Self-adaptive control method and system for animal husbandry pollutant treatment

Abstract

The invention relates to the technical field of adaptive control, in particular to an adaptive control method and system for animal husbandry pollutant treatment. The method comprises the following steps: based on a historical viscosity sequence, a real-time change rate of feces viscosity, environmental temperature and humidity data and an association relationship between the environmental temperature and humidity data and viscosity change, obtaining viscosity memory effect strength reflecting an influence degree of a fluctuation mode on a current decision and a viscosity trend prediction index used for predicting a future viscosity change trend; dynamically adjusting a gain parameter of an adaptive controller according to the viscosity memory effect strength and the viscosity trend prediction index; and by utilizing the adjusted gain parameter and the current viscosity deviation, generating a preliminary control instruction for cooperatively adjusting the dilution water amount and the stirring intensity through a multi-objective optimization algorithm. According to the method, the system can sense the viscosity rising trend caused by environment change or feed change in advance, the control gain is intelligently adjusted based on the historical fluctuation mode, and the response time of the system is shortened.

Description

Technical Field

[0001] This invention relates to the field of adaptive control technology, and more specifically, to an adaptive control method and system for treating livestock pollutants. Background Technology

[0002] In large-scale livestock farming, the efficient treatment and resource utilization of manure are crucial for ensuring environmental safety and sustainable development. The viscosity of manure is its core rheological property, directly affecting pump efficiency, mixing energy consumption, solid-liquid separation, and the efficiency of subsequent biochemical reactions such as anaerobic digestion. Excessively high viscosity can lead to pipe blockage and equipment overload; conversely, excessively low viscosity often means excessive water consumption, increasing the load on subsequent treatment processes and operating costs.

[0003] Currently, the industry's control of manure viscosity has the following shortcomings:

[0004] The perception and decision-making dimensions are too narrow. Most current methods rely solely on instantaneous viscosity measurements for feedback control (such as PID), completely ignoring the two key dimensions of historical behavior pattern memory of fecal viscosity and time-varying characteristics coupled with the environment. This results in a lack of predictability in control decisions, which can only respond passively after viscosity anomalies occur, causing regulation lag and overshoot oscillation.

[0005] The rigidity of control objectives and parameters in existing solutions, which typically employ fixed gain parameters and a single viscosity control objective, makes it impossible to dynamically balance multiple objectives such as "rapid viscosity stabilization," "saving dilution water," and "reducing operating energy consumption." Furthermore, these solutions cannot adaptively adjust the aggressiveness of the control strategy based on changes in operating conditions (such as changes in feed formulation or seasonal shifts). This results in three major drawbacks for existing systems in the dynamic and complex real-world livestock farming environment: low control precision, significant resource waste, and poor adaptability. Consequently, they fail to meet the urgent needs of modern livestock farming for refined and intelligent processing.

[0006] Therefore, an adaptive control method and system for treating livestock pollutants are provided. Summary of the Invention

[0007] The purpose of this invention is to provide an adaptive control method and system for treating livestock pollutants, so as to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention aims to provide an adaptive control method for treating livestock pollutants, comprising the following steps:

[0009] S1. Multi-dimensional state information, including real-time fecal viscosity, parameters characterizing the source of fecal matter, environmental temperature and humidity data, and historical viscosity sequences, is collected synchronously through a sensor array.

[0010] S2. Based on the historical viscosity sequence, dynamically adjust the viscosity fluctuation pattern within the time window to obtain the viscosity memory effect intensity that reflects the degree of influence of the fluctuation pattern on the current decision.

[0011] S3. Based on the real-time viscosity change rate of the fecal waste, the environmental temperature and humidity data and their correlation with viscosity change, obtain a viscosity trend prediction index for predicting future viscosity change trends.

[0012] S4. Dynamically adjust the gain parameters of the adaptive controller based on the viscosity memory effect intensity and the viscosity trend prediction index;

[0013] S5. Using the adjusted gain parameter and the current viscosity deviation, a preliminary control command for coordinating the adjustment of dilution water volume and stirring intensity is generated through a multi-objective optimization algorithm. The preliminary control command is then dynamically corrected based on the real-time viscosity change rate, ultimately generating a coordinating control command for the dilution water volume and stirring intensity.

