Resource integration management methods and systems for the co-treatment of livestock and poultry manure

Abstract

This invention relates to the field of agricultural waste resource utilization and smart environmental protection technology, specifically a resource integration management method and system for the co-treatment of livestock and poultry manure. The method includes: initializing and establishing a breeding characteristic fingerprint mapping model and a biochemical potential time-varying model; collecting business behavior data from breeding nodes; generating pollution-generating characteristic vectors; acquiring environmental temperature data and transportation time data; calculating real-time biochemical potential energy; acquiring current process status data at the treatment end; inputting the real-time biochemical potential energy and current process status data into a preset virtual reaction equilibrium simulation logic to calculate the disturbance residual; acquiring physical transportation costs; constructing a total cost objective function by combining biochemical penalty costs and physical transportation costs; optimizing the process by minimizing the total cost objective function to generate resource scheduling instructions. This invention eliminates the data timeliness lag and distortion under the traditional static scheduling mode, ensuring the real-time and accurate biochemical potential energy data of materials upon arrival at the treatment end.

Description

Technical Field

[0001] This invention relates to the field of agricultural waste resource utilization and smart environmental protection technology, specifically to a resource integration management method and system for the co-treatment of livestock and poultry manure. Background Technology

[0002] In the current resource integration scenario of co-processing of livestock and poultry manure, there are complex logistics and biochemical interactions between the dispersed breeding nodes and the centralized processing terminals. The source data mostly comes from user reports or basic IoT devices, and only includes business behavior data such as the scale of stock and the manure cleaning cycle. Moreover, the manure continues to undergo biochemical reactions after it is generated and during transportation.

[0003] For the scheduling and management of such resources, existing solutions generally adopt a linear scheduling mode based on static physical quantities, that is, point-to-point transportation matching based solely on weight or volume. Due to the lack of direct detection methods for the biochemical components at the source and the failure to quantify the natural fermentation attenuation effect of manure under different environmental temperatures and transportation durations, the biochemical potential data of materials arriving at the treatment end suffers from time lag and distortion. This data blind spot makes it impossible for scheduling instructions to predict the biochemical impact of newly added materials on the current anaerobic digestion system at the treatment end, which can easily lead to acid-base imbalance in the reactor or reduced microbial activity, resulting in frequent process failures and a surge in operation and maintenance costs. Therefore, how to accurately deduce biochemical characteristics through indirect business data, correct energy evolution deviations during transportation in real time, and achieve dynamic adaptation and global optimization scheduling based on the carrying capacity of the treatment end has become an urgent technical problem to be solved. Summary of the Invention

[0004] The purpose of this invention is to provide a resource integration management method and system for the co-treatment of livestock and poultry manure, solving the problem of resource scheduling mismatch caused by the lack of source biochemical data and energy decay during transportation, avoiding biochemical shocks to the treatment process system due to blind scheduling, and making it easier to achieve coordinated optimization of global cost and system steady state based on the treatment end's carrying capacity. Specifically, the technical solution of this invention is as follows:

[0005] Resource integration management methods for the co-treatment of livestock and poultry manure include:

[0006] Initialize and establish a breeding characteristic fingerprint mapping model and a time-varying model of biochemical potential;

[0007] Step 1: Collect business behavior data of aquaculture nodes; use the aquaculture feature fingerprint mapping model to perform weighted mapping processing on the business behavior data to generate pollution feature vectors.

[0008] Step 2: Obtain ambient temperature data and transportation time data; combine the ambient temperature data and transportation time data, use the biochemical potential time-varying model to evolve and correct the pollution-generating feature vector, and calculate the real-time biochemical potential.

[0009] Step 3: Obtain the current process status data of the processing end; input the real-time biochemical potential energy and the current process status data into the preset virtual reaction equilibrium simulation logic to calculate the disturbance residual;

[0010] Step 4: Calculate the biochemical penalty cost based on the perturbation residual; obtain the physical transportation cost; construct the total cost objective function by combining the biochemical penalty cost and the physical transportation cost; perform optimization with minimizing the total cost objective function as a constraint to generate resource scheduling instructions.

[0011] Optional business behavior data includes: stock size data, feed composition index data, average age of livestock data, and manure removal cycle data;

[0012] The pollution-generating characteristic vector includes: initial carbon-to-nitrogen ratio, initial moisture content, and initial biodegradation potential.

