Monitoring Method for Fermentation of Livestock and Poultry Manure Based on Digital Twin

Abstract

The present invention discloses a monitoring method for livestock and poultry manure fermentation based on digital twin, which relates to the technical field of livestock and poultry manure treatment; by constructing a high-precision digital twin model, real-time synchronizing the sensing data on the fermentation site, simulating the fermentation state and predicting the development trend; setting an intelligent control strategy that integrates data relevance and control priority to dynamically identify and avoid rule conflicts; on this basis, automatically executing control actions such as turning, spraying and ventilation, recording feedback data and continuously optimizing the parameters of the twin model based on the feedback results; this method realizes the intelligent regulation of the fermentation process and the adaptive update of the model, and significantly improves the stability, energy utilization efficiency and prediction accuracy of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of livestock and poultry manure treatment, and particularly to a monitoring method for livestock and poultry manure fermentation based on digital twin. Background Art

[0002] Monitoring of livestock and poultry manure fermentation refers to the process of real-time or periodic monitoring and regulation of various parameters during the fermentation process of livestock and poultry manure, aiming to ensure that the fermentation process is efficient, stable and meets environmental protection requirements. The system is set with intelligent control based on a rule engine, such as "if the temperature > 65°C and the humidity < 30%, then turn on the turning pile + spraying", and there is another rule "if the temperature < 55°C after turning the pile, then stop spraying".

[0003] The prior art has the following deficiencies: In extreme states such as continuous high temperature and dryness, the two rules of the system may have cyclic conflicts, constantly triggering mutually exclusive conditions, thus entering a logical dead loop, resulting in frequent startup and shutdown of the turning pile / spraying equipment in the execution layer. In addition, if there are logical conflicts in the rule engine, it will not only cause frequent startup and shutdown of equipment and a sharp increase in energy consumption, but also seriously interfere with the accurate mapping and prediction judgment of the fermentation state by the digital twin model. Summary of the Invention

[0004] The purpose of the present invention is to provide a monitoring method for livestock and poultry manure fermentation based on digital twin to solve the deficiencies in the background art.

[0005] To achieve the above purpose, the present invention provides the following technical solutions: A monitoring method for livestock and poultry manure fermentation based on digital twin, including:

[0006] Construct a digital twin model of the livestock and poultry manure fermentation process, collect real-time data of the fermentation heap through sensors deployed at the fermentation site, and synchronize the real-time data into the digital twin model;

[0007] Simulate the actual fermentation state in the digital twin model, and predict the fermentation development trend and potential anomalies;

[0008] Set an intelligent control strategy, which combines the mutual influence degree between real-time data and the execution conditions of the control action priority, judges whether there are logical conflicts in the rule engine, and if so, optimizes the control strategy;

[0009] The control strategy constructs a comprehensive feature vector based on the coupling relationship between the predicted value of oxygen utilization efficiency and the estimated value of surface drying rate, combines the preset control action priority, predicts the risk value of logical conflicts in the rule engine through a machine learning model, and compares it with a predetermined threshold to judge whether there are logical conflicts;

[0010] Automatically perform turning, spraying, and ventilation operations according to the optimized control strategy, and record control behaviors and feedback data in real time. Continuously adjust the digital twin model based on the recorded feedback data.

[0011] Preferably, the real-time data collected by the sensor includes the temperature, humidity, oxygen concentration, pH value, and gas release amount at different layers of the compost pile, which are used to construct the input data set of the multi-dimensional and hierarchical twin model.

[0012] Preferably, a graph neural network is used to predict the fermentation development trend and potential anomalies, specifically including:

[0013] Divide the entire fermentation compost pile into multiple monitoring sub-regions, and each sub-region corresponds to a node in a graph; the connection relationship between nodes is established based on the spatial adjacency relationship of the compost pile, heat or moisture migration paths; each node is associated with multiple sensor data to form a feature vector of the node; construct the initial graph structure of the compost pile G=(V,E,X), where: V is the set of nodes; E is the set of edges; X is the node feature matrix;

[0014] Represent the state of the compost pile at each moment as a graph Gt, and continuous time forms a graph sequence; combine the node feature changes and graph structure invariance in each graph to construct the input of time-series graph data; the label data is the target variable or anomaly status mark of each node within a future time window.

[0015] Use a graph convolutional network or a graph attention network to extract the spatio-temporal features of the nodes; the GNN model performs convolutional operations on the graph Gt at each time step to obtain the embedded representation Ht of the nodes.

