximize the sum of comprehensive suitability scores, and satisfy the following constraints: for each task to be assigned, there is exactly one binary decision variable corresponding to this task among all binary decision variables; for each available dryer, the total processing burden of all tasks assigned to this dryer does not exceed its preset processing capacity limit. A heuristic search algorithm was used to solve the optimization model online, resulting in an optimal or near-optimal task-dryer redistribution scheme. By comparing this redistribution scheme with the current actual operating status, the in-process tasks that need to be migrated from the current dryer to another dryer are identified. For each task that needs to be migrated, a scheduling adjustment instruction is generated, which contains a series of ordered control instructions.

[0012] In a preferred embodiment, the process of comparing the predicted final moisture content distribution of all batches of grain in processing with a preset global moisture control target range and calculating the adjustment amount for fine-tuning the target total weight of each dryer in the control optimization module is as follows: The control optimization module first periodically obtains the predicted final moisture content of all batches being processed in the entire plant from the central processing module, forming the current moisture prediction dataset. Based on this current moisture prediction dataset, robust distribution parameter estimation is performed. The robust distribution parameter estimation process includes: A1. Calculating the initial arithmetic mean and initial standard deviation of the current moisture prediction dataset; A2. Calculating a robust weight for each predicted final moisture content value in the current moisture prediction dataset; A3. Using all the calculated robust weights, weighting all predicted final moisture content values ​​in the current moisture prediction dataset, recalculating the weighted arithmetic mean as the robust center value, and calculating the weighted standard deviation as the robust dispersion. Next, a global consistency deviation index is calculated, which is based on the preset global moisture control target interval, robust center value, and robust dispersion. Finally, for each dryer in operation, calculate the adjustment amount for the target total weight using the following steps: B1. The product of a negative proportional control coefficient and the difference between the predicted final moisture content of the dryer and the center value of the target interval, minus the product of a differential control coefficient and the rate of change of the predicted final moisture content of the dryer, is used as the intended value for local moisture adjustment. B2. Calculate a gradient value based on the robust central value, robust dispersion, target interval central value, and the dryer's own predicted final moisture content. B3. Subtract the product of a global coordination gain coefficient and the gradient value from the local moisture adjustment intention value to obtain a comprehensive adjustment intention value; then divide this comprehensive adjustment intention value by the sensitivity of the dryer's predicted final moisture content to the change in the target total weight, and the resulting quotient is the initial adjustment amount for the target total weight of the dryer. B4. Compare the initial adjustment amount with a preset maximum allowable positive adjustment amount and a preset maximum allowable negative adjustment amount. If the initial adjustment amount is greater than the maximum allowable positive adjustment amount, the maximum allowable positive adjustment amount is taken as the final adjustment amount. If the initial adjustment amount is less than the maximum allowable negative adjustment amount, the maximum allowable negative adjustment amount is taken as the final adjustment amount. Otherwise, the initial adjustment amount is taken as the final adjustment amount. The final adjustment amount is the adjustment amount for the target total weight to be issued.

[0013] In a preferred embodiment, the process of sending the adjustment amount to the edge computing module of the corresponding dryer specifically involves: The control optimization module constructs a target weight update instruction for each dryer that needs to adjust its target total weight. The instruction includes the dryer's device identifier, the new target total weight value, the instruction sequence number, and the timestamp. The new target total weight value is formed by adding the adjustment amount of the target total weight to the dryer's current target total weight. Before the target weight update command is issued, the control optimization module executes security verification logic; Through the factory control network, the target weight update command that has passed the verification is sent to the edge computing module of the corresponding dryer, i.e. its programmable logic controller, using a secure and reliable industrial communication protocol. Upon receiving the target weight update command, the programmable logic controller immediately replaces the original target total weight stored in its internal memory with the new target total weight value contained in the target weight update command. This new target total weight value is used as the new benchmark for determining the stop of the subsequent drying process, and the programmable logic controller records that this target total weight update comes from global optimization control. At the same time, the programmable logic controller sends a confirmation report to the central processing module. After receiving the confirmation report, the central processing module updates the target total weight parameter stored in the digital twin of the corresponding dryer to the new target total weight value. After updating the target total weight, the programmable logic controller also initiates a temporary high-frequency process status monitoring to observe the dynamic response of the drying process after adjusting the target total weight.

[0014] The beneficial effects of this invention are as follows: Through real-time weight monitoring and intelligent analysis, the directness and accuracy of grain moisture measurement are significantly improved; based on dynamic prediction and collaborative scheduling, the task allocation and load of multiple dryers are optimized, the overall operation time is shortened, and the factory capacity is increased; through real-time monitoring and fine-tuning of global moisture distribution, the final moisture content of each batch of grain is highly consistent, improving the uniformity of product quality; at the same time, precise process control avoids over-drying and reduces energy consumption; the entire design can adapt to changes in equipment status and production tasks, realizing the upgrade from single-machine automation to factory-level collaborative intelligence. Attached Figure Description

[0015] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a block diagram of the system structure of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0017] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0018] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0019] Example 1 This embodiment provides, for example Figure 1-2 The present invention discloses an intelligent measurement and control system for grain moisture in a dryer based on weight variation, which specifically includes: an edge computing module deployed in each dryer, a central processing module deployed in the factory control center, a collaborative scheduling module, and a control optimization module, wherein; Edge computing module: In response to the start of grain loading in the dryer, it obtains the dryer's own weight, the total initial weight after loading, and the initial moisture content of the grain, calculates the net weight of the loaded grain, and calculates the target total weight when drying is completed based on the preset target moisture content; during the drying process, the real-time total weight is continuously collected by the weight sensor installed in the dryer, and the remaining dewatering equivalent representing the current remaining dewatering workload and the instantaneous dewatering power representing the instantaneous drying efficiency are dynamically calculated, and both are output to the central processing module; The central processing module receives and integrates the remaining dehydration equivalent and instantaneous dehydration power data uploaded by all edge computing modules, establishes and maintains a dynamic digital twin for each dryer, and updates the state of the digital twin driven by the real-time received remaining dehydration equivalent and instantaneous dehydration power. Based on the updated state of the digital twin, a dynamic prediction algorithm is applied to estimate the expected completion time of each dryer, and further, based on the received remaining dehydration equivalent and instantaneous dehydration power, the predicted final moisture content of each batch of grain is predicted. Collaborative scheduling module: Based on the estimated completion time and instantaneous dewatering power of each dryer provided by the central processing module, it performs online multi-objective optimization decision-making, allocates suitable dryers to newly arrived drying tasks, and generates scheduling adjustment instructions for drying tasks that are already being processed; Control optimization module: Based on the predicted final moisture content distribution of all batches of grain in processing provided by the central processing module, it compares it with the preset global moisture control target range and calculates the adjustment amount used to fine-tune the target total weight of each dryer; the adjustment amount is sent to the edge computing module of the corresponding dryer, so that the edge computing module controls the drying process according to the adjusted target total weight.