[0013] Optionally, the aquaculture feature fingerprint mapping model includes: a preset feature mapping weight matrix and a preset bias vector;

[0014] Using aquaculture feature fingerprint mapping model, business behavior data is weighted and mapped to generate pollution-generating feature vectors, including:

[0015] By using the feature mapping weight matrix, the business behavior data is linearly weighted to obtain the weighted intermediate quantity;

[0016] The weighted intermediate values ​​are calibrated using the bias vector to generate a pollution-generating feature vector.

[0017] Optionally, the pollution-generating feature vector is evolved and corrected using a time-varying biochemical potential model to obtain the real-time biochemical potential, including:

[0018] The absolute process temperature is calculated by superimposing the ambient temperature data with the preset compost self-heating temperature rise value.

[0019] The reaction rate constant is calculated based on the absolute process temperature and the pre-defined Arrhenius equation.

[0020] By using reaction rate constant and transport time data, the initial biodegradation potential value in the pollution generation feature vector is calculated by integral decay to generate real-time biochemical potential energy.

[0021] Optionally, the virtual reaction equilibrium simulation logic is constructed based on the principle of component conservation;

[0022] The calculated disturbance residuals include:

[0023] Analyze the current process status data to obtain the current volume and current component concentration of the reaction system;

[0024] Analyze real-time biochemical potential energy to obtain the mass of materials to be added;

[0025] Calculate the concentration of the key components after mixing, based on the current volume, current component concentration, and mass of the material to be added.

[0026] Calculate the combined biochemical indicators after mixing based on the concentrations of key components;

[0027] The perturbation residuals were calculated based on the concentrations of key components and complex biochemical indicators.

[0028] Optionally, based on the concentrations of key components and complex biochemical indicators, the perturbation residuals are calculated, including:

[0029] Obtain the preset optimal carbon-to-nitrogen ratio threshold and the preset target biochemical potential threshold;

[0030] Calculate the first deviation of the composite biochemical index from the optimal carbon-nitrogen ratio threshold;

[0031] Calculate the second deviation of the energy density corresponding to the concentration of the key components after mixing from the target biochemical potential threshold.

[0032] The first deviation and the second deviation are weighted and summed to generate the disturbance residual.

[0033] Optional, the total cost objective function includes: physical transportation cost and biochemical penalty cost;

[0034] The calculation steps for the biochemical penalty cost item include:

[0035] Obtain the preset unit risk disposal base price;

[0036] Risk factors are calculated based on perturbation residuals and a pre-defined nonlinear risk normalization function.

[0037] Using the unit risk disposal base price and risk factors, a biochemical penalty cost item is generated.

[0038] Optionally, a closed-loop feedback step may also be included:

[0039] The actual output data from the acquisition and processing end;

[0040] Calculate the prediction error between the actual output data and the system's predicted value;

[0041] Determine whether the prediction error is within the preset safety error range;

[0042] In response to the prediction error being outside the safe error range, the feature mapping weight matrix in the aquaculture feature fingerprint mapping model is corrected using the gradient descent method, and the pre-exponential factor parameters in the biochemical potential time-varying model are also corrected.

[0043] In response to the prediction error being within the safe error range, the feature mapping weight matrix and pre-exponential factor parameters remain unchanged.

[0044] A resource integration management system for the co-treatment of livestock and poultry manure includes:

[0045] The source mapping module is used to collect business behavior data of aquaculture nodes and generate pollution feature vectors using the aquaculture feature fingerprint mapping model.

[0046] The dynamic evolution module is used to acquire ambient temperature data and transportation time data, and to calculate real-time biochemical potential energy using a time-varying biochemical potential model.

[0047] The simulation matching module is used to acquire the current process state data of the processing end and input the real-time biochemical potential energy and the current process state data into the virtual reaction equilibrium simulation logic to determine the disturbance residual.

[0048] The optimization scheduling module is used to construct the total cost objective function based on the disturbance residual, and generate resource scheduling instructions with minimizing the total cost objective function as a constraint.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] 1. This invention establishes a breeding feature fingerprint mapping model to transform easily obtainable business behavior data such as stock size and feed composition into pollution production feature vectors. This effectively solves the data gap problem caused by the lack of direct biochemical detection methods at the breeding end, and realizes low-cost and high-precision derivation from indirect business indicators to key biochemical component data, laying a digital foundation for subsequent precise scheduling.