[0016] Input the node representation Ht into the subsequent time series prediction module; predict the temperature or humidity variable of each node at future time steps; combine the actual historical distribution to perform anomaly scoring on the prediction results; if the predicted value of a certain node deviates from the normal change trend, it is marked as a potential anomaly.

[0017] Preferably, the method for obtaining the predicted value of oxygen utilization efficiency is: construct a causal relationship graph of oxygen utilization, determine the variable nodes related to oxygen utilization efficiency, including the temperature, humidity or moisture content, oxygen concentration, and ventilation volume of the fermentation compost pile, establish the causal dependence relationship between variables to form a directed acyclic graph, and set the predicted value of oxygen utilization efficiency as the target variable node among them.

[0018] Collect real-time monitoring data during the historical fermentation process for training the Bayesian network.

[0019] During the fermentation process, the observed values ​​of temperature, humidity, oxygen concentration and ventilation volume are collected in real time; the observed values ​​are input into the Bayesian network as known nodes; the conditional propagation algorithm in the Bayesian network is triggered using the input of the observed nodes; the network automatically updates the posterior probabilities of all relevant nodes; the probability distribution result of the oxygen utilization efficiency prediction value node is obtained as the current oxygen utilization efficiency prediction value.

[0020] Preferably, the method for obtaining the surface drying rate estimate is:

[0021] High-sensitivity humidity sensors are deployed on the surface of the fermentation pile to collect data at fixed time intervals; each data sequence represents the change trajectory of surface humidity within a fixed time window, forming a humidity time series sample; an autoencoder is used to perform unsupervised feature learning on the humidity time series: the encoder part automatically extracts low-dimensional feature vectors representing the drying change trend in each time series; the decoder part attempts to restore the feature vector to the original humidity curve; after training, the encoder compresses any newly collected humidity change sequence into a fixed-dimensional feature representation;

[0022] A clustering algorithm is used to cluster the feature vectors extracted by the autoencoder; each clustering result represents a typical drying mode; after clustering, a corresponding drying rate label is assigned to each type of drying mode; during the fermentation process, the system continuously collects the current surface humidity time series; the trained autoencoder is used to perform feature compression on the current sequence; the feature vector is matched with the cluster center to determine which drying mode the current sequence belongs to; based on the category, the surface drying rate estimate for the current time period is output accordingly.

[0023] Preferably, the predicted value of oxygen utilization efficiency and the estimated surface drying rate are converted into a comprehensive feature vector, and the comprehensive feature vector is used as the input of the machine learning model. Combined with the currently set control action priority, the machine learning model uses each group of comprehensive feature vectors to predict the risk value label of logical conflict in the rule engine as the prediction target, and takes minimizing the sum of prediction errors of risk value labels of logical conflict in all rule engines as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The risk value of logical conflict in the rule engine is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0024] Preferably, the acquired risk value of the logical conflict in the rule engine is compared with a predetermined threshold value. If the risk value of the logical conflict in the rule engine is greater than or equal to the predetermined threshold value, it means that the logical conflict exists in the rule engine. At this time, a warning signal is generated and the control strategy is optimized; if the risk value of the logical conflict in the rule engine is less than the predetermined threshold value, it means that the logical conflict does not exist in the rule engine. At this time, no warning signal is generated and no additional adjustment is required.

[0025] Preferably, the digital twin model is continuously adjusted based on the recorded feedback data, specifically including: in the initial stage, a digital twin model is constructed to simulate the livestock and poultry manure fermentation process, and a data structure for recording is established simultaneously to associate the corresponding relationship between each control behavior, environmental response, and model output; multiple small sample task sets are constructed using historical feedback data, each task represents a specific fermentation sub-scenario, and each task contains a training set and a validation set; the digital twin model is gradually fine-tuned in each task; through repeated iterations of multiple tasks, the initial parameters of the model are optimized. In actual operation, after each control behavior is executed, the collected feedback data is immediately constructed into a new task; the new task is input into the trained meta-learning framework for iteration to locally fine-tune the current twin model, and the fine-tuned model is retained as the current scenario sub-model; the system regularly verifies the prediction performance of the current twin sub-model; if the error decreases or meets the self-set threshold, the system will extract the changed parameters after fine-tuning, feedback them to the main model parameter set, and update the main model or add them to the empirical model pool.