[0020] In this embodiment, it is necessary to specifically explain the process in the edge computing module for calculating the net weight of the grain and, based on the preset target moisture content, calculating the target total weight upon completion of drying: First, subtract the dryer's own weight from the total initial weight of the loaded grain to obtain the net weight of the loaded grain, which is typically between 5,000 and 20,000 kilograms. Then, multiply the net weight of the loaded grain by the difference between the initial moisture content of the grain and the net weight of the loaded grain, obtaining an intermediate value. This intermediate value represents the mass of the dry matter portion of the net weight of the loaded grain; its calculation excludes the initial moisture content and forms the basis for subsequent calculations. Next, divide this intermediate value by the difference between the preset target moisture content and the target moisture content, obtaining a calculation result. This calculation result represents the total mass of the grain's dry matter and remaining moisture when the target moisture content is reached. The preset target moisture content is usually set between 12.5% ​​and 14.5% according to storage safety standards. Finally, add this calculation result to the dryer's own weight; the sum obtained is the target total weight when drying is complete. This calculation process is equivalent to executing the following logic: Target total weight = Dryer's own weight + [Net weight of loaded grain × (1 - Initial moisture content of grain) / (1 - Preset target moisture content)]; Based on this, the dynamic calculation process for the remaining dewatering equivalent, representing the current remaining dewatering workload, and the instantaneous dewatering power, representing the instantaneous drying efficiency, is as follows: During the drying process, the real-time total weight is continuously collected at a preset fixed time interval, preferably one second. The remaining water removal equivalent is obtained by subtracting the target total weight when drying is completed from the latest collected real-time total weight. A negative value indicates over-drying, a positive value indicates incomplete drying, and a zero value triggers a stop signal. Instantaneous water removal power is calculated by subtracting the total real-time weight collected at the previous sampling time from the total real-time weight collected at the current sampling time to obtain a weight difference. Dividing this weight difference by a fixed time interval yields an initial rate of change value. A dynamic smoothing coefficient, determined by a base smoothing factor and a sensitivity coefficient, is introduced. The base smoothing factor, ranging from 0.1 to 0.3, sets the minimum smoothing intensity; the sensitivity coefficient, ranging from 0.5 to 2, controls the response intensity to changing acceleration. The initial rate of change value at the current sampling time is weighted and mixed with the smoothed rate of change value calculated at the previous sampling time. The weight of the initial rate of change value during weighted mixing is the dynamic smoothing coefficient, and the weight of the smoothed rate of change value at the previous sampling time is... The smoothed rate of change at the current sampling moment is calculated by subtracting the dynamic smoothing coefficient from the original rate of change and then performing a mixed calculation. This is used as the instantaneous dewatering power. The value of the dynamic smoothing coefficient is determined by the product of the base smoothing factor, the sensitivity coefficient, and the absolute value of the acceleration of the original rate of change. This makes the dynamic smoothing coefficient close to the base smoothing factor when the drying rate is stable, and the algorithm tends to smooth historical data to suppress noise. When the drying rate changes drastically, the increase in the absolute value of the acceleration leads to an increase in the dynamic smoothing coefficient, and the algorithm adopts the current original rate of change value more quickly to ensure response speed. The absolute value of the acceleration of the original rate of change value is obtained by subtracting the original rate of change value at the previous sampling moment from the original rate of change value at the current sampling moment, taking the absolute value, and then dividing by a fixed time interval. The edge computing module packages the remaining water removal equivalent and instantaneous water removal power along with the device identifier and time information into a unified data frame and uploads it. The data frame contains at least the following fields: start identifier, device unique identifier, data acquisition timestamp, remaining water removal equivalent, instantaneous water removal power, and checksum. This data frame is uploaded to the central processing module via industrial Ethernet, providing real-time and noise-resistant key process status information for global collaborative control.