[0051] 2. This invention utilizes a time-varying biochemical potential model, combined with ambient temperature and transportation time, to dynamically evolve and correct the pollution-generating feature vector. This mechanism accurately quantifies the energy decay effect caused by natural fermentation during transportation of manure, eliminates the data timeliness lag and distortion under the traditional static scheduling mode, and ensures the real-time and accuracy of the biochemical potential data of the material when it arrives at the treatment end.

[0052] 3. This invention applies virtual reaction equilibrium simulation logic to couple and analyze real-time biochemical potential energy with the current process state at the treatment end, and calculates disturbance residuals; by simulating the mixing state after the introduction of exogenous materials, it predicts the degree of biochemical impact of new materials on the existing anaerobic digestion system in advance, effectively avoiding process failures such as acid-base imbalance or reduced bacterial activity, and ensuring the stable operation of the treatment system.

[0053] 4. This invention constructs a total cost objective function that includes physical transportation costs and biological and chemical penalty costs, and generates scheduling instructions with minimizing the total cost as a constraint; by introducing a nonlinear risk normalization function, potential process fluctuation risks are transformed into quantifiable economic costs, realizing a global trade-off between logistics costs and process safety costs, improving resource utilization efficiency and reducing overall operation and maintenance expenditures. Attached Figure Description

[0054] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0055] Figure 1 This is a flowchart of the method of the present invention;

[0056] Figure 2 This is a structural diagram of the system of the present invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0058] Example 1:

[0059] Please see Figure 1 Resource integration management methods for the co-treatment of livestock and poultry manure include:

[0060] Initialize and establish a breeding characteristic fingerprint mapping model and a time-varying model of biochemical potential;

[0061] Step 1: Collect business behavior data of aquaculture nodes; use the aquaculture feature fingerprint mapping model to perform weighted mapping processing on the business behavior data to generate pollution feature vectors.

[0062] Step 2: Obtain ambient temperature data and transportation time data; combine the ambient temperature data and transportation time data, use the biochemical potential time-varying model to evolve and correct the pollution-generating feature vector, and calculate the real-time biochemical potential.

[0063] Step 3: Obtain the current process status data of the processing end; input the real-time biochemical potential energy and the current process status data into the preset virtual reaction equilibrium simulation logic to calculate the disturbance residual;

[0064] Step 4: Calculate the biochemical penalty cost based on the perturbation residual; obtain the physical transportation cost; construct the total cost objective function by combining the biochemical penalty cost and the physical transportation cost; perform optimization with minimizing the total cost objective function as a constraint to generate resource scheduling instructions.

[0065] This embodiment relates to a resource integration and management method for the co-treatment of livestock and poultry manure. This method solves the problem of resource scheduling mismatch caused by data loss at the breeding end and biochemical degradation during transportation through digital means. At the beginning of system operation, a breeding characteristic fingerprint mapping model and a biochemical potential time-varying model are initialized and established in the computing platform.

[0066] Business behavior data of aquaculture nodes is collected, which comes from user reporting terminals or IoT basic collection devices; using a preset aquaculture feature fingerprint mapping model, the collected business behavior data is weighted and mapped to derive and generate pollution feature vectors; this process uses data regression technology to transform easily obtainable business indicators into biochemical component data that are difficult to measure directly;

[0067] Ambient temperature data and transportation time data are acquired through vehicle positioning and sensing devices. Combining these two dynamic variables, the biochemical potential time-varying model is used to evolve and correct the pollution characteristic vector, and the real-time biochemical potential energy is calculated. This process aims to quantify the energy decay of sewage due to natural fermentation during transportation and ensure the timeliness of the data.