[0026] In the above technical solution, the technical effects and advantages provided by the present invention are as follows:

[0027] 1. By constructing a digital twin model, the present invention performs high-fidelity and dynamic simulation on the whole process of livestock and poultry manure fermentation, and combines advanced algorithms such as graph neural network, Bayesian network, and autoencoder to realize the intelligent acquisition of key parameters such as fermentation trend prediction, potential anomaly identification, oxygen utilization efficiency evaluation, and surface drying rate judgment. By fusing the above key features into a comprehensive vector and introducing a polynomial regression model for predicting the risk of rule engine conflicts, the logical mutual exclusion problem existing in the traditional rule engine under extreme working conditions can be effectively identified, early warnings can be generated in advance, and control strategies can be optimized to avoid the risk of the system entering an infinite loop or the equipment starting and stopping frequently.

[0028] 2. The present invention introduces a meta-learning algorithm to continuously and adaptively optimize the digital twin model, enabling the model to quickly adjust under variable fermentation materials, climate environments, and operating conditions, and improving the adaptability and prediction accuracy of the model in new scenarios. This method realizes a closed-loop intelligent control process, which not only enhances the stability, intelligence, and energy efficiency level of the fermentation process, but also provides an expandable and highly reliable digital solution for the resource utilization of livestock and poultry breeding waste. Description of the Drawings

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0030] Figure 1 This is the flowchart of the method of the present invention. Specific embodiments

[0031] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0032] For the embodiments, please refer to Figure 1 As shown, the method for monitoring the fermentation of livestock and poultry manure based on digital twin in this embodiment includes:

[0033] Construct a digital twin model of the livestock and poultry manure fermentation process, collect real-time data of the fermentation heap through sensors deployed at the fermentation site, and synchronize the real-time data into the digital twin model;

[0034] Simulate the actual fermentation state in the digital twin model to predict the fermentation development trend and potential anomalies;

[0035] Set intelligent control strategies. The control strategies combine the mutual influence degree between real-time data and the execution conditions of the control action priorities, and judge whether there are logical conflicts in the rule engine. If so, optimize the control strategies;

[0036] Automatically execute turning, spraying and ventilation operations according to the optimized control strategies, and record the control behaviors and feedback data in real time, and continuously adjust the digital twin model based on the recorded feedback data.

[0037] Constructing a digital twin model of the livestock and poultry manure fermentation process means performing high-fidelity modeling and simulation of the actual fermentation system in a virtual environment, specifically including:

[0038] Establish a geometric model of the fermentation heap, including parameters such as the length, width, height, hierarchical structure, and material stacking density of the heap; simulate the heat conduction path, moisture migration channel and ventilation hole layout inside the heap; consider the influence of different material types (such as pig manure, chicken manure, cow manure) on the heap structure stability and heat flow characteristics.

[0039] Based on the principle of aerobic fermentation, introduce a microbial metabolic kinetics model to simulate processes such as organic matter decomposition, carbon-nitrogen cycle, and heat release; simulate the laws of the reaction rate changing with temperature, pH, and moisture in different fermentation stages (heating stage, high-temperature stage, cooling stage); introduce an oxygen consumption model and by-product release models such as carbon dioxide and ammonia to enhance the environmental response ability of the twin model.

[0040] Simulate the effects of environmental temperature, humidity, wind speed, etc. on the fermentation process of the heap; introduce weather data interfaces or edge meteorological station information to enable the model to have the ability to dynamically perceive the external environment; consider the differences in ventilation efficiency and heat exchange efficiency caused by the on-site layout.

[0041] To achieve the dynamic update of the digital twin model, various types of sensors need to be deployed at the fermentation site to continuously monitor the internal and environmental parameters of the heap, mainly including:

[0042] Temperature sensors: used to monitor the temperature distribution at different depths (surface layer, middle layer, bottom layer) of the heap; humidity sensors: monitor the moisture content of the material or the relative humidity of the heap to judge the drying trend; oxygen sensors: monitor the ventilation status of the heap and whether it is in an aerobic state; gas sensors (NH 3 、CH 4 etc.): used to detect odor substances or anaerobic reaction by-products.

[0043] Arrange multiple points according to the vertical and horizontal distribution of the heap structure to obtain more comprehensive heap parameters; deploy sensors in different fermentation areas (starting section, middle section, ending section) to achieve regional comparison and anomaly identification; all sensors are connected to the edge acquisition gateway or data acquisition controller (DAQ) via wireless (such as LoRa, NB-IoT) or wired methods.