[0014] As a further improvement to this technical solution, in S1, the parameters characterizing the source characteristics of feces include: feed composition characteristic data obtained through optical analysis, pre-stored viscosity reference coefficients corresponding to different animal species, and real-time detected fecal pH values.

[0015] As a further improvement to this technical solution, the specific process for obtaining the viscosity memory effect intensity, which reflects the degree of influence of the fluctuation mode on the current decision, in step S2 is as follows:

[0016] S21. Set the initial length of the historical data backtracking time window. The historical viscosity sequence within the time window is segmented, the viscosity change rate of each time period is calculated, and a time decay weight is assigned to each historical data point.

[0017] S22. Based on the absolute value of the viscosity change rate in each time period and its corresponding time decay weight, a preliminary memory effect intensity is generated. ;

[0018] S23, regarding the intensity of the preliminary memory effect. After normalization, the final viscosity memory effect intensity is obtained. .

[0019] As a further improvement to this technical solution, in step S21, the length of the historical data backtracking time window is dynamically adjusted based on an adaptive time window management mechanism, specifically as follows:

[0020] S211. Real-time calculation of the current viscosity change rate ;

[0021] S212, The current viscosity change rate The fluctuation is compared with a preset fluctuation threshold, which includes a low fluctuation threshold. and high volatility threshold ;

[0022] S213, if If the time window is too long, it is determined to be a period of fluctuation, and the historical data is traced back to the specified time window length. Shorten to the first preset value ;

[0023] S214, if If the time window is determined to be stable, then the historical data will be traced back to the specified length. Extended to the second preset value ;

[0024] S215, if If so, the current time window length remains unchanged.

[0025] As a further improvement to this technical solution, the specific steps in S3 for obtaining the viscosity trend prediction index used to predict future viscosity change trends are as follows:

[0026] S31, Based on the current viscosity change rate Get the current viscosity change acceleration ;

[0027] S32. Based on the ambient temperature and humidity data, calculate the deviation of the data relative to the reference ambient conditions;

[0028] S33. Based on the environmental deviation and the pre-trained environmental-viscosity coupling relationship, calculate the environmental coupling factor. ;

[0029] S34, the rate of change , change acceleration and environmental coupling factors The input is fed into a pre-trained, attention-based trend prediction model to output the future. Viscosity prediction sequence for each control cycle ;

[0030] S35. Standardize the trend prediction value into the viscosity trend prediction index. .

[0031] As a further improvement to this technical solution, the specific steps for dynamically adjusting the gain parameters of the adaptive controller in step S4 are as follows:

[0032] S41. Predefine the grade classification range of the viscosity memory effect intensity and the viscosity trend prediction index;

[0033] S42. Establish a gain adjustment decision table with the viscosity memory effect intensity level and the viscosity trend prediction index level as inputs;

[0034] Among them, the viscosity memory effect intensity is divided into the first memory effect level, the second memory effect level, and the third memory effect level; the viscosity trend prediction index is divided into the first trend level, the second trend level, and the third trend level.

[0035] S43. Viscosity memory effect intensity calculated in real time With viscosity trend prediction index Based on the assigned level, query the gain adjustment decision table to determine the current gain mode to be adopted. The gain modes include a first gain mode, a second gain mode, and a third gain mode, specifically:

[0036] If the viscosity memory effect intensity level is the first memory effect level and the viscosity trend prediction index level is the first trend level, then the first gain mode is adopted;

[0037] If the viscosity memory effect intensity level is the third memory effect level and the viscosity trend prediction index level is the third trend level, then the third gain mode is adopted;

[0038] Otherwise, the second gain mode is used.

[0039] S44. The preset base gain value according to the determined gain adjustment mode. Calculate the gain parameter for the current control cycle. .

[0040] When the first gain mode is used , The first gain coefficient;

[0041] When the second gain mode is used , This is the second gain coefficient;

[0042] When the third gain mode is used , This is the third gain coefficient.

[0043] As a further improvement to this technical solution, in step S5, the specific steps for generating preliminary control commands to coordinately adjust the dilution water volume and stirring intensity using a multi-objective optimization algorithm based on the adjusted gain parameter and the current viscosity deviation are as follows:

[0044] S51. Construct a multi-objective optimization problem, the objective of which is to minimize the cumulative viscosity deviation, cumulative dilution water consumption, and cumulative energy consumption of system equipment in the prediction time domain.

