A pollutant emission prediction and management system for livestock farms

Abstract

The present application relates to the technical field of pollution control of livestock and poultry breeding, in particular to a pollution emission prediction and management system for livestock and poultry farms, which comprises a data acquisition and storage unit for acquiring multi-source data of the farm in real time; an influencing factor calculation unit for obtaining key influencing factors of water quality change and malodorous gas based on production data and facility operation data of the farm; a model prediction and analysis unit for generating a predicted value of pollution emission by using a dynamic multi-dimensional pollution emission prediction model and correcting in the prediction process by introducing the key influencing factors; and a management decision unit for determining an early warning level and generating a control instruction according to the early warning level. The system realizes accurate prediction of malodorous gas and water pollution emission by constructing a dynamic multi-dimensional pollution emission prediction model, and at the same time, the system not only prevents over-standard pollution emission in advance, but also deeply integrates with breeding production management to guide actual breeding operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of livestock and poultry breeding pollution control, in particular to a pollution emission prediction and management system for livestock and poultry farms. BACKGROUND

[0002] With the rapid development of large-scale and intensive livestock and poultry breeding industry, a large amount of malodorous gas and high-concentration organic wastewater generated in the breeding process has become an important source of agricultural non-point source pollution, which has caused continuous and serious pressure on the surrounding atmosphere, water body and soil environment. Precise control and emission reduction of breeding pollutants is a key challenge to promote the green and sustainable development of livestock and poultry industry.

[0003] However, the development of existing pollution emission management system for livestock and poultry farms is obviously lagging behind the expansion of production scale, which mainly has the following defects:

[0004] First, the prediction model parameters are updated slowly, which cannot respond to the dynamic changes of the farm. When the daily changes such as increasing the discharge amount, changing the feed formula, and increasing the age of livestock and poultry occur, the system still uses the inherent coefficient, resulting in a large deviation between the prediction results and the actual emission situation, and the early warning function is basically invalid.

[0005] Second, the management measures are single and disjointed, often only focusing on end-of-pipe treatment (such as increasing the addition of sewage treatment chemicals after exceeding the standard), while ignoring more economic and effective source control (such as optimizing the feed ratio) and process management (such as timely manure cleaning).

[0006] Therefore, a pollution emission prediction and management system for livestock and poultry farms is designed. SUMMARY

[0007] The present application aims to provide a pollution emission prediction and management system for livestock and poultry farms to solve the problems raised in the background art.

[0008] To achieve the above-mentioned purpose, the present application aims to provide a pollution emission prediction and management system for livestock and poultry farms, comprising:

[0009] A data acquisition and storage unit is used to acquire multi-source data of the farm in real time; the multi-source data includes production data, pollutant concentration data and facility operation data;

[0010] An influencing factor calculation unit is used to obtain key influencing factors affecting water quality changes and malodorous gas based on the production data and facility operation data of the farm;

[0011] A model prediction analysis unit is configured to generate a predicted value of pollutant emission by using a dynamic multi-dimensional pollutant emission prediction model based on real-time multi-source data, and introduce key influencing factors for correction in the prediction process.

[0012] A management decision unit is configured to obtain a ratio of a predicted emission value to an emission standard based on the predicted value of the malodorous gas concentration and the predicted value of the water pollutant concentration, determine an early warning level according to the ratio, and generate a control instruction according to the early warning level.

[0013] As a further improvement of the technical solution, the data acquisition and storage unit comprises a production data acquisition module, a pollution concentration monitoring module, an operation data acquisition module, and a data storage module.

[0014] The production data acquisition module is configured to obtain data representing the scale of livestock and poultry breeding and the nutrition supply of the farm.

[0015] The scale of livestock and poultry breeding data includes livestock and poultry species, inventory, current average age, and effective area of livestock and poultry house; and the nutrition supply data includes feed formula and crude protein content in the feed formula.

[0016] The pollution concentration monitoring module is configured to monitor real-time water index concentration data reflecting the degree of water pollution and real-time gas index concentration data reflecting the degree of gas pollution.

[0017] The operation data acquisition module is configured to acquire key operation parameters of the water treatment facility.

[0018] The key operation data of the facility include operation data of aerators, sludge return pumps, and fans, and dissolved oxygen content of the aeration tank.

[0019] The data storage module is configured to store real-time multi-source data and historical multi-source data in the form of a time series database, and also store baseline parameters preset in the dynamic multi-dimensional pollutant emission prediction model, including baseline fecal pollutant generation rate and baseline maximum specific degradation rate of sewage.

[0020] As a further improvement of the technical solution, the influencing factor calculation unit comprises a gas factor influencing module and a water body factor influencing module.

[0021] The gas factor influencing module is configured to obtain livestock density influencing factors and feed composition influencing factors based on production data of the farm.

[0022] The water body factor influencing module is configured to obtain process efficiency influencing factors based on key operation data of the facility.

[0023] As a further improvement of the technical solution, in the gas factor influencing module, the specific method for obtaining livestock density influencing factors is as follows:

[0024] S21, obtaining a breed conversion coefficient from a preset breed coefficient table according to the livestock and poultry breed, and converting the inventory into a standard livestock and poultry unit;

[0025] S22, calculating the number of livestock and poultry per unit effective area to obtain a basic breeding density;

[0026] S23, calculating a daily age weight factor according to the current average daily age of the livestock and poultry and the physiological mature daily age of the breed;

[0027] S24, obtaining a breeding density influence factor based on the basic breeding density and the daily age weight factor.