[0044] The multi-source heterogeneous data collected needs to be standardized and then synchronized to the digital twin platform in real time to drive model updates and predictive analysis: unify the format of sensor data, denoise, align time series, and eliminate anomalies; filters, moving average algorithms, and data interpolation techniques can be used to enhance data stability; establish a real-time data caching mechanism to ensure data integrity and traceability.

[0045] Adopt the collaborative method of edge computing and cloud platform to achieve two-way synchronization of data reporting and the twin model; support updating model input variables at second-level or minute-level time granularity; for key indicators (such as sudden temperature changes), a high-frequency synchronization strategy can be triggered to achieve dynamic response.

[0046] For each data update, the model will automatically run simulation calculations to deduce the current internal state of the heap and predict future fermentation trends; if the predicted value deviates from the set target range, it will be used as a decision support input for the control layer; the model is also used to identify abnormal patterns (such as fermentation temperature loss, local anaerobiosis, etc.) to achieve early warning.

[0047] When simulating the actual fermentation state in a digital twin model to predict the fermentation development trend and potential anomalies, a graph neural network can be used. The application advantages are as follows: The heap is regarded as a "spatial graph" composed of multiple points, where the nodes represent sensor points and the edges represent heat conduction or moisture migration paths; it can capture the spatial correlation and mutual influence between different regions of the heap; it has strong modeling ability for spatial heterogeneous features in multi-sensor data.

[0048] The entire fermentation heap is divided into multiple monitoring sub-regions, and each sub-region corresponds to a node (Node) in a graph; the connection relationship between nodes (edges, Edge) is established based on the spatial adjacency relationship, heat, or moisture migration path of the heap; each node is associated with multiple sensor data (such as temperature, humidity, oxygen concentration) to form a feature vector of the node; the initial graph structure G=(V,E,X) of the heap is constructed, where: V is the set of nodes (fermentation sub-regions); E is the set of edges (spatial coupling relationship); X is the node feature matrix (multi-dimensional environmental variables).

[0049] The state of the heap at each moment is represented as a graph Gt, and a sequence of graphs is formed in continuous time; by combining the change of node features and the graph structure invariance in each graph, the input of time-series graph data is constructed; the label data is the target variable (such as temperature, humidity, etc.) or anomaly status mark of each node within a future time window.

[0050] Use a graph convolutional network (GCN) or a graph attention network (GAT) to extract the spatio-temporal features of nodes; the GNN model performs a convolutional operation on the graph Gt at each time step to obtain the embedded representation Ht of the nodes; the representation of each node is not only based on local sensor data but also integrates the features of adjacent nodes, reflecting the spatial coupling relationship inside the heap.

[0051] Input the node representation Ht into the subsequent time series prediction module (such as LSTM, GRU, or Transformer module); predict variables such as temperature / humidity for each node at future time steps (such as t+1 to t+6 hours); combine the actual historical distribution to perform anomaly scoring on the prediction results (such as residual analysis, probability distribution shift judgment); if the predicted value of a certain node deviates from the normal change trend, it is marked as a potential anomaly.

[0052] Feed the prediction results and anomaly recognition results back to the digital twin platform for updating the thermal map, trend line, and status label in the twin heap; trigger control strategies in advance, such as early warning for turning the heap and spraying operations; at the same time, compare the recognition results with the subsequent actual monitoring results to dynamically adjust the GNN model parameters (online learning or retraining) to improve the prediction accuracy.

[0053] There are multiple control actions in the system, such as turning the pile, spraying, ventilation, etc. The priorities of each control action are preset (for example, spraying takes precedence over turning the pile), which are used as judgment weights in conflict detection.

[0054] Extract real-time sensor data closely related to control actions, including the predicted value of the oxygen utilization efficiency inside the pile, which is used to reflect the real-time metabolic ability of microorganisms to oxygen during the fermentation process, is highly related to whether turning the pile is needed, reflects whether the ventilation of the pile is effective, and the estimated value of the drying rate of the pile surface, which represents the speed of water loss on the surface per unit time and is closely related to whether spraying is needed, and can reflect the drying speed rather than just the absolute humidity value.

[0055] Traditional systems mostly use oxygen concentration as the judgment basis, but this value is greatly affected by ventilation and location and is inaccurate; the predicted value of oxygen utilization efficiency can more directly reflect "whether turning the pile is needed" and is a more dynamic indicator reflecting the state of microorganisms.