[0021] In this embodiment, it is specifically necessary to explain that in the central processing module, the process of establishing and maintaining a dynamic digital twin for each dryer, and applying a dynamic prediction algorithm based on the updated digital twin status to estimate the expected completion time is as follows: The central processing module constructs a corresponding digital twin for each dryer. Each digital twin continuously receives and stores pairing information uploaded from the corresponding edge computing module, consisting of the remaining water removal equivalent and instantaneous water removal power. This drives the digital twin's status to synchronize with the dryer in real time and maintains a dynamic data set that stores multiple sets of such pairing information recently. For example, the dynamic data set typically maintains pairing information within the last 20 to 30 minutes, and its capacity can be set to 500 to 1000 sets. It is updated using a first-in-first-out queue mechanism. This set provides a continuous and stable time-series data foundation for subsequent model construction and parameter identification. Based on the recent historical pairing information stored in the dynamic dataset, a nonlinear dynamic model is constructed for each dryer to describe the change of its dehydration efficiency with the drying process. The purpose of constructing this nonlinear dynamic model is to more accurately characterize the characteristic that the dehydration efficiency is not constant during the drying process, but rather decays nonlinearly as the remaining moisture decreases. This is fundamentally different from the prediction method based on fixed average efficiency and can significantly improve the prediction accuracy. A recursive parameter estimation algorithm with a forgetting factor is adopted. Utilizing continuously updated historical pairing information, the algorithm aims to make the model's dewatering power approximate the actual instantaneous dewatering power received. This allows for online real-time identification and correction of the initial dewatering power amplitude parameter, efficiency decay rate parameter, and steady-state dewatering power basis parameter in the nonlinear dynamic model. The preferred recursive parameter estimation algorithm with a forgetting factor is the recursive least squares method with a forgetting factor. The forgetting factor is between 0.95 and 0.99, which is used to give higher weight to recent data when updating parameters, thereby enabling the model to adapt to the slow drift of the dryer's efficiency characteristics caused by changes in operating conditions, grain variety, or environment. After obtaining the nonlinear dynamic model at the current moment, and after the parameters have been identified and corrected online, the remaining water removal equivalent at the current moment is used as the starting value for integration, and zero is used as the ending value for integration. The reciprocal of the model's water removal power is integrated to obtain the expected completion time considering the nonlinear decay characteristics of future efficiency. This integration operation is the dynamic prediction algorithm applied. Within the range of the remaining water removal equivalent from the starting value to the ending value, the reciprocal of the model's water removal power defined by the nonlinear dynamic model with completed parameter identification is accumulated and summed. The physical meaning of this integration operation is: to calculate the total time required to complete the remaining water removal equivalent under the efficiency change law described by the current nonlinear dynamic model. Compared with the simple division prediction that directly divides the remaining water removal equivalent by the current instantaneous water removal power, the integration method mathematically rigorously handles the nonlinearity of efficiency changes with the process, making the prediction results more in line with physical reality. Especially when the efficiency decays significantly in the middle and later stages, the prediction accuracy is improved more significantly. The nonlinear dynamic model defines the relationship between the model's dehydration power and the remaining dehydration equivalent. It indicates that the model's dehydration power equals an initial dehydration power amplitude parameter multiplied by a power of the natural constant base, where the exponent is the product of the negative efficiency decay rate parameter and the total amount of moisture removed, plus a steady-state dehydration power base parameter. The total amount of moisture removed is obtained by subtracting the current remaining dehydration equivalent from the remaining dehydration equivalent at the start of drying. The remaining dehydration equivalent at the start of drying is calculated by subtracting the target total weight at the end of drying from the total initial weight after loading. In this nonlinear dynamic model, the initial dehydration power amplitude parameter reflects the ability of the grain surface moisture to evaporate rapidly in the early stages of drying. The efficiency decay rate parameter characterizes the rate at which dehydration difficulty increases and efficiency decreases as internal moisture migration becomes dominant; this value is usually positive. The steady-state dehydration power base parameter represents the low-speed dehydration capability corresponding to the extremely difficult-to-remove residual moisture at the end of drying. This nonlinear dynamic model structure can well fit the general law that drying efficiency is initially fast, then slows down, and tends to stabilize. The specific process for predicting the final moisture content of each batch of grain is as follows: First, based on the target total weight and machine weight of the dryer received from the corresponding edge computing module, and the preset target moisture content obtained from the production order; the preset target moisture content is obtained from the production order through the application programming interface provided by the factory manufacturing execution system or enterprise resource planning system, with the data exchange format being JSON or XML, and a dry matter mass estimation operation is performed. This dry matter mass estimation operation subtracts the machine weight from the target total weight to obtain the target net weight, and then multiplies the target net weight by the difference between the target net weight and the preset target moisture content. The calculation result is used as the dry matter mass estimation value of the current batch of grain. Secondly, using the estimated dry matter mass, the real-time total weight received from the corresponding edge computing module, and the machine's own weight, a current instantaneous moisture content estimation operation is performed. This current instantaneous moisture content estimation operation subtracts the machine's own weight from the real-time total weight to obtain the current total grain mass. The difference between this and the quotient of the estimated dry matter mass divided by the current total grain mass is used as the current instantaneous moisture content estimation value of the grain. This current instantaneous moisture content estimation operation is based on the principle of mass conservation. Assuming that the estimated dry matter mass is accurate, it calculates the current total grain mass through the real-time total weight and the machine's own weight, and then indirectly calculates the current instantaneous moisture content estimation value. This current instantaneous moisture content estimation value is a key step in integrating dry matter information with process dynamic weight information. Next, the remaining water removal equivalent at the current moment is obtained from the digital twin of the corresponding dryer. This remaining water removal equivalent is identified as the total amount of water that is expected to be removed from the current moment until the drying is completed. Finally, the final moisture content prediction calculation is performed. This calculation first multiplies the current total mass of grain by the current instantaneous moisture content estimate to obtain an intermediate product. Then, the total amount of moisture expected to be removed is subtracted from this intermediate product to obtain the numerator of the prediction formula. Simultaneously, the total mass of grain is subtracted from the total amount of moisture expected to be removed to obtain the denominator of the prediction formula. Finally, the ratio of the numerator to the denominator is used as the predicted final moisture content of this batch of grain. This final moisture content prediction calculation formula strictly follows the law of conservation of mass. The numerator calculates the mass of the remaining moisture in the grain after the prediction, and the denominator calculates the total mass of the grain after the prediction. The innovation of this formula lies in its comprehensive use of dry matter estimation based on order targets, current state diagnosis based on real-time weight, and future moisture removal prediction based on a nonlinear dynamic model. This enables dynamic and forward-looking prediction of the final moisture content before drying is completed, providing a key input for proactive quality control.