[0045] S52. An evolutionary algorithm based on Pareto sort is used to solve the optimization problem online to obtain a set of Pareto optimal solutions;

[0046] S53. Based on the preset weight coefficients used to characterize the priority of each optimization objective, select the optimal solution from the Pareto optimal solution set and output the corresponding preliminary set value of dilution water flow rate. Initial set value of mixer speed .

[0047] As a further improvement to this technical solution, the constraints of the multi-objective optimization problem include those based on the gain parameter. The flow and power constraints are as follows: the change in dilution water flow rate between adjacent control cycles shall not exceed... The change in mixer speed between adjacent control cycles does not exceed ,in This represents the maximum allowable flow rate of the dilution water. This represents the maximum permissible speed of the mixer.

[0048] As a further improvement to this technical solution, in step S5, the initial control command is dynamically corrected based on the real-time viscosity change rate, ultimately generating a coordinated control command for the dilution water volume and stirring intensity. The specific method is as follows:

[0049] Obtain the absolute value of the current viscosity change rate;

[0050] If the absolute value of the rate of change is greater than the first rate threshold Then the suppression mode is triggered:

[0051] Multiply the dilution water flow rate setpoint in the initial control command by the inhibition coefficient. The corrected dilution water flow rate is obtained and the execution command is executed, while the set value of the mixer speed remains unchanged.

[0052] If the absolute value of the rate of change is less than the second rate threshold If so, the acceleration mode is triggered:

[0053] Multiply the dilution water flow rate setpoint by the acceleration factor. The mixer speed setting value is increased proportionally to obtain the corrected collaborative control command;

[0054] If the absolute value of the rate of change is within the first rate threshold With the second rate threshold In between, the preliminary control command serves as the final coordinated control command to be executed.

[0055] On the other hand, the present invention provides an adaptive control system for livestock pollutant treatment, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the adaptive control method for livestock pollutant treatment described in any of the above-mentioned embodiments.

[0056] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0057] In this adaptive control method and system for livestock pollutant treatment, the introduction of a dual sensing dimension—the intensity of historical memory effect and the environmental coupling trend prediction index—enables the system to anticipate viscosity increases caused by environmental changes or feed modifications, and intelligently adjust the control gain based on historical fluctuation patterns. This not only transforms the system from a passive response to an active prediction system, but also significantly shortens the response time in the face of sudden operating conditions, and substantially improves the anticipation and accuracy of the control.

[0058] This adaptive control method and system for livestock pollutant treatment utilizes an embedded multi-objective optimization algorithm to automatically optimize each control decision, achieving an optimal dynamic balance between the three core objectives of viscosity control accuracy, water conservation, and reduced equipment energy consumption. This mechanism can widely adapt to complex operating conditions caused by different animal species, feed formulations, and seasonal climates, fundamentally solving the common problem of traditional systems struggling to balance control performance and operational economy, and exhibiting poor adaptability. It provides an innovative solution for achieving comprehensive intelligent and resource-efficient livestock pollutant treatment. Attached Figure Description

[0059] Figure 1 This is a flowchart of the overall method of the present invention. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0061] Example 1: Please refer to Figure 1 As shown, this embodiment provides an adaptive control method for treating livestock pollutants, including the following steps:

[0062] S1. Multi-dimensional state information, including real-time fecal viscosity, parameters characterizing the source of fecal matter, environmental temperature and humidity data, and historical viscosity sequences, is collected synchronously through a sensor array.

[0063] In S1, the parameters characterizing the source of feces include: feed composition characteristic data obtained through optical analysis, pre-stored viscosity reference coefficients corresponding to different animal species, and real-time detected pH values ​​of feces.

[0064] The specific methods for obtaining it are as follows:

[0065] Thirty minutes after daily feeding, an automatic sampler was used to collect 50g of feed samples from the feed trough. The feed samples were then optically scanned in the wavelength range of 400-2500nm using a near-infrared spectrometer to obtain spectral characteristic data reflecting the cellulose, protein and starch content in the feed.

[0066] Cellulose content: Calculated using the absorption peak at 1700nm wavelength, it is 58.3% (standard value: 55-62%).

[0067] Protein content: 18.7% (standard value: 16-20%), calculated from the absorption peak at 1650nm wavelength.