[0056] Among them, the method for obtaining the predicted value of oxygen utilization efficiency is as follows: construct a causal relationship graph of oxygen utilization, determine the variable nodes related to oxygen utilization efficiency, including the temperature, humidity or moisture content of the fermentation pile, oxygen concentration (inlet air, inside the pile), and ventilation volume; based on the fermentation principle and on-site experience, establish the causal dependence relationship between variables to form a directed acyclic graph (DAG); set the predicted value of oxygen utilization efficiency as the target variable node among them, which is jointly affected by other variables;

[0057] Collect a large amount of real-time monitoring data during the historical fermentation process for training the Bayesian network;

[0058] For each pair of variables with a dependence relationship, establish a conditional probability distribution through statistical methods; if some data are missing or sparse, prior knowledge can be introduced to set the initial probability, and then it can be corrected through Bayesian updating.

[0059] During the fermentation process, collect the observed values of temperature, humidity, oxygen concentration, and ventilation volume in real time; use the observed values as known nodes to input into the Bayesian network; trigger the conditional propagation algorithm in the Bayesian network using the input of the observed nodes; the network automatically updates the posterior probabilities of all relevant nodes; obtain the probability distribution result of the predicted value node of oxygen utilization efficiency as the current predicted value of oxygen utilization efficiency.

[0060] Different from the absolute humidity, the estimated value of the surface drying rate reflects the dynamic trend of water loss and can judge whether it is "continuously drying"; it is the core variable for judging whether spraying is needed, avoiding misjudgment of "low humidity but actually not dry".

[0061] The method for obtaining the estimated value of the surface drying rate is as follows:

[0062] High-sensitivity humidity sensors are deployed on the surface of the fermentation pile to collect data at fixed time intervals (such as every 5 minutes or 10 minutes); each data sequence represents the change trajectory of surface humidity within a certain time window, forming a large number of humidity time series samples;

[0063] The autoencoder is used to perform unsupervised feature learning on the above humidity time series: the encoder part automatically extracts the low-dimensional feature vector that best represents the dryness change trend of each time series; the decoder part attempts to restore the feature vector to the original humidity curve for training stability; after training, the encoder can compress any newly collected humidity change sequence into a feature representation of fixed dimensionality.

[0064] The feature vectors extracted by the autoencoder are clustered using a clustering algorithm (such as K-Means, DBSCAN, or DTW-based hierarchical clustering); each clustering result represents a typical drying mode, such as: fast drying, slow drying, drying stagnation, and rewetting after drying (such as after rain); after clustering, a corresponding drying rate label (such as "high", "medium", "low" or numerical level) is assigned to each type of drying mode.

[0065] During the fermentation process, the system continuously collects the current surface humidity time series; uses the trained autoencoder to compress the features of the current sequence; matches the feature vector with the cluster center to determine which drying mode the current sequence belongs to; and outputs the estimated surface drying rate for the current time period based on the category.

[0066] The predicted value of oxygen utilization efficiency and the estimated surface drying rate are converted into a comprehensive feature vector, which is used as the input of the machine learning model. The feature vector reflects the linkage trend between the two types of control behaviors (such as high OUE + high SDR indicates that compost turning and spraying may be triggered at the same time). OUE is the predicted value of oxygen utilization efficiency, and SDR is the estimated surface drying rate. Combined with the currently set control action priority (such as spraying takes precedence over compost turning), the machine learning model uses each group of comprehensive feature vectors to predict the risk value label of logical conflict in the rule engine as the prediction target, and minimizes the sum of prediction errors of risk value labels with logical conflicts in all rule engines as the training target. The machine learning model is trained until the sum of prediction errors reaches convergence, and the model training is stopped. The risk value of logical conflict in the rule engine is determined based on the model output results. Among them, the machine learning model is a polynomial regression model.

[0067] Compare the risk value of logical conflicts in the obtained rule engine with a predetermined threshold. If the risk value of logical conflicts in the rule engine is greater than or equal to the predetermined threshold, it indicates that there are logical conflicts in the rule engine. At this time, a warning signal is generated and the control strategy is optimized. If the risk value of logical conflicts in the rule engine is less than the predetermined threshold, it indicates that there are no logical conflicts in the rule engine. At this time, no warning signal is generated and no additional adjustment is required.