[0022] In this embodiment, it is specifically necessary to explain that the process of performing online multi-objective optimization decision-making and allocating a suitable dryer to a newly arrived drying task in the collaborative scheduling module is as follows: Upon receiving a new drying task, the collaborative scheduling module immediately performs data acquisition and organization operations. These operations include: obtaining the latest status data stored in the digital twins of each dryer in the entire plant from the central processing module, including estimated completion time and instantaneous dehydration power; extracting the initial moisture content of each batch of grain recently processed by the dryer from the historical task records maintained by the central processing module and associated with each dryer, and calculating a mean and a standard deviation based on these initial moisture contents, which together serve as the historical initial moisture content statistical characteristics of the dryer; and then, from the new task's corresponding raw material... In the production order, the initial moisture content, target moisture content, and net weight information of the grain for the new task are obtained; the estimated completion time, instantaneous dehydration power, and historical initial moisture content statistical characteristics of each dryer, along with the initial moisture content, target moisture content, and net weight information of the grain for the new task, are organized together to form a plant-wide decision status set; "Recent" historical task records usually refer to the data of the most recent week or the most recent 50 processing batches, whichever comes first. When maintaining the digital twin, the central processing module synchronously records the task characteristics of each processing batch, forming a queryable historical record to provide data support for scheduling decisions; For each dryer marked as available in the new task and decision state set, a comprehensive suitability score is calculated. The calculation process is as follows: First, a base score component is calculated, which is equal to the weighted sum of a time advantage component and an efficiency advantage component. The time advantage component is obtained by calculating the exponent of the negative estimated completion time divided by a preset time normalization constant, using the natural constant as the base. The efficiency advantage component is obtained by dividing the instantaneous dewatering power of the dryer by the average instantaneous dewatering power of all available dryers in the decision state set. The time advantage component is multiplied by a first time weighting coefficient between zero and one, and the efficiency advantage component is multiplied by a second time weighting coefficient between zero and one, with the sum of the first and second time weighting coefficients being one. The sum of these two products is the base score component. Then, a characteristic matching component is calculated. The calculation process for this characteristic matching component is as follows: Calculate the new... The absolute value of the difference between the initial moisture content of the grain in the task and the mean value in the historical initial moisture content statistical characteristics of the dryer is then divided by a preset adjustment coefficient multiplied by the standard deviation in the historical initial moisture content statistical characteristics of the dryer to obtain a quotient value. Finally, using the natural constant as the base and the negative of the quotient value as the exponent, the value of the exponent is calculated, which is the characteristic matching degree component. Next, a load balancing component is calculated by adding a preset inhibition intensity coefficient to the number one and multiplying it by the total number of tasks currently assigned to the dryer to obtain a sum, and then taking the reciprocal of the sum. Finally, the basic score component, the characteristic matching degree component, and the load balancing component are multiplied consecutively, and the resulting product is the overall suitability score of the dryer for the new task. A preset time normalization constant, whose value can be referenced to the average duration of drying operations in the factory, for example, is set to 2 hours (7200 seconds) to normalize the dimensions of the expected completion time, so that the time advantage component is within a reasonable order of magnitude. The first and second time weighting coefficients are used to weigh the relative importance of time advantage and efficiency advantage. They can be dynamically adjusted according to production strategy. For example, during rush periods, the first time weighting coefficient can be set to 0.7 and the second time weighting coefficient to 0.3, so as to give more preference to equipment that is about to be idle. The preset adjustment coefficient is usually between 1.5 and 3, which is used to adjust the sensitivity of the influence of the dispersion of historical data on the matching degree. When it is desired that the equipment is more focused on processing grains with common moisture ranges, a smaller adjustment coefficient can be selected. "Total number of currently assigned tasks" includes tasks being processed on the dryer as well as assigned tasks that have not yet started. The preset suppression intensity coefficient is usually set between 0.1 and 0.5 to control the intensity of load balancing. The larger the coefficient, the stronger the suppression effect on high-load equipment. After calculating the overall suitability score of all available dryers for the new task, the dryer with the highest overall suitability score is selected as the assignment target for the new task. This overall suitability scoring method achieves intelligent trade-offs among multiple objectives by integrating multiple dimensions such as time, efficiency, historical matching degree and load balancing. Compared with the strategy of only selecting the earliest completed equipment, this method can also take into account the stability of processing efficiency and the load balancing among equipment, thereby improving production efficiency and system robustness at the global level. The process of generating scheduling adjustment instructions for drying tasks that are already being processed is as follows: The collaborative scheduling module periodically or in response to preset events triggers plant-wide rescheduling calculations. Preset events include, but are not limited to: the arrival of new high-priority tasks, the failure or maintenance of a critical dryer, or the monitoring of significant changes in the plant's average load or efficiency exceeding a set threshold. The rescheduling calculation first obtains the plant-wide decision status set and includes all currently processed tasks as objects to be reassigned. Based on this set of decision states, calculate the corresponding comprehensive suitability score for each combination of task to be assigned and each available dryer; To formally describe the allocation relationship between tasks and dryers, a binary decision variable with a value of zero or one is defined for each pair of tasks and available dryers. When the binary decision variable has a value of one, it means that the task is allocated to the dryer, and when it has a value of zero, it means that the task is not allocated. The optimization objective is to maximize the sum of the overall suitability scores of all successfully assigned task-dryer combinations in the entire plant. The sum of the overall suitability scores is obtained by multiplying the overall suitability scores of all tasks and all available dryer combinations by the corresponding binary decision variables, and then adding all the products together. Construct an optimization model whose objective is to maximize the sum of the overall suitability scores, and satisfy the following constraints: For each task to be assigned, there is exactly one binary decision variable corresponding to this task among all binary decision variables; For each available dryer, the total processing load of all tasks assigned to this dryer does not exceed its preset processing capacity limit; The "processing capacity limit" can be defined as the maximum number of tasks that the dryer can process in the next scheduling cycle, such as 3; or, it can be defined as the maximum total net processing weight or total estimated time limit estimated based on its average processing efficiency and cycle duration. A heuristic search algorithm is used to solve the optimization model online, and an optimal or near-optimal task-dryer redistribution scheme is obtained. The heuristic search algorithm can be the tabu search algorithm, because it can find a high-quality near-optimal solution for medium-sized integer programming problems in a reasonable time, and avoids getting trapped in local optima through the tabu list mechanism, which is suitable for the real-time requirements of online scheduling. By comparing this redistribution scheme with the current actual operating status, tasks that need to be migrated from their current dryer to another dryer are identified. For each task requiring migration, a scheduling adjustment instruction is generated. This instruction contains a series of ordered control instructions: First, a control instruction is sent to the programmable logic controller (PLC) of the source dryer where the task is currently located, ordering it to perform an early grain discharge preparation operation; simultaneously, another control instruction is sent to the PLC of the target dryer specified in the redistribution scheme, ordering it to prepare to receive grain and continue the drying operation; then, the current process status data of the task at the source dryer, including its current instantaneous moisture content estimate and remaining water removal equivalent, is transmitted to the PLC of the target dryer; the target dryer... After receiving this status data, the programmable logic controller (PLC) of the dryer uses it as a virtual initial state of operation, recalculates the target total weight required to complete the drying operation, and takes over and continues the drying control of this batch of grain based on this target total weight. Finally, all control commands are sent to the corresponding PLCs through the factory control network and wait for confirmation signals from each controller. The task migration will only be actually executed if the following conditions are met: the current remaining dehydration equivalent of the in-process task is greater than a preset migration safety threshold to ensure sufficient drying capacity to complete the subsequent process on the target equipment; and both the source and target equipment are in a state where they can be safely interrupted and received. The migration safety threshold can be set according to the grain variety and process experience, for example, 20% of the initial dehydration amount. The programmable logic controller of the target dryer uses the received "estimated real-time moisture content" as the virtual initial moisture content of the migration task and the "remaining water removal equivalent" as the virtual initial remaining water removal amount. Combined with the dry matter mass of the batch of grain (which can be obtained from the original task order or transferred by the central processing module), the target controller recalculates the remaining target total weight of this drying operation, thereby achieving a smooth relay of the drying process and ensuring the consistency of the final moisture content.