[0068] Starch content: Calculated from the absorption peak at 1450nm wavelength, it is 27.5% (standard value: 25-30%).

[0069] Based on the animal species information provided by the farm, the corresponding animal species coefficients are retrieved from the pre-stored database. The viscosity reference coefficient for pigs ranges from 1.0 to 1.3, for cattle from 0.7 to 0.9, and for poultry from 1.4 to 1.6. In this scheme, all animal species obtained are pigs, and the coefficient is determined to be 1.15 based on the average weight of the pig herd on that day (150 kg) and the breed (Landrace pig).

[0070] A pH sensor is installed at the outlet of the sewage collection tank. It automatically detects the pH once per minute. The daily detection value is 6.8 (normal range: 6.5-7.5). The system automatically calculates the average pH value of the past 10 minutes as 6.75, which is used as the current pH value input.

[0071] Spectral eigenvectors: (Cellulose, protein, and starch content);

[0072] Animal species coefficient: 1.15;

[0073] pH value: 6.75;

[0074] Final feature vector: ;

[0075] The final feature vector is input into a pre-trained manure characteristic classification model (based on the random forest algorithm, with training data from 10,000 samples from 100 farms). The system outputs a viscosity characteristic label of "medium viscosity - stable type", with a corresponding viscosity benchmark value of 22.5 mPa·s.

[0076] Furthermore, the viscosity of fecal waste was collected based on multi-mode viscosity fusion measurement, the specific process of which included:

[0077] S11: Obtain the first viscosity measurement value using an ultrasonic viscosity sensor. The ultrasonic viscosity sensor operates at a frequency of 1-10MHz.

[0078] S12: Obtain the second viscosity measurement value via a rotary viscosity sensor. The shear rate of the rotary viscosity sensor is set to 10-100 s. -1 ;

[0079] S13: Obtain the estimated solids content of the current sewage using a solids content sensor. ;

[0080] S14: Select the fusion strategy based on the estimated solids content S:

[0081] like (Range 5%-8%), then use the ultrasonic viscosity measurement value. Main, weight The set range is 0.7-0.9, with the rotational viscosity measurement value as an auxiliary factor and weighted accordingly. The set range is 0.1-0.3;

[0082] like (If the range is 8%-12%), then equal-weighted fusion will be used. ;

[0083] like Then the rotational viscosity measurement value is used. Main, weight The set range is 0.1-0.3, with ultrasonic viscosity measurement as the auxiliary value and weight. The set range is 0.7-0.9;

[0084] S15: Calculate the viscosity value after fusion. , which is the real-time viscosity of the feces used by the system.

[0085] S2. Based on historical viscosity sequences, dynamically adjust the viscosity fluctuation pattern within a time window to obtain the viscosity memory effect intensity, which reflects the degree of influence of the fluctuation pattern on the current decision.

[0086] In S2, the specific process for obtaining the intensity of the viscosity memory effect, which reflects the degree of influence of this fluctuation pattern on the current decision, is as follows:

[0087] S21. Set the initial length of the historical data backtracking time window. The historical viscosity sequence within the time window is segmented, the viscosity change rate for each time period is calculated, and a time decay weight is assigned to each historical data point. The initial length... It lasts for 120 minutes;

[0088] S22. Based on the absolute value of the viscosity change rate in each time period and its corresponding time decay weight, a preliminary memory effect intensity is generated. ;

[0089]

[0090]

[0091] In the formula, The number of data points within the time window; For the first Viscosity values ​​over a specific time period; For the first The time for each data point; For the first Time decay weight for each time period; For the first The absolute value of the viscosity change rate over a time period; This is the attenuation factor, with a value ranging from 0.01 to 0.1. ; ; Let be the viscosity change value during the i-th time period. , For the first Viscosity values ​​for each data point; Let i be the time interval corresponding to the i-th time period. , For the first The time for each data point;

[0092] S23, Intensity of the initial memory effect After normalization, the viscosity memory effect intensity is mapped to the [0,1] interval to obtain the final viscosity memory effect intensity. .