[0068] Dynamically adjust the execution order of control actions (priority adjustment): According to the actual working conditions reflected by the current OUE and SDR, reallocate the priorities of control actions. For example, if the current OUE is small → increase the priority of turning the pile; if the current SDR is large → increase the priority of spraying; avoid both conflicting actions (such as turning the pile and spraying) being executed when the trigger conditions are simultaneously met. If the weights of the two actions are similar, the system can insert a "delay window" to execute the second action (refer to the "shortest cooling time" model).

[0069] Introduce a "control mutual exclusion time locking mechanism" and set the minimum interval time threshold between actions. For example, spraying is not allowed within 20 minutes after turning the pile; turning the pile is not allowed again within 15 minutes after spraying; avoid frequent start-stop of equipment or abnormal disturbance of the pile body caused by rapid switching; the time locking threshold can be dynamically adjusted based on historical feedback.

[0070] After the system detects the conflict risk, it automatically performs the following threshold optimization: increase the temperature threshold for turning the pile trigger (e.g., from 65°C to 68°C); lower the lower limit of humidity for spraying trigger (e.g., from 30% to 28%); stagger the trigger conditions to reduce the probability of simultaneous satisfaction; the range of optimized values is controlled by the "risk value" intensity, and the higher the risk, the greater the adjustment amplitude.

[0071] Rather than triggering two conflicting actions in parallel, it is better to adopt a "light first, heavy later" progressive control strategy: First, only execute the action with higher priority or more significant risk; by observing the changes in feedback parameters, decide whether to execute the second action. For example, observe whether the SDR decreases after the initial spraying for 5 minutes; if it is still dry, trigger the next stage of turning the pile; essentially, a dynamic feedback closed-loop control logic is introduced.

[0072] The optimized control strategy is generated by the central control module and includes: control action types (turning the pile, spraying, ventilation); execution timing (execute immediately or delay execution); action parameters (such as spraying duration, turning depth, air volume intensity); the control strategy is encoded as a standard control instruction and sent to the execution equipment through a PLC or remote IO module, such as: turning pile robotic arm or turning machine; spraying solenoid valve control system; ventilation fan group.

[0073] The control system monitors the action status of the execution layer in real time to ensure that instructions are accurately responded to. For continuous control actions (such as spraying and ventilation), the system starts duration control or closed-loop feedback regulation according to the set parameters. If the equipment execution is abnormal (such as startup failure, action interruption), the system will trigger an abnormal termination mechanism and record the fault log.

[0074] The system structurally records the information of each control action execution, including but not limited to: action type (turning pile / spraying / ventilation); execution start and end times; action intensity parameters (such as wind speed, water spray volume, turning pile frequency); control reason (such as turning pile triggered by low OUE, spraying triggered by high SDR); decision source (such as model output / manual intervention); The data is stored in the "control behavior log module" for historical backtracking and model learning.

[0075] After the action is executed, the system collects the corresponding environmental feedback data in real time, including: whether the temperature distribution after turning the pile is evenly increased; the change trend of the surface humidity after spraying; whether the oxygen concentration and OUE are improved after ventilation; compare the parameter changes before and after control to quantify the control effect; Write the feedback data and execution data into the "feedback data archiving module" together as the basis for optimizing subsequent control strategies and calibrating the model.

[0076] The system conducts closed-loop regulation analysis based on the feedback data: if the feedback result reaches the expected target, the current strategy is solidified; if the feedback deviation is large, the model is triggered for update, parameter fine-tuning or recommended manual intervention; at the same time, the feedback data is used to train or fine-tune prediction models (such as OUE prediction, drying rate identification, conflict risk assessment model) to improve the accuracy of the next round of strategies.

[0077] This method enables the digital twin model to have the ability of self-tuning to adapt to different fermentation scenarios, material differences and environmental disturbances by constructing a fast adaptation mechanism between control behavior and feedback data, specifically including:

[0078] In the initial stage, a differentiable and transferable digital twin model is constructed to simulate the livestock manure fermentation process (temperature and humidity dynamics, microbial metabolism, etc.); at the same time, a set of data structures for recording are established to associate the corresponding relationships between each control behavior, environmental response and model output; Each record contains: control input (such as spraying duration, turning pile frequency); model predicted value (such as temperature rise rate); actual feedback value (measured by sensor); environmental context variables (such as external air temperature, ventilation mode, etc.).