[0023] In this embodiment, it is specifically necessary to explain that in the control optimization module, the process of comparing the predicted final moisture content distribution of all batches of grain in processing with the preset global moisture control target range and calculating the adjustment amount used to fine-tune the target total weight of each dryer is as follows: The control optimization module first periodically retrieves the predicted final moisture content of all batches being processed in the entire plant from the central processing module, forming the current moisture prediction dataset. Based on this current moisture prediction dataset, robust distribution parameter estimation is performed. This robust distribution parameter estimation process includes: A1, calculating the initial arithmetic mean and initial standard deviation of the current moisture prediction dataset; A2, calculating a robust weight for each predicted final moisture content value in the current moisture prediction dataset. The calculation process for this robust weight is as follows: using the natural constant as the base and a negative fraction as the exponent, calculate the value of the exponent power, where the numerator of the fraction is... The square of the difference between the predicted final moisture content and the initial arithmetic mean is calculated, with the denominator being the square of the product of twice the initial standard deviation and a preset robustness factor. A3. Using all the calculated robustness weights, the values ​​of all predicted final moisture content in the current moisture prediction dataset are weighted, and the weighted arithmetic mean is recalculated as the robust center value. The weighted standard deviation is also calculated as the robust dispersion. The preset robustness factor is usually between 2.0 and 3.0, used to define what degree of deviation is considered an "outlier". The larger the factor, the higher the tolerance of the algorithm to outliers, and the more robust the estimation results. Next, a global consistency deviation index is calculated. This index is based on a preset global moisture control target interval, robust center value, and robust dispersion. Specifically: the global moisture control target interval has a lower bound and an upper bound; the target interval center value is obtained by adding the lower bound and the upper bound and dividing by two; the target interval half-width is obtained by subtracting the lower bound from the upper bound and dividing by two; the difference between the robust center value and the target interval center value yields the distribution center offset; the robust dispersion is divided by the target interval half-width to obtain the relative dispersion; and a preset mean offset weighting coefficient is multiplied by the distribution center offset to obtain the mean offset. The dispersion term is obtained by multiplying a preset dispersion weight coefficient by the relative dispersion. The mean offset term and the dispersion term are added together to form the global consistency deviation index. Both the mean offset weight coefficient and the dispersion weight coefficient are real numbers between zero and one, and their sum is one. The setting of the mean offset weight coefficient and the dispersion weight coefficient reflects the trade-off between the two optimization objectives of "overall moisture position" and "moisture concentration". In practical applications, in order to emphasize the consistency of finished product moisture, the dispersion weight coefficient is usually set to a larger value, such as 0.6 to 0.8, while the mean offset weight coefficient is set to 0.4 to 0.2 accordingly. Finally, for each dryer in operation, calculate the adjustment amount for the target total weight using the following steps: B1. The product of a negative proportional control coefficient and the difference between the predicted final moisture content of the dryer and the center value of the target interval, minus the product of a differential control coefficient and the rate of change of the predicted final moisture content of the dryer, is used as the intended value for local moisture adjustment. The proportional control coefficient and the differential control coefficient can be obtained through simulation or on-site tuning based on the dynamic characteristics of the drying process. The proportional control coefficient mainly determines the correction strength for static deviations, with a typical value range, for example, between 0.5 and 2.0. The differential control coefficient is used to suppress overshoot that may be caused by rapid changes in the predicted value, with a typical value range, for example, between 0.1 and 0.5. B2. Based on the robust center value, robust dispersion, target interval center value, and the dryer's own predicted final moisture content, calculate a gradient value. This gradient value characterizes the degree of influence of the dryer's predicted final moisture content on the global consistency deviation index. A specific calculation method for the gradient value is: Gradient value = (predicted final moisture content - robust center value) / (robust dispersion^2) + η * (robust center value - target interval center value) / (target interval half-width^2), where η is a coordination coefficient. The first term of this formula gives batches that deviate further from the distribution center a greater adjustment driving force (positive value reduces it, negative value increases it), thereby compressing the distribution width. The second term provides a unidirectional fine-tuning trend for all batches when the overall distribution center deviates from the target, jointly pulling the distribution center closer to the target. B3. Subtract the product of a global coordination gain coefficient and the gradient value from the local moisture adjustment intention value to obtain a comprehensive adjustment intention value. Then divide this comprehensive adjustment intention value by the sensitivity of the dryer's predicted final moisture content to changes in the target total weight. The sensitivity indicates that an increase in the target total weight will lead to an increase in the predicted final moisture content. The resulting quotient is the initial adjustment amount for the target total weight of the dryer. The sensitivity of the predicted final moisture content to changes in the target total weight can be derived offline through the mass conservation model of the drying process. Its theoretical expression is related to the dry matter mass of the grain. Alternatively, it can be estimated online by analyzing the historical operating data of the dryer and statistically analyzing the average trend of its predicted final moisture content changing with the target total weight. B4. Compare the initial adjustment amount with a preset maximum allowable positive adjustment amount and a preset maximum allowable negative adjustment amount. If the initial adjustment amount is greater than the maximum allowable positive adjustment amount, the maximum allowable positive adjustment amount is taken as the final adjustment amount. If the initial adjustment amount is less than the maximum allowable negative adjustment amount, the maximum allowable negative adjustment amount is taken as the final adjustment amount. Otherwise, the initial adjustment amount is taken as the final adjustment amount. The final adjustment amount is the adjustment amount for the target total weight to be issued. The maximum allowable positive adjustment amount and the maximum allowable negative adjustment amount are calculated in reverse based on the maximum allowable moisture change range for a single adjustment. For example, if the allowable moisture change caused by a single adjustment is no more than 0.5%, the corresponding target total weight adjustment limit can be estimated based on the dry matter mass of the current batch. This ensures the smoothness of the control action and avoids excessive disturbance to the drying process. The specific process of sending the adjustment amount to the edge computing module of the corresponding dryer is as follows: The control optimization module constructs a target weight update instruction for each dryer that needs to adjust its target total weight. The instruction includes the dryer's device identifier, the new target total weight value, the instruction sequence number, and the timestamp. The new target total weight value is formed by adding the adjustment amount of the target total weight to the dryer's current target total weight. Before the target weight update command is issued, the control optimization module executes a safety verification logic. This safety verification logic ensures that the new target total weight value is greater than the current real-time total weight obtained from the dryer's edge computing module, and that the new target total weight value is less than a preset safe weight upper limit. The preset safe weight upper limit is usually set as a safe percentage of the dryer's rated capacity (e.g., 85%), or is determined based on the dryer's mechanical structure and historical maximum load records. This is used to prevent equipment overload or safety accidents caused by setting the target too high. Through the factory control network, the target weight update command that has passed the verification is sent to the edge computing module of the corresponding dryer, i.e. its programmable logic controller, using a secure and reliable industrial communication protocol. The secure and reliable industrial communication protocol can be the OPCUA protocol based on time-sensitive networking. Such protocols can guarantee the real-time performance, determinism and data integrity of command transmission, meeting the stringent requirements of industrial control scenarios. Upon receiving the target weight update command, the programmable logic controller immediately replaces the original target total weight stored in its internal memory with the new target total weight value contained in the target weight update command. This new target total weight value is used as the new benchmark for determining the stop of the subsequent drying process, and the programmable logic controller records that this target total weight update comes from global optimization control. At the same time, the programmable logic controller sends a confirmation report to the central processing module. After receiving the confirmation report, the central processing module updates the target total weight parameter stored in the digital twin of the corresponding dryer to the new target total weight value. After updating the target total weight, the programmable logic controller (PLC) also initiates a temporary high-frequency process status monitoring to observe the dynamic response of the drying process after adjusting the target total weight. The temporary high-frequency process status monitoring refers to increasing the sampling frequency of the weight sensor to several times the original frequency (e.g., 5Hz) for a short period of time (e.g., 5 minutes) and closely monitoring the changing trend of the remaining water equivalent. If the changing trend is found to be abnormal or seriously inconsistent with the expectations, the PLC can trigger an alarm or automatically revert to the target total weight before the adjustment, thus achieving safety self-protection in closed-loop control.