[0068] Acquire current process status data from the processing end, including reactor liquid level, pH, and microbial activity indicators; input real-time biochemical potential energy and current process status data into a preset virtual reaction equilibrium simulation logic; this logic calculates the concentration of key components after mixing by simulating the mixing state after the introduction of exogenous materials, based on the law of conservation of mass. The calculation formula is as follows:

[0069]

[0070] in, For the quality of the raw materials to be used, The density of the raw material; Characterizing the volume of raw materials ; For the first reaction system Current concentrations of key components; This represents the current volume of the reaction system. The first batch of raw materials to be put into use Concentration of key components;

[0071] Furthermore, based on the concentration of key components after mixing Using a preset component calorific value conversion coefficient vector Calculate the energy density per unit mass of the mixture. The specific calculation logic is as follows:

[0072]

[0073] in, For the first The theoretical calorific value per unit mass of key components, such as total organic carbon, is expressed in units of... ; The average density of the mixed system is approximately taken as [value missing]. This step aims to map the volumetric concentration data to the biochemical potential energy data in the mass-energy dimension, in order to align with the target threshold in terms of dimensionality; based on the mixed composite biochemical index, i.e., the mixed carbon-nitrogen ratio. and the energy density calculated above ,in, This represents the total organic carbon concentration after mixing. The total nitrogen concentration after mixing is compared with the steady-state target of the system to calculate the disturbance residual, so as to quantify the impact of the new material on the existing process system.

[0074] Based on the perturbation residual, the cost of biochemical punishment is calculated using a nonlinear risk normalization function to characterize the potential cost of process risk management; physical transportation costs are obtained based on road network geographic information; a total cost objective function is constructed by combining the cost of biochemical punishment and the physical transportation cost, and global optimization is performed with minimizing this function as a constraint to generate resource scheduling instructions pointing to specific processing nodes.

[0075] Example 2:

[0076] Business behavior data includes: stock size data, feed composition index data, average age of livestock data, and manure removal cycle data;

[0077] The pollution-generating characteristic vector includes: initial carbon-to-nitrogen ratio, initial moisture content, and initial biodegradation potential.

[0078] Business behavior data The input variables for the model consist of the following dimensions: Shelf size data. The current livestock inventory of the farm, derived from breeding records, is used to estimate the total pollution production baseline; feed composition index data. The percentage of protein or fiber content in feed, derived from a feed supply chain database, determines the basic abundance of nitrogen and phosphorus in manure; average age of livestock data. These are values ​​representing the livestock's growth stage, used to correct for the degree of manure digestibility; manure removal cycle data. The time interval between two manure cleaning operations is derived from the equipment operation log and is used to determine the initial anaerobic state of the manure before it leaves the plant.

[0079] Pollution-generating feature vector The output variables of the model include the following standardized parameters: initial carbon-to-nitrogen ratio. A dimensionless ratio, it is a key parameter determining the type of subsequent fermentation process; initial moisture content value. These are percentage values, affecting transportation efficiency and thermal balance at the processing end; initial biodegradability potential values. The energy density that can theoretically be generated per unit mass of sewage, in units of... .

[0080] Example 3:

[0081] Aquaculture feature fingerprint mapping model, including: a preset feature mapping weight matrix and a preset bias vector;

[0082] Using aquaculture feature fingerprint mapping model, business behavior data is weighted and mapped to generate pollution-generating feature vectors, including:

[0083] By using the feature mapping weight matrix, the business behavior data is linearly weighted to obtain the weighted intermediate quantity;

[0084] The weighted intermediate values ​​are calibrated using the bias vector to generate a pollution-generating feature vector.

[0085] The aquaculture feature fingerprint mapping model includes a pre-defined feature mapping weight matrix. and the preset bias vector This model is based on the principle of affine transformation in linear algebra.

[0086] Generate pollution feature vectors The process is as follows: using the feature mapping weight matrix Business behavior data vector Perform linear weighted calculation to obtain the weighted intermediate quantity. ;matrix The dimensions are designed as the ratio of output units to input units to eliminate the difference in physical dimensions of the input data; a bias vector is used. The weighted intermediate values ​​are calibrated to generate a pollution-generating feature vector. Bias vector Used to introduce regional baseline corrections, such as eliminating overall activity bias in specific climatic regions.

[0087] Example 4:

[0088] The evolutionary correction of the pollution-generating feature vector is performed using a time-varying biochemical potential model, and the real-time biochemical potential is calculated, including:

[0089] The absolute process temperature is calculated by superimposing the ambient temperature data with the preset compost self-heating temperature rise value.

[0090] The reaction rate constant is calculated based on the absolute process temperature and the pre-defined Arrhenius equation.