[0079] Construct multiple small-sample task sets using historical feedback data. Each task represents a specific fermentation sub-scenario, such as: wet pig manure turning; rapid drying in high summer heat; rapid air supply for local hypoxia, etc. Each task contains a training set (for model update) and a validation set (for effect evaluation) to evaluate the generalization and adaptation capabilities of the twin model; simulate the goal of "whether a model can quickly adapt to different fermentation environments".

[0080] Gradually fine-tune the digital twin model in each task; through repeated iterations of multiple tasks, optimize the initial parameters of the model so that it can quickly adapt to new tasks; essentially, instead of training "one optimal model", train "one model that is easy to be fine-tuned".

[0081] During actual operation, after each control action is executed, the collected feedback data is immediately constructed into a new task; input this task into the trained meta-learning framework, and with only a few iterations, the current twin model can be locally fine-tuned; the goal of fine-tuning is: to make the mapping relationship between the model and the current heap body and control response more accurate; retain the fine-tuned model as the current scenario sub-model for short-term prediction and control.

[0082] The system periodically or according to a trigger mechanism verifies the prediction performance of the current twin sub-model (for example, through subsequent real sensor data); if the error drops significantly or meets the self-set threshold, the system will: extract the parameter changes after fine-tuning; feedback to the main model parameter set, update the main model or add it to the empirical model pool; achieve the gradual evolution of the main model and the enhancement of its generalization ability.

[0083] Through continuous feedback, fine-tuning and induction processes, the system gradually constructs a digital twin model library applicable to: different fermentation materials; different climate environments; different heap structures; when formulating future control strategies, the system can quickly call the most matching sub-model from the model pool for high-precision prediction and control strategy deduction.

[0084] The above formulas are all dimensionless and take their numerical calculations. The formula is a formula obtained by collecting a large amount of data for software simulation to get the closest real situation. The preset parameters in the formula are set by technicians in this field according to the actual situation.

[0085] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired or wireless (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

Claims

1. A method for monitoring livestock and poultry manure fermentation based on digital twins, characterized in that: include: Build a digital twin model of the livestock and poultry manure fermentation process, collect real-time data of the fermentation pile through sensors deployed at the fermentation site, and synchronize the real-time data to the digital twin model; Simulate the actual fermentation status in the digital twin model to predict fermentation development trends and potential anomalies; Set up an intelligent control strategy, which combines the mutual influence between real-time data and the execution conditions of the control action priority to determine whether there is a logical conflict in the rule engine. If so, optimize the control strategy, specifically: The control strategy constructs a comprehensive feature vector based on the coupling relationship between the predicted value of oxygen utilization efficiency and the estimated value of surface drying rate, and combines the preset control action priority to predict the risk value of logical conflict in the rule engine through a machine learning model, and compares it with a predetermined threshold to determine whether there is a logical conflict; According to the optimized control strategy, the pile turning, spraying and ventilation operations are automatically performed, and the control behavior and feedback data are recorded in real time. The digital twin model is continuously adjusted based on the recorded feedback data.

2. The method for monitoring livestock and poultry manure fermentation based on digital twin according to claim 1, characterized in that: The real-time data collected by the sensor include temperature, humidity, oxygen concentration, pH value and gas release at different layers of the pile, which are used to construct a multi-dimensional, hierarchical twin model input data set.

3. The method for monitoring livestock and poultry manure fermentation based on digital twin according to claim 1, characterized in that: Graph neural networks are used to predict fermentation trends and potential anomalies, including: The entire fermentation pile is divided into multiple monitoring sub-areas, each of which corresponds to a node in the graph; the connection relationship between nodes is established based on the spatial adjacency relationship of the pile and the heat or moisture migration path; each node is associated with multiple sensor data to form a feature vector of the node; the initial graph structure G=(V,E,X) of the pile is constructed, where: V is the node set; E is the edge set; X is the node feature matrix; The state of the heap at each moment is represented as a graph Gt, and the continuous time forms a graph sequence; combining the node feature changes in each graph with the graph structure invariance, the time series graph data input is constructed; the label data is the target variable or abnormal state mark of each node in a certain time window in the future; Use graph convolutional networks or graph attention networks to extract the spatiotemporal features of nodes; the GNN model performs convolution operations on the graph Gt at each time step to obtain the embedded representation Ht of the node; The node representation Ht is input into the subsequent time series prediction module; the temperature or humidity variables of each node in the future time step are predicted; the prediction results are scored for anomalies based on the actual historical distribution; if the predicted value of a node deviates from the normal change trend, it is marked as a potential anomaly.