[0024] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0025] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0026] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0027] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0028] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0029] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0030] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A smart measurement and control system for grain moisture in a dryer based on weight change, characterized in that, Specifically, it includes: The system includes edge computing modules deployed in each dryer, a central processing module deployed in the factory control center, a collaborative scheduling module, and a control optimization module. Edge computing module: In response to the start of grain loading in the dryer, it obtains the dryer's own weight, the total initial weight after loading, and the initial moisture content of the grain, calculates the net weight of the loaded grain, and calculates the target total weight when drying is completed based on the preset target moisture content. During the drying process, the total weight is continuously collected in real time by a weight sensor installed in the dryer. Based on this, the remaining dewatering equivalent, which represents the current remaining dewatering workload, and the instantaneous dewatering power, which represents the instantaneous drying efficiency, are dynamically calculated and output to the central processing module. The central processing module receives and integrates the remaining dehydration equivalent and instantaneous dehydration power data uploaded by all edge computing modules, establishes and maintains a dynamic digital twin for each dryer, and updates the state of the digital twin driven by the real-time received remaining dehydration equivalent and instantaneous dehydration power. Based on the updated state of the digital twin, a dynamic prediction algorithm is applied to estimate the expected completion time of each dryer, and further, based on the received remaining dehydration equivalent and instantaneous dehydration power, the predicted final moisture content of each batch of grain is predicted. Collaborative scheduling module: Based on the estimated completion time and instantaneous dewatering power of each dryer provided by the central processing module, it performs online multi-objective optimization decision-making, allocates suitable dryers to newly arrived drying tasks, and generates scheduling adjustment instructions for drying tasks that are already being processed; Control optimization module: Based on the predicted final moisture content distribution of all batches of grain in processing provided by the central processing module, it compares it with the preset global moisture control target range and calculates the adjustment amount used to fine-tune the target total weight of each dryer; the adjustment amount is sent to the edge computing module of the corresponding dryer, so that the edge computing module controls the drying process according to the adjusted target total weight.

2. The intelligent measurement and control system for grain moisture in a dryer based on weight change as described in claim 1, characterized in that: In the edge computing module, the specific process of calculating the net weight of the grain and, based on the preset target moisture content, calculating the target total weight upon completion of drying is as follows: First, subtract the dryer's own weight from the total initial weight after loading the grain to obtain the net weight of the loaded grain. Then, multiply the net weight of the loaded grain by the difference between the initial moisture content of the grain and the net weight of the grain to obtain an intermediate value. Next, divide this intermediate value by the difference between the preset target moisture content and the net weight of the grain to obtain a calculation result. Finally, add this calculation result to the dryer's own weight, and the sum obtained is the target total weight when drying is completed.

3. The intelligent measurement and control system for grain moisture in a dryer based on weight change as described in claim 2, characterized in that: The process of dynamically calculating the remaining dewatering equivalent, which represents the current remaining dewatering workload, and the instantaneous dewatering power, which represents the instantaneous drying efficiency, is as follows: During the drying process, the real-time total weight is continuously collected at preset fixed time intervals. The remaining water removal equivalent is obtained by subtracting the target total weight when drying is completed from the latest collected real-time total weight. Instantaneous water removal power is calculated by subtracting the total real-time weight collected at the previous sampling time from the total real-time weight collected at the current sampling time to obtain a weight difference. The weight difference is then divided by a fixed time interval to obtain an original rate of change value. A dynamic smoothing coefficient, determined by a basic smoothing factor and a sensitivity coefficient, is introduced. The original rate of change value at the current sampling time is then weighted and mixed with the smoothed rate of change value calculated at the previous sampling time. The weight of the original rate of change value at the current sampling time is the dynamic smoothing coefficient, and the weight of the smoothed rate of change value at the previous sampling time is one minus the dynamic smoothing coefficient. The smoothed rate of change value at the current sampling time is obtained by mixing and calculating as the instantaneous water removal power. The edge computing module packages the remaining water removal equivalent and instantaneous water removal power, along with the device identifier and time information, into a unified data frame and uploads it.