[0091] By using reaction rate constant and transport time data, the initial biodegradation potential value in the pollution generation feature vector is calculated by integral decay to generate real-time biochemical potential energy.

[0092] The time-varying model of biochemical potential is constructed based on the Arrhenius equation of chemical kinetics; environmental temperature data is used. Superimposed pre-set compost self-heating temperature rise The temperature was then converted to the Kelvin scale to calculate the absolute process temperature. Self-heating temperature rise of compost Depending on the material's stacking volume, it is usually set to to empirical constants;

[0093] Based on absolute process temperature Calculate the reaction rate constant using the pre-defined reaction kinetic equation. ;in Pre-exponential factor, The ideal gas constant; activation energy parameter The calibration was achieved by fitting a pre-collected laboratory isothermal decay dataset. During this calibration process, different constant laboratory temperatures were recorded. The reaction rate at the following levels Determined through regression analysis The value;

[0094] Using the reaction rate constant And transportation time data, for the initial biodegradation potential value in the pollution-generating feature vector Perform integral decay calculations to generate real-time biochemical potential energy. .

[0095] Example 5:

[0096] The virtual reaction equilibrium simulation logic is constructed based on the principle of component conservation.

[0097] The calculated disturbance residuals include:

[0098] Analyze the current process status data to obtain the current volume and current component concentration of the reaction system;

[0099] Analyze real-time biochemical potential energy to obtain the mass of materials to be added;

[0100] Calculate the concentration of the key components after mixing, based on the current volume, current component concentration, and mass of the material to be added.

[0101] Calculate the combined biochemical indicators after mixing based on the concentrations of key components;

[0102] The perturbation residuals were calculated based on the concentrations of key components and complex biochemical indicators.

[0103] The virtual reaction equilibrium simulation logic is built based on the principle of component conservation; it analyzes the current process state data to obtain the current volume of the reaction system. and current concentration of key components Key components include total organic carbon and total nitrogen; real-time biochemical potential energy is analyzed to obtain the mass of materials to be input. and its component concentration Due to the pollution-generating feature vector Includes only standardized parameters and component concentrations. Specifically, this includes total organic carbon concentration. and total nitrogen concentration It needs to be obtained by solving a system of simultaneous equations in reverse; the specific solution steps are as follows:

[0104] By definition, the initial carbon-to-nitrogen ratio in the pollution-generating characteristic vector And the real-time biochemical potential energy obtained from step 2 The following physical constraints must be met:

[0105]

[0106] in, and These are the preset theoretical heat production values ​​per unit mass of total organic carbon and total nitrogen, respectively; by solving the above system of equations, the specific component concentrations of the materials to be input can be obtained:

[0107]

[0108]

[0109] Calculated and As Substitute into subsequent mixed calculations;

[0110] Based on the law of conservation of mass, calculate the concentration of the key components after mixing. The calculation formula is: ,in The density of the raw materials is derived from the combined results of weighbridge data and volumetric measurement equipment. Based on the concentrations of key components after mixing, the composite biochemical index of the mixture is calculated, specifically the carbon-to-nitrogen ratio. Based on the concentrations of key components and complex biochemical indicators, the perturbation residuals are calculated by comparing them with the steady-state targets of the system.

[0111] Example 6:

[0112] Based on the concentrations of key components and complex biochemical indicators, the perturbation residuals were calculated, including:

[0113] Obtain the preset optimal carbon-to-nitrogen ratio threshold and the preset target biochemical potential threshold;

[0114] Calculate the first deviation of the composite biochemical index from the optimal carbon-nitrogen ratio threshold;

[0115] Calculate the second deviation of the energy density corresponding to the concentration of the key components after mixing from the target biochemical potential threshold.

[0116] The first deviation and the second deviation are weighted and summed to generate the disturbance residual.

[0117] Disturbance residual The calculation employs a multi-objective weighted bias method; the preset optimal carbon-to-nitrogen ratio threshold is obtained from the process database. and the preset target biochemical potential threshold ; Set as to It is determined based on the optimal nutrient range for microbial growth and metabolism; Set as the minimum energy density required to maintain thermal equilibrium in the reactor;

[0118] Calculation of composite biochemical indicators The first deviation is obtained from the normalized variance relative to the optimal carbon-to-nitrogen ratio threshold. ,in, The optimal carbon-to-nitrogen ratio threshold is preset; the energy density corresponding to the concentration of the key components after mixing is calculated. The second deviation is obtained from the normalized variance relative to the target biochemical potential threshold. The first and second deviations are weighted and summed to generate the disturbance residual. ,in, and These are the weighting coefficients; the weighting coefficients satisfy the normalization constraint conditions. In this embodiment, given the crucial role of carbon-nitrogen ratio balance in the microbial initiation phase, the following is set: , This value was obtained through statistical analysis of historical process failure data using the Analytic Hierarchy Process (AHP).