[0093] In S21, the length of the historical data backtracking time window is dynamically adjusted based on an adaptive time window management mechanism, and the viscosity volatility is recalculated every 15 minutes. Specifically:

[0094] S211. Real-time calculation of the current viscosity change rate ;

[0095] S212, Change the current viscosity rate The fluctuation is compared with a preset fluctuation threshold, which includes a low fluctuation threshold. (0.15 mPa·s / min) and high fluctuation threshold (0.65 mPa·s / min);

[0096] S213, if If the time window is too long, it is determined to be a period of fluctuation, and the historical data is traced back to the specified time window length. Shorten to the first preset value , For 30-60 minutes;

[0097] S214, if If the time window is determined to be stable, then the historical data will be traced back to the specified length. Extended to the second preset value , It lasts 180-300 minutes, and ;

[0098] S215, if If so, the current time window length remains unchanged.

[0099] S3. Based on the real-time viscosity change rate of fecal waste, environmental temperature and humidity data, and their correlation with viscosity change, obtain a viscosity trend prediction index for predicting future viscosity change trends.

[0100] In S3, the specific steps for obtaining the viscosity trend prediction index used to predict future viscosity changes are as follows:

[0101] S31, Based on the current viscosity change rate Get the current viscosity change acceleration ;

[0102] S32. Calculate the deviation of the environmental temperature and humidity data relative to the reference environmental conditions.

[0103]

[0104]

[0105] In the formula, The preset reference ambient temperature; This is the difference between the current ambient temperature and the reference temperature. This represents the difference between the current ambient humidity and the reference humidity. The preset baseline relative humidity;

[0106] S33. Calculate the environmental coupling factor based on the environmental deviation and the pre-trained environmental-viscosity coupling relationship. ;

[0107]

[0108] In the formula, The weighting of the effect of unit temperature deviation on the viscosity change trend; The weights represent the influence of unit humidity deviation on the viscosity change trend; both of these influence weights are environmental-viscosity coupling coefficients.

[0109] S34, Rate of Change , change acceleration and environmental coupling factors The input is fed into a pre-trained, attention-based trend prediction model, which is a neural network containing three Transformer encoder layers, each with four attention heads. The model is trained using an Adam optimizer by minimizing the mean squared error between the predicted and actual viscosity. The output is used to predict the future viscosity. Viscosity prediction sequence for each control cycle ;

[0110] S35. Standardize the trend forecast value into a viscosity trend forecast index. ,in A higher value indicates a stronger upward trend in viscosity.

[0111] S4. Dynamically adjust the gain parameters of the adaptive controller based on the viscosity memory effect intensity and viscosity trend prediction index;

[0112] In S4, the specific steps for dynamically adjusting the gain parameters of the adaptive controller are as follows:

[0113] S41. Predefined range of grades for viscosity memory effect intensity and viscosity trend prediction index;

[0114] S42. Establish a gain adjustment decision table with viscosity memory effect intensity level and viscosity trend prediction index level as inputs;

[0115] Among them, the viscosity memory effect intensity is divided into the first memory effect level, the second memory effect level, and the third memory effect level; the viscosity trend prediction index is divided into the first trend level, the second trend level, and the third trend level.

[0116] Specifically, the first level of memory effect (weak): Second level of memory effect (medium): Third level of memory effect (strong): ;

[0117] First trend level (stable): Second trend level (rising): Third trend level (rapid rise): .

[0118] S43. Viscosity memory effect intensity calculated in real time With viscosity trend prediction index Based on the assigned level, query the gain adjustment decision table to determine the appropriate gain mode. Gain modes include the first gain mode, the second gain mode, and the third gain mode, specifically:

[0119] If the viscosity memory effect intensity level is the first memory effect level and the viscosity trend prediction index level is the first trend level, then the first gain mode is adopted.

[0120] If the viscosity memory effect intensity level is the third memory effect level and the viscosity trend prediction index level is the third trend level, then the third gain mode is adopted;

[0121] Otherwise, the second gain mode is used.

[0122] S44. The preset base gain value according to the determined gain adjustment mode. Calculate the gain parameter for the current control cycle. .

[0123] When using the first gain mode , This is the first gain coefficient, with a value ranging from 0.5 to 0.8;

[0124] When using the second gain mode , This is the second gain coefficient, with a value ranging from 0.8 to 1.2;

[0125] When using the third gain mode , This is the third gain coefficient, with a value ranging from 1.2 to 2.0.