4. The intelligent measurement and control system for grain moisture in a dryer based on weight change as described in claim 3, characterized in that: In the central processing module, the process of establishing and maintaining a dynamic digital twin for each dryer, and applying a dynamic prediction algorithm based on the updated digital twin status to estimate the expected completion time, is as follows: The central processing module constructs a corresponding digital twin for each dryer. Each digital twin continuously receives and stores pairing information consisting of the remaining water removal equivalent and instantaneous water removal power uploaded from the corresponding edge computing module. This drives the digital twin's status to synchronize with the dryer in real time and maintains a dynamic data set that stores multiple sets of such pairing information recently. Based on the recent historical pairing information stored in the dynamic dataset, a nonlinear dynamic model is constructed for each dryer to describe the change in its dehydration efficiency as the drying process progresses. A recursive parameter estimation algorithm with a forgetting factor is adopted. By utilizing continuously updated historical pairing information, the goal is to make the model's dewatering power approximate the actual received instantaneous dewatering power. The initial dewatering power amplitude parameter, efficiency decay rate parameter, and steady-state dewatering power basis parameter in the nonlinear dynamic model are identified and corrected online in real time. After obtaining the nonlinear dynamic model at the current moment, and after the parameters have been identified and corrected online, the remaining water removal equivalent at the current moment is used as the starting value for integration, and zero is used as the ending value for integration. The inverse of the water removal power of the model is integrated to obtain the expected completion time considering the nonlinear decay characteristics of future efficiency. This integration operation is the dynamic prediction algorithm used.

5. The intelligent measurement and control system for grain moisture in a dryer based on weight change as described in claim 4, characterized in that: The nonlinear dynamic model defines the relationship between the model's dewatering power and the remaining dewatering equivalent. It indicates that the model's dewatering power is equal to an initial dewatering power amplitude parameter multiplied by a power of the natural constant base, where the exponent of the power is the product of the negative efficiency decay rate parameter and the total amount of moisture removed, plus a steady-state dewatering power base parameter. The total amount of moisture removed is obtained by subtracting the remaining dewatering equivalent at the start of drying from the remaining dewatering equivalent at the current moment. The remaining dewatering equivalent at the start of drying is calculated by subtracting the target total weight at the end of drying from the total initial weight after loading.

6. The intelligent measurement and control system for grain moisture in a dryer based on weight change as described in claim 5, characterized in that: The process for predicting the final moisture content of each batch of grain is as follows: First, based on the target total weight of the dryer, the machine body weight received from the corresponding edge computing module, and the preset target moisture content obtained from the production order, a dry matter mass estimation operation is performed. This dry matter mass estimation operation subtracts the machine body weight from the target total weight to obtain the target net weight, and then multiplies the target net weight by the difference between the target net weight and the preset target moisture content. The calculation result is used as the dry matter mass estimation value of the current batch of grain. Secondly, using the estimated dry matter mass, the real-time total weight received from the corresponding edge computing module, and the machine's own weight, the current instantaneous moisture content estimation operation is performed. This current instantaneous moisture content estimation operation subtracts the machine's own weight from the real-time total weight to obtain the current total grain mass. Then, the difference between the estimated dry matter mass and the quotient of the current total grain mass is used as the estimated instantaneous moisture content of the grain at the current moment. Next, the remaining water removal equivalent at the current moment is obtained from the digital twin of the corresponding dryer. This remaining water removal equivalent is identified as the total amount of water that is expected to be removed from the current moment until the drying is completed. Finally, the final moisture content prediction calculation is performed. This final moisture content prediction calculation first multiplies the current total mass of grain by the current instantaneous moisture content estimate to obtain an intermediate product. Then, the total amount of moisture expected to be removed is subtracted from this intermediate product to obtain the numerator of the prediction formula. At the same time, the total mass of grain is subtracted from the total amount of moisture expected to be removed to obtain the denominator of the prediction formula. Finally, the ratio obtained by dividing the numerator by the denominator is used as the predicted final moisture content of this batch of grain.

7. The intelligent measurement and control system for grain moisture in a dryer based on weight change as described in claim 6, characterized in that: In the collaborative scheduling module, the process of performing online multi-objective optimization decision-making and allocating a suitable dryer to a newly arrived drying task is as follows: Upon receiving a new drying task, the collaborative scheduling module immediately performs data acquisition and organization operations. These operations include: obtaining the latest status data stored in the digital twins of each dryer in the entire plant from the central processing module, including the estimated completion time and instantaneous dehydration power; extracting the initial moisture content of each batch of grain recently processed by the dryer from the historical task records maintained by the central processing module and associated with each dryer, and calculating a mean and a standard deviation based on these initial moisture contents, which together serve as the historical initial moisture content statistical features of the dryer; obtaining the initial moisture content, target moisture content, and net weight information of the grain for the new task from the production order corresponding to the new task; and organizing the estimated completion time, instantaneous dehydration power, historical initial moisture content statistical features of each dryer, and the initial moisture content, target moisture content, and net weight information of the grain for the new task into a plant-wide decision status set. For each dryer marked as available in the new task and decision state set, a comprehensive suitability score is calculated. The calculation process is as follows: First, a base score component is calculated, which is equal to the weighted sum of a time advantage component and an efficiency advantage component. The time advantage component is obtained by calculating the exponent of the negative estimated completion time divided by a preset time normalization constant, using the natural constant as the base. The efficiency advantage component is obtained by dividing the instantaneous dewatering power of the dryer by the average instantaneous dewatering power of all available dryers in the decision state set. The time advantage component is multiplied by a first time weighting coefficient between zero and one, and the efficiency advantage component is multiplied by a second time weighting coefficient between zero and one. The two products are then added together, and the result is the base score component. Then, the calculation... A characteristic matching degree component is calculated as follows: The absolute value of the difference between the initial moisture content of the grain for the new task and the mean value in the historical initial moisture content statistical characteristics of the dryer is calculated; this absolute value is then divided by the product of a preset adjustment coefficient and the standard deviation in the historical initial moisture content statistical characteristics of the dryer, yielding a quotient; finally, using the natural constant as the base and the negative of the quotient as the exponent, the value of the exponent is calculated, which is the characteristic matching degree component; next, a load balancing component is calculated, which is obtained by adding a preset inhibition intensity coefficient to the product of the number one and the total number of tasks currently assigned to the dryer, obtaining a sum, and then taking the reciprocal of the sum; finally, the basic score component, the characteristic matching degree component, and the load balancing component are multiplied consecutively, and the resulting product is the overall suitability score of the dryer for the new task; After calculating the overall suitability score of all available dryers for the new task, the dryer with the highest overall suitability score is selected as the assignment target for the new task.