4. The method for monitoring livestock and poultry manure fermentation based on digital twin according to claim 3 is characterized in that: The method for obtaining the predicted value of oxygen utilization efficiency is as follows: constructing a causal relationship diagram of oxygen utilization, determining variable nodes related to oxygen utilization efficiency, including fermentation pile temperature, humidity or moisture content, oxygen concentration, and ventilation volume, establishing a causal dependency relationship between variables, forming a directed acyclic graph, and setting the predicted value of oxygen utilization efficiency as the target variable node therein; Collect real-time monitoring data from historical fermentation processes for training Bayesian networks; During the fermentation process, the observed values ​​of temperature, humidity, oxygen concentration and ventilation volume are collected in real time; the observed values ​​are input into the Bayesian network as known nodes; The input of the observation node is used to trigger the conditional propagation algorithm in the Bayesian network; the network automatically updates the posterior probabilities of all relevant nodes; and the probability distribution result of the oxygen utilization efficiency prediction value node is obtained as the current oxygen utilization efficiency prediction value.

5. The method for monitoring livestock and poultry manure fermentation based on digital twin according to claim 4, characterized in that: An estimate of the surface drying rate is obtained as follows: High-sensitivity humidity sensors are deployed on the surface of the fermentation pile to collect data at fixed time intervals; each data sequence represents the change trajectory of surface humidity within a fixed time window, forming a humidity time series sample; an autoencoder is used to perform unsupervised feature learning on the humidity time series: the encoder part automatically extracts low-dimensional feature vectors representing the drying change trend in each time series; the decoder part attempts to restore the feature vector to the original humidity curve; after training, the encoder compresses any newly collected humidity change sequence into a fixed-dimensional feature representation; A clustering algorithm is used to cluster the feature vectors extracted by the autoencoder; each clustering result represents a typical drying mode; after clustering, a corresponding drying rate label is assigned to each type of drying mode; During the fermentation process, the system continuously collects the current surface humidity time series and uses the trained autoencoder to perform feature compression on the current series; Match the feature vector with the cluster center to determine which drying mode the current sequence belongs to; According to the category, the estimated surface drying rate for the current time period is output accordingly.

6. The method for monitoring livestock and poultry manure fermentation based on digital twin according to claim 5, characterized in that: The predicted value of oxygen utilization efficiency and the estimated surface drying rate are converted into a comprehensive feature vector, which is used as the input of the machine learning model. Combined with the currently set control action priority, the machine learning model uses each group of comprehensive feature vectors to predict the risk value label of logical conflict in the rule engine as the prediction target, and takes minimizing the sum of prediction errors of risk value labels of logical conflict in all rule engines as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The risk value of logical conflict in the rule engine is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

7. The method for monitoring livestock and poultry manure fermentation based on digital twin according to claim 6, characterized in that: The obtained risk value of logical conflict in the rule engine is compared with the preset threshold. If the risk value of logical conflict in the rule engine is greater than or equal to the preset threshold, it means that there is a logical conflict in the rule engine. At this time, a warning signal is generated and the control strategy is optimized; if the risk value of logical conflict in the rule engine is less than the preset threshold, it means that there is no logical conflict in the rule engine. At this time, no warning signal is generated and no additional adjustment is required.

8. The method for monitoring livestock and poultry manure fermentation based on digital twin according to claim 1, characterized in that: The digital twin model is continuously adjusted based on the recorded feedback data, specifically including: in the initial stage, a digital twin model is constructed to simulate the fermentation process of livestock and poultry manure, and a set of data structures for recording is established to associate the correspondence between each control behavior, environmental response and model output; multiple small sample task sets are constructed using historical feedback data, each task represents a specific fermentation sub-scenario, and each task contains a training set and a verification set; the digital twin model is gradually fine-tuned in each task; the initial parameters of the model are optimized through repeated iterations of multiple tasks. In actual operation, after each control behavior is executed, the collected feedback data is immediately constructed as a new task; the new task is input into the trained meta-learning framework for iteration, the current twin model is locally fine-tuned, and the fine-tuned model is retained as the current scene sub-model; the system regularly verifies the prediction performance of the current twin sub-model; if the error decreases or meets the self-set threshold, the system will extract the fine-tuned parameter changes, feed them back to the main model parameter set, and update the main model or add it to the experience model pool.

Images