[0126] The above method integrates the historical viscosity memory effect intensity with a viscosity trend prediction index coupled with the environment to construct an intelligent decision-making logic that fully covers nine levels of combinations, achieving forward-looking and context-adaptive control gain. When the system identifies the highest-risk condition where "strong memory effect" and "rapid upward trend" overlap, it can immediately switch to high-gain mode for strong intervention, reducing response time by more than 50% and effectively curbing viscosity runaway. For other medium-risk combinations, a medium-gain mode is intelligently adopted to balance response speed and stability. Only when the system is in the ideal state of "weak memory and stable trend" is a low-gain mode used to save energy and reduce consumption.

[0127] S5. Using the adjusted gain parameters and the current viscosity deviation, a preliminary control command for the coordinated adjustment of dilution water volume and stirring intensity is generated through a multi-objective optimization algorithm. The preliminary control command is then dynamically corrected based on the real-time viscosity change rate, and finally, a coordinated control command for the dilution water volume and stirring intensity is generated.

[0128] In S5, the specific steps for generating preliminary control commands to coordinately adjust the dilution water volume and stirring intensity using a multi-objective optimization algorithm based on the adjusted gain parameters and the current viscosity deviation are as follows:

[0129] S51. Construct a multi-objective optimization problem; the objective is to minimize the cumulative viscosity deviation, cumulative dilution water consumption, and cumulative energy consumption of system equipment in the prediction time domain, and the constraints include viscosity constraints, flow constraints, and power constraints.

[0130] During implementation, Within each control cycle, establish the objective function: ;

[0131] ;

[0132] ;

[0133] ;

[0134] In the formula, The objective function is one; The objective function is two; The objective function is three; As a discount factor (0 < γ ≤ 1), more emphasis is placed on recent bias; For at any time Predicted future The viscosity value of the step; To control the time domain, ; For at any time The Future of Decision Making The dilution water flow rate of the step; This refers to the pump energy consumption coefficient. The energy consumption coefficient of the mixer; For at any time The Future of Decision Making The speed of the mixer in step; Set the target value for viscosity; The time interval is used to control the cycle; the cubic term approximately reflects the relationship between stirring power consumption and rotation speed.

[0135] Constraints for multi-objective optimization problems include those based on gain parameters. The flow and power constraints are as follows: the change in dilution water flow rate between adjacent control cycles shall not exceed... The change in mixer speed between adjacent control cycles does not exceed ,in This represents the maximum allowable flow rate of the dilution water. This represents the maximum permissible speed of the mixer.

[0136] The constraints are:

[0137] ;

[0138] ;

[0139] ;

[0140] In the formula, This is the minimum allowable viscosity. This represents the maximum allowable viscosity. This is the minimum allowable flow rate of the dilution water; This represents the maximum allowable flow rate of the dilution water. This is the minimum allowable speed of the mixer; This represents the maximum permissible speed of the mixer.

[0141] S52. An evolutionary algorithm based on Pareto sort is used to solve the optimization problem online to obtain a set of Pareto optimal solutions;

[0142] The algorithm parameters were set as follows: population size of 120, crossover probability of 0.85, mutation probability of 0.1, and maximum number of iterations of 150 generations. The selection operator was set using binary tournament selection.

[0143] S53. Based on the preset weighting coefficients used to characterize the priority of each optimization objective, select the optimal solution from the Pareto optimal solution set and output the corresponding preliminary set value of dilution water flow rate. Initial set value of mixer speed .

[0144] In this embodiment, the preset weights are used. .

[0145] Let the Pareto front solution set be For each objective, calculate the minimum value in the frontier solution. and maximum value For each solution Calculate the normalized target value :

[0146] ;

[0147] Calculate the weighted composite score for each solution. .

[0148] Select the solution with the lowest score .

[0149] Take the first control action of the solution as the initial instruction: , ).

[0150] In S5, the initial control command is dynamically corrected based on the real-time viscosity change rate, ultimately generating a coordinated control command for the dilution water volume and stirring intensity. The specific method is as follows:

[0151] Obtain the absolute value of the current viscosity change rate;

[0152] If the absolute value of the rate of change is greater than the first rate threshold If the range is 0.6-1.0 mPa·s / min, then the inhibition mode is triggered.

[0153] Multiply the dilution water flow rate setpoint in the initial control command by the inhibition coefficient. The value ranges from 0.5 to 0.7, and the corrected dilution water flow rate is obtained by executing the command, while keeping the mixer speed setting unchanged.