[0119] Example 7:

[0120] The total cost objective function includes: physical transportation cost and biological penalty cost;

[0121] The calculation steps for the biochemical penalty cost item include:

[0122] Obtain the preset unit risk disposal base price;

[0123] Risk factors are calculated based on perturbation residuals and a pre-defined nonlinear risk normalization function.

[0124] Using the unit risk disposal base price and risk factors, a biochemical penalty cost item is generated.

[0125] Total cost objective function It includes physical transportation costs and biological / chemical penalty costs, and the formula is as follows: Among them, physical transportation costs The specific calculation formula is as follows:

[0126]

[0127] in, This represents the number of road segments included in the scheduling path. The first [item] calculated using Dijkstra's algorithm based on road network geographic information GIS. The shortest distance of each section of the road, in km; The unit price is the preset unit distance transportation price, in yuan / km; Estimated loading and unloading waiting time, in hours (h). The unit price for the time cost of a vehicle shift is expressed in yuan / hour.

[0128] The calculation of the biochemical penalty cost item is as follows: Obtain the preset unit risk disposal base price. The unit of this value is yuan per ton, which represents the average reagent dosage and operation and maintenance cost required at the processing end to maintain steady state when the system experiences a unit degree of biochemical fluctuation.

[0129] Based on perturbation residual Calculate the risk factor using a pre-defined nonlinear risk normalization function. The calculation formula is:

[0130]

[0131] in, This is a dimensionless system tolerance limit coefficient, which is expressed as the maximum acid-base tolerance deviation of the reaction system, such as the maximum allowable tolerance. The fluctuation range is normalized to the allowable deviation ratio relative to the optimal steady state. The specific normalization calculation formula is as follows:

[0132]

[0133] in, This is the critical pH threshold before the reaction system goes rancid and collapses; it is typically set to 6.8 or 8.5. The optimal steady-state pH for the process is typically set to 7.2; this coefficient... Used to define the maximum self-regulation boundary of a system without human intervention; The penalty index is set to a constant greater than 1 to reflect the non-linear growth of risk;

[0134] Using unit risk disposal base price, risk factors and unloading quality at key points Generate a biochemical penalty cost item; considering The unit is yuan per ton, and the dimensions need to be consistent during calculations:

[0135]

[0136] If the data collected on-site is the unloading volume Then through Convert to quality data, unit: The constants in the formula are involved in the calculation. Used to convert kilograms to tons, to ensure The final result is in standard currency units; the specific steps for global optimization with the constraint of minimizing the total cost objective function are as follows: establish a set of candidate processing nodes. ; Traverse each candidate node in the set The physical transportation cost is calculated based on the node's geographical location, and the biochemical penalty cost is calculated based on the node's current process status data. The summation yields the total cost of the node. Compare the total cost values ​​of all candidate nodes and select... The node with the smallest value is selected as the target node, and a resource scheduling instruction containing the target node ID and a suggested transportation path is generated.

[0137] Example 8:

[0138] This method also includes a closed-loop feedback step:

[0139] The actual output data from the acquisition and processing end;

[0140] Calculate the prediction error between the actual output data and the system's predicted value;

[0141] Determine whether the prediction error is within the preset safety error range;

[0142] In response to the prediction error being outside the safe error range, the feature mapping weight matrix in the aquaculture feature fingerprint mapping model is corrected using the gradient descent method, and the pre-exponential factor parameters in the biochemical potential time-varying model are also corrected.

[0143] In response to the prediction error being within the safe error range, the feature mapping weight matrix and pre-exponential factor parameters remain unchanged.