8. The intelligent measurement and control system for grain moisture in a dryer based on weight change as described in claim 7, characterized in that: The process of generating scheduling adjustment instructions for drying tasks that are already being processed specifically involves: The collaborative scheduling module periodically or in response to preset events triggers a plant-wide rescheduling calculation. The rescheduling calculation first obtains the decision status set of the entire plant and includes all currently processed tasks as objects to be reassigned. Based on this set of decision states, calculate the corresponding comprehensive suitability score for each combination of task to be assigned and each available dryer; To formally describe the allocation relationship between tasks and dryers, a binary decision variable with a value of zero or one is defined for each pair of tasks and available dryers. When the binary decision variable has a value of one, it means that the task is allocated to the dryer, and when it has a value of zero, it means that the task is not allocated. The optimization objective is to maximize the sum of the overall suitability scores of all successfully assigned task-dryer combinations in the entire plant. The sum of the overall suitability scores is obtained by multiplying the overall suitability scores of all tasks and all available dryer combinations by the corresponding binary decision variables, and then adding all the products together. Construct an optimization model whose optimization objective is to maximize the sum of comprehensive suitability scores, and satisfy the following constraints: for each task to be assigned, there is exactly one binary decision variable corresponding to this task among all binary decision variables; for each available dryer, the total processing burden of all tasks assigned to this dryer does not exceed its preset processing capacity limit. A heuristic search algorithm was used to solve the optimization model online, resulting in an optimal or near-optimal task-dryer redistribution scheme. By comparing this redistribution scheme with the current actual operating status, the in-process tasks that need to be migrated from the current dryer to another dryer are identified. For each task that needs to be migrated, a scheduling adjustment instruction is generated, which contains a series of ordered control instructions.

9. The intelligent measurement and control system for grain moisture in a dryer based on weight change as described in claim 8, characterized in that: In the control optimization module, the process of comparing the predicted final moisture content distribution of all batches of grain in processing with the preset global moisture control target range and calculating the adjustment amount for fine-tuning the target total weight of each dryer is as follows: The control optimization module first periodically obtains the predicted final moisture content of all batches being processed in the entire plant from the central processing module, forming the current moisture prediction dataset. Based on this current moisture prediction dataset, robust distribution parameter estimation is performed. The robust distribution parameter estimation process includes: A1. Calculating the initial arithmetic mean and initial standard deviation of the current moisture prediction dataset; A2. Calculating a robust weight for each predicted final moisture content value in the current moisture prediction dataset; A3. Using all the calculated robust weights, weighting all predicted final moisture content values ​​in the current moisture prediction dataset, recalculating the weighted arithmetic mean as the robust center value, and calculating the weighted standard deviation as the robust dispersion. Next, a global consistency deviation index is calculated, which is based on the preset global moisture control target interval, robust center value, and robust dispersion. Finally, for each dryer in operation, calculate the adjustment amount for the target total weight using the following steps: B1. The product of a negative proportional control coefficient and the difference between the predicted final moisture content of the dryer and the center value of the target interval, minus the product of a differential control coefficient and the rate of change of the predicted final moisture content of the dryer, is used as the intended value for local moisture adjustment. B2. Calculate a gradient value based on the robust central value, robust dispersion, target interval central value, and the dryer's own predicted final moisture content. B3. Subtract the product of a global coordination gain coefficient and the gradient value from the local moisture adjustment intention value to obtain a comprehensive adjustment intention value; then divide this comprehensive adjustment intention value by the sensitivity of the dryer's predicted final moisture content to the change in the target total weight, and the resulting quotient is the initial adjustment amount for the target total weight of the dryer. B4. Compare the initial adjustment amount with a preset maximum allowable positive adjustment amount and a preset maximum allowable negative adjustment amount. If the initial adjustment amount is greater than the maximum allowable positive adjustment amount, the maximum allowable positive adjustment amount is taken as the final adjustment amount. If the initial adjustment amount is less than the maximum allowable negative adjustment amount, the maximum allowable negative adjustment amount is taken as the final adjustment amount. Otherwise, the initial adjustment amount is taken as the final adjustment amount. The final adjustment amount is the adjustment amount for the target total weight to be issued.

10. The intelligent measurement and control system for grain moisture in a dryer based on weight change according to claim 9, characterized in that: The process of sending the adjustment amount to the edge computing module of the corresponding dryer is as follows: The control optimization module constructs a target weight update instruction for each dryer that needs to adjust its target total weight. The instruction includes the dryer's device identifier, the new target total weight value, the instruction sequence number, and the timestamp. The new target total weight value is formed by adding the adjustment amount of the target total weight to the dryer's current target total weight. Before the target weight update command is issued, the control optimization module executes security verification logic; Through the factory control network, the target weight update command that has passed the verification is sent to the edge computing module of the corresponding dryer, i.e. its programmable logic controller, using a secure and reliable industrial communication protocol. Upon receiving the target weight update command, the programmable logic controller immediately replaces the original target total weight stored in its internal memory with the new target total weight value contained in the target weight update command. This new target total weight value is used as the new benchmark for determining the stop of the subsequent drying process, and the programmable logic controller records that this target total weight update comes from global optimization control. At the same time, the programmable logic controller sends a confirmation report to the central processing module. After receiving the confirmation report, the central processing module updates the target total weight parameter stored in the digital twin of the corresponding dryer to the new target total weight value. After updating the target total weight, the programmable logic controller also initiates a temporary high-frequency process status monitoring to observe the dynamic response of the drying process after adjusting the target total weight.

Images