[0154] If the absolute value of the rate of change is less than the second rate threshold If the range is 0.1-0.2 mPa·s / min, then the acceleration mode is triggered.

[0155] Multiply the dilution water flow rate setpoint by the acceleration factor. The value ranges from 1.2 to 1.5, and the set value of the mixer speed is increased proportionally by 10% to 20% to obtain the corrected collaborative control command.

[0156] If the absolute value of the rate of change is within the first rate threshold With the second rate threshold In between, the preliminary control command serves as the final coordinated control command to be executed.

[0157] In this embodiment, the first rate threshold Second rate threshold Suppression coefficient acceleration coefficient Speed ​​increase ratio .

[0158] The initial instructions obtained from multi-objective optimization are: , .

[0159] If the calculation is Triggering suppression mode:

[0160] ;

[0161] ;

[0162] If the calculation is Trigger acceleration mode

[0163] ;

[0164] ;

[0165] If the calculation is If so, then no correction is needed.

[0166] ;

[0167] ;

[0168] Ultimately, Send to the variable frequency dilution pump and variable frequency agitator for execution.

[0169] Example 2: This example provides an adaptive control system for livestock pollutant treatment, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the adaptive control method for livestock pollutant treatment described in any of the above examples.

[0170] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. An adaptive control method for treating livestock pollutants, characterized in that, Includes the following steps: S1. Multi-dimensional state information, including real-time fecal viscosity, parameters characterizing the source of fecal matter, environmental temperature and humidity data, and historical viscosity sequences, is collected synchronously through a sensor array. S2. Based on the historical viscosity sequence, dynamically adjust the viscosity fluctuation pattern within the time window to obtain the viscosity memory effect intensity that reflects the degree of influence of the fluctuation pattern on the current decision. S3. Based on the real-time viscosity change rate of the fecal waste, the environmental temperature and humidity data and their correlation with viscosity change, obtain a viscosity trend prediction index for predicting future viscosity change trends. S4. Dynamically adjust the gain parameters of the adaptive controller based on the viscosity memory effect intensity and the viscosity trend prediction index; S5. Using the adjusted gain parameter and the current viscosity deviation, a preliminary control command for coordinating the adjustment of dilution water volume and stirring intensity is generated through a multi-objective optimization algorithm. The preliminary control command is then dynamically corrected based on the real-time viscosity change rate, ultimately generating a coordinating control command for the dilution water volume and stirring intensity.

2. The adaptive control method for livestock pollutant treatment according to claim 1, characterized in that: In S1, the parameters characterizing the source of feces include: feed composition characteristic data obtained through optical analysis, pre-stored viscosity reference coefficients corresponding to different animal species, and real-time detected pH value of feces.

3. The adaptive control method for livestock pollutant treatment according to claim 2, characterized in that: In step S2, the specific process for obtaining the intensity of the viscosity memory effect, which reflects the degree of influence of the fluctuation pattern on the current decision, is as follows: S21. Set the initial length of the historical data backtracking time window. The historical viscosity sequence within the time window is segmented, the viscosity change rate of each time period is calculated, and a time decay weight is assigned to each historical data point. S22. Based on the absolute value of the viscosity change rate in each time period and its corresponding time decay weight, a preliminary memory effect intensity is generated. ; S23, regarding the intensity of the preliminary memory effect After normalization, the final viscosity memory effect intensity is obtained. .

4. The adaptive control method for livestock pollutant treatment according to claim 3, characterized in that: In step S21, the length of the historical data backtracking time window is dynamically adjusted based on an adaptive time window management mechanism, specifically as follows: S211. Real-time calculation of the current viscosity change rate ; S212, The current viscosity change rate The fluctuation is compared with a preset fluctuation threshold, which includes a low fluctuation threshold. and high volatility threshold ; S213, if If the time window is too long, it is determined to be a period of fluctuation, and the historical data is traced back to the specified time window length. Shorten to the first preset value ; S214, if If the time window is determined to be stable, then the historical data will be traced back to the specified length. Extended to the second preset value ; S215, if If so, the current time window length remains unchanged.