[0144] The closed-loop feedback step aims to calibrate model parameters; it also collects actual output data from the processing end. Such as daily biogas production; the predicted value calculated by the system based on the model. The method for obtaining this information is as follows: Obtain the total mass of the mixture fed into the reaction system in this scheduling task. ,Right now and the energy density per unit mass of the mixture calculated in Example 1. Using formulas Calculate the theoretical gas production, where The preset energy-to-gas conversion coefficient, in units of This coefficient is derived from gas production calibration data under ideal laboratory conditions; the actual output data is compared with the predicted values ​​calculated by the system based on the model. Prediction error between ;

[0145] Determine the prediction error Is it within the preset safety error range? Within; this range is set based on the sensor measurement accuracy and the allowable fluctuation range of the process; in response to the prediction error being outside the safe error range, the feature mapping weight matrix in the aquaculture feature fingerprint mapping model is corrected in reverse using the gradient descent method. And the pre-exponential factor parameters in the time-varying model of biochemical potential. The corrected formula follows ,in, These are the model parameters before correction. For the learning rate, the partial derivatives The calculation is based on the chain rule:

[0146] Pre-exponential factors in time-varying models of biochemical potential The calculation path is For the feature mapping weight matrix in the aquaculture feature fingerprint mapping model The calculation path is Through the above chain-like differentiation, the output error of the terminal is allocated to the transportation stage model and the source mapping model for synchronous updating; in response to the prediction error being within the safe error range, the feature mapping weight matrix and the pre-exponential factor parameters remain unchanged.

[0147] Example 9:

[0148] Please see Figure 2 A resource integration management system for the co-treatment of livestock and poultry manure includes:

[0149] The source mapping module is used to collect business behavior data of aquaculture nodes and generate pollution feature vectors using the aquaculture feature fingerprint mapping model.

[0150] The dynamic evolution module is used to acquire ambient temperature data and transportation time data, and to calculate real-time biochemical potential energy using a time-varying biochemical potential model.

[0151] The simulation matching module is used to acquire the current process state data of the processing end and input the real-time biochemical potential energy and the current process state data into the virtual reaction equilibrium simulation logic to determine the disturbance residual.

[0152] The optimization scheduling module is used to construct the total cost objective function based on the disturbance residual, and generate resource scheduling instructions with minimizing the total cost objective function as a constraint.

[0153] The resource integration management system for the co-treatment of livestock and poultry manure includes the following functional modules:

[0154] The source mapping module is configured to collect business behavior data from aquaculture nodes and generate pollution feature vectors using an aquaculture feature fingerprint mapping model.

[0155] The dynamic evolution module is configured to acquire ambient temperature data and transportation time data, and use the biochemical potential time-varying model to calculate the real-time biochemical potential energy.

[0156] The simulation matching module is configured to acquire the current process state data of the processing end and input the real-time biochemical potential energy and the current process state data into the virtual reaction equilibrium simulation logic to determine the disturbance residual.

[0157] The optimization scheduling module is configured to construct a total cost objective function based on the disturbance residual, and generate resource scheduling instructions with minimizing the total cost objective function as a constraint.

[0158] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A resource integration management method for the co-treatment of livestock and poultry manure, characterized in that, include: Initialize and establish a breeding characteristic fingerprint mapping model and a time-varying model of biochemical potential; Step 1: Collect business behavior data of aquaculture nodes; use the aquaculture feature fingerprint mapping model to perform weighted mapping processing on the business behavior data to generate pollution feature vectors. Step 2: Obtain ambient temperature data and transportation time data; By combining ambient temperature data and transportation time data, the biochemical potential time-varying model is used to evolve and correct the pollution-generating characteristic vector, and the real-time biochemical potential is calculated. Step 3: Obtain the current process status data of the processing end; The real-time biochemical potential energy and current process state data are input into the preset virtual reaction equilibrium simulation logic to calculate the disturbance residual. Step 4: Calculate the biochemical penalty cost based on the perturbation residual; obtain the physical transportation cost; A total cost objective function is constructed by combining the costs of biological and chemical penalties and physical transportation; optimization is performed by minimizing the total cost objective function to generate resource scheduling instructions.

2. The resource integration and management method for the co-treatment of livestock and poultry manure according to claim 1, characterized in that, Business behavior data includes: stock size data, feed composition index data, average age of livestock data, and manure removal cycle data; The pollution-generating characteristic vector includes: initial carbon-to-nitrogen ratio, initial moisture content, and initial biodegradation potential.