5. The adaptive control method for livestock pollutant treatment according to claim 4, characterized in that: In step S3, the specific steps for obtaining the viscosity trend prediction index used to predict future viscosity changes are as follows: S31, Based on the current viscosity change rate Get the current viscosity change acceleration ; S32. Based on the ambient temperature and humidity data, calculate the deviation of the data relative to the reference ambient conditions; S33. Based on the environmental deviation and the pre-trained environmental-viscosity coupling relationship, calculate the environmental coupling factor. ; S34, the rate of change , change acceleration and environmental coupling factors The input is fed into a pre-trained, attention-based trend prediction model to output the future. Viscosity prediction sequence for each control cycle ; S35. Standardize the trend prediction value into the viscosity trend prediction index. .

6. The adaptive control method for livestock pollutant treatment according to claim 5, characterized in that: In step S4, the specific steps for dynamically adjusting the gain parameters of the adaptive controller are as follows: S41. Predefine the grade classification range of the viscosity memory effect intensity and the viscosity trend prediction index; S42. Establish a gain adjustment decision table with the viscosity memory effect intensity level and the viscosity trend prediction index level as inputs; Among them, the viscosity memory effect intensity is divided into the first memory effect level, the second memory effect level, and the third memory effect level; the viscosity trend prediction index is divided into the first trend level, the second trend level, and the third trend level. S43. Viscosity memory effect intensity calculated in real time With viscosity trend prediction index Based on the assigned level, query the gain adjustment decision table to determine the current gain mode to be adopted. The gain modes include a first gain mode, a second gain mode, and a third gain mode, specifically: If the viscosity memory effect intensity level is the first memory effect level and the viscosity trend prediction index level is the first trend level, then the first gain mode is adopted; If the viscosity memory effect intensity level is the third memory effect level and the viscosity trend prediction index level is the third trend level, then the third gain mode is adopted; Otherwise, the second gain mode is used. S44. The preset base gain value according to the determined gain adjustment mode. Calculate the gain parameter for the current control cycle. .

7. The adaptive control method for livestock pollutant treatment according to claim 6, characterized in that: In step S5, the specific steps for generating preliminary control commands to coordinately adjust the dilution water volume and stirring intensity using a multi-objective optimization algorithm based on the adjusted gain parameter and the current viscosity deviation are as follows: S51. Construct a multi-objective optimization problem, the objective of which is to minimize the cumulative viscosity deviation, cumulative dilution water consumption, and cumulative energy consumption of system equipment in the prediction time domain. S52. An evolutionary algorithm based on Pareto sort is used to solve the optimization problem online to obtain a set of Pareto optimal solutions; S53. Based on the preset weight coefficients used to characterize the priority of each optimization objective, select the optimal solution from the Pareto optimal solution set and output the corresponding preliminary set value of dilution water flow rate. Initial set value of mixer speed .

8. The adaptive control method for livestock pollutant treatment according to claim 7, characterized in that: The constraints of the multi-objective optimization problem include those based on the gain parameter. The flow and power constraints are as follows: the change in dilution water flow rate between adjacent control cycles shall not exceed... The change in mixer speed between adjacent control cycles does not exceed ,in This represents the maximum allowable flow rate of the dilution water. This represents the maximum permissible speed of the mixer.

9. The adaptive control method for livestock pollutant treatment according to claim 8, characterized in that: In step S5, the initial control command is dynamically corrected based on the real-time viscosity change rate, ultimately generating a coordinated control command for the dilution water volume and stirring intensity. The specific method is as follows: Obtain the absolute value of the current viscosity change rate; If the absolute value of the rate of change is greater than the first rate threshold Then the suppression mode is triggered: Multiply the dilution water flow rate setpoint in the initial control command by the inhibition coefficient. The corrected dilution water flow rate is obtained and the execution command is executed, while the set value of the mixer speed remains unchanged. If the absolute value of the rate of change is less than the second rate threshold If so, the acceleration mode is triggered: Multiply the dilution water flow rate setpoint by the acceleration factor. The mixer speed setting value is increased proportionally to obtain the corrected collaborative control command; If the absolute value of the rate of change is within the first rate threshold With the second rate threshold In between, the initial control command serves as the final coordinated control command to be executed.

10. An adaptive control system for treating livestock pollutants, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes a computer program to implement the adaptive control method for livestock pollutant treatment as described in any one of claims 1-9.

Images