3. The resource integration and management method for the co-treatment of livestock and poultry manure according to claim 2, characterized in that, Aquaculture feature fingerprint mapping model, including: a preset feature mapping weight matrix and a preset bias vector; Using aquaculture feature fingerprint mapping model, business behavior data is weighted and mapped to generate pollution-generating feature vectors, including: By using the feature mapping weight matrix, the business behavior data is linearly weighted to obtain the weighted intermediate quantity; The weighted intermediate values ​​are calibrated using the bias vector to generate a pollution-generating feature vector.

4. The resource integration and management method for the co-treatment of livestock and poultry manure according to claim 2, characterized in that, The evolutionary correction of the pollution-generating feature vector is performed using a time-varying biochemical potential model, and the real-time biochemical potential is calculated, including: The absolute process temperature is calculated by superimposing the ambient temperature data with the preset compost self-heating temperature rise value. The reaction rate constant is calculated based on the absolute process temperature and the pre-defined Arrhenius equation. By using reaction rate constant and transport time data, the initial biodegradation potential value in the pollution generation feature vector is calculated by integral decay to generate real-time biochemical potential energy.

5. The resource integration and management method for the co-treatment of livestock and poultry manure according to claim 1, characterized in that, The virtual reaction equilibrium simulation logic is constructed based on the principle of component conservation. The calculated disturbance residuals include: Analyze the current process status data to obtain the current volume and current component concentration of the reaction system; Analyze real-time biochemical potential energy to obtain the mass of materials to be added; Calculate the concentration of the key components after mixing, based on the current volume, current component concentration, and mass of the material to be added. Calculate the combined biochemical indicators after mixing based on the concentrations of key components; The perturbation residuals were calculated based on the concentrations of key components and complex biochemical indicators.

6. The resource integration and management method for the co-treatment of livestock and poultry manure according to claim 5, characterized in that, Based on the concentrations of key components and complex biochemical indicators, the perturbation residuals were calculated, including: Obtain the preset optimal carbon-to-nitrogen ratio threshold and the preset target biochemical potential threshold; Calculate the first deviation of the composite biochemical index from the optimal carbon-nitrogen ratio threshold; Calculate the second deviation of the energy density corresponding to the concentration of the key components after mixing from the target biochemical potential threshold. The first deviation and the second deviation are weighted and summed to generate the disturbance residual.

7. The resource integration and management method for the co-treatment of livestock and poultry manure according to claim 1, characterized in that, The total cost objective function includes: physical transportation cost and biological penalty cost; The calculation steps for the biochemical penalty cost item include: Obtain the preset unit risk disposal base price; Risk factors are calculated based on perturbation residuals and a pre-defined nonlinear risk normalization function. Using the unit risk disposal base price and risk factors, a biochemical penalty cost item is generated.

8. The resource integration and management method for the co-treatment of livestock and poultry manure according to claim 4, characterized in that, It also includes a closed-loop feedback step: The actual output data of the acquisition and processing terminal; Calculate the prediction error between the actual output data and the system's predicted value; Determine whether the prediction error is within the preset safety error range; In response to the prediction error being outside the safe error range, the feature mapping weight matrix in the aquaculture feature fingerprint mapping model is corrected using the gradient descent method, and the pre-exponential factor parameters in the biochemical potential time-varying model are also corrected. In response to the prediction error being within the safe error range, the feature mapping weight matrix and pre-exponential factor parameters remain unchanged.

9. A resource integration management system for the co-treatment of livestock and poultry manure, applied to the resource integration management method for the co-treatment of livestock and poultry manure as described in any one of claims 1-8, characterized in that, include: The source mapping module is used to collect business behavior data of aquaculture nodes and generate pollution feature vectors using the aquaculture feature fingerprint mapping model. The dynamic evolution module is used to acquire ambient temperature data and transportation time data, and to calculate real-time biochemical potential energy using a time-varying biochemical potential model. The simulation matching module is used to acquire the current process state data of the processing end and input the real-time biochemical potential energy and the current process state data into the virtual reaction equilibrium simulation logic to determine the disturbance residual. The optimization scheduling module is used to construct the total cost objective function based on the disturbance residual, and generate resource scheduling instructions with minimizing the total cost objective function as a constraint.

Images