[0028] As a further improvement of the technical solution, the specific method for obtaining the feed composition influence factor in the gas factor influence module is as follows:

[0029] S25, based on the data of nutrient supply, calculating the ratio of the crude protein content of the feed on the day to the standard crude protein reference content of the current livestock and poultry growth stage, to obtain a crude protein influence base value;

[0030] S26, obtaining the ratio of the actual content of the key limiting amino acid in the feed formula on the day to the standard demand, and calculating an amino acid balance degree coefficient;

[0031] S27, based on the crude protein influence base value and the amino acid balance degree coefficient, obtaining the feed composition influence factor.

[0032] As a further improvement of the technical solution, the implementation method for obtaining the process efficiency influence factor in the water body factor influence module is as follows:

[0033] Based on the key operation data of the facility, the actual operation energy consumption, oxygen transfer efficiency and sludge reflux efficiency index are obtained, and the process efficiency influence factor is constructed according to the actual operation energy consumption, oxygen transfer efficiency and sludge reflux efficiency index.

[0034] As a further improvement of the technical solution, the model prediction analysis unit includes a first prediction branch module and a second prediction branch module;

[0035] The first prediction branch module is used to obtain an odor gas concentration prediction value based on the reference fecal pollutant generation rate, combined with the breeding density influence factor, the feed composition influence factor and the real-time gas index concentration data, using a long short-term memory network;

[0036] The second prediction branch module is used to obtain a water body pollutant concentration prediction value based on the reference sewage maximum specific degradation rate, combined with the process efficiency influence factor and the real-time water body index concentration data, using a water quality dynamics model.

[0037] As a further improvement of the technical solution, the specific steps of the first prediction branch module obtaining the odor gas concentration prediction value are as follows:

[0038] S311, based on the breeding density influence factor and the feed composition influence factor, the reference fecal pollution production rate is corrected by weighted deviation coupling to obtain the corrected fecal pollution production rate ;

[0039] S312, the corrected fecal pollution production rate , combined with the current inventory, the odor gas production intensity is obtained ;

[0040] S313, the odor gas production intensity, the real-time gas index concentration data sequence and the environmental data are input into the trained long short-term memory network as input features, and the concentration prediction value of the odor gas in the farm in the future preset time is output .

[0041] As a further improvement of the technical solution, the specific steps of the second prediction branch module obtaining the water body pollutant concentration prediction value are as follows:

[0042] S321, based on the process efficiency influence factor, the reference maximum specific degradation rate of wastewater is corrected to obtain the corrected maximum specific degradation rate ;

[0043] S322, the corrected maximum specific degradation rate as a kinetic parameter, is input into a water quality kinetic model, and the water quality kinetic model takes the pollutant load, the real-time inflow and the corrected maximum specific degradation rate as inputs;

[0044] S323, the water quality kinetic model is run to simulate the degradation process of the pollutant in the treatment, and finally the concentration value of the effluent at the outlet in the future preset time is predicted .

[0045] As a further improvement of the technical solution, in the management decision unit, the specific steps of generating the control instruction are:

[0046] S41, the emission standard limit value is obtained, and the instantaneous exceeding ratio of the water body index and the instantaneous exceeding ratio of the gas are calculated according to the emission standard limit value, and the maximum value of the instantaneous exceeding ratio of the water body index and the instantaneous exceeding ratio of the gas is selected as the dominant exceeding ratio ;

[0047] S42, the dominant exceeding ratio is counted Length of time continuously exceeding preset threshold in future prediction period , and converting the length of time into a time persistence factor ;

[0048] S43, obtaining a final early warning level value based on the dominant over-standard ratio and the time persistence factor ;

[0049] S44, defining an optimization threshold , a standard limit threshold and an emergency response threshold , if , determining blue early warning, triggering feed ratio fine-tuning instruction;

[0050] if , triggering manure cleaning control instruction and feed ratio adjustment instruction;

[0051] if , triggering sewage treatment process parameter adjustment instruction, deodorization device control instruction and manure cleaning control instruction;

[0052] if , triggering highest-level sewage treatment instruction, highest-level deodorization instruction, production limit and suspension suggestion and artificial emergency intervention alarm.

[0053] Compared with the prior art, the present application has the following beneficial effects:

[0054] 1. In the pollutant emission prediction and management system for livestock and poultry farms, by constructing a dynamic multi-dimensional pollutant emission prediction model, the breeding density influence factor, the feed ingredient influence factor and the process efficiency influence factor are introduced into the prediction process for dynamic correction, realizing the accurate prediction of malodorous gas and water pollutant emission.

[0055] 2. In the pollutant emission prediction and management system for livestock and poultry farms, based on the prediction result, the early warning level is automatically determined and the targeted control instruction (such as feed ratio fine-tuning, manure cleaning frequency adjustment, sewage treatment process optimization, etc.) is generated, not only preventing the over-standard emission of pollutants in advance, but also deeply integrating with breeding production management, guiding actual breeding operation, and significantly improving the pollutant treatment efficiency and breeding environment management level. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 is the overall flowchart of the present application;

[0057] The meanings of various labels in the figure are as follows:

[0058] ​1, data acquisition and storage unit; 11, production data acquisition module; 12, pollution concentration monitoring module; 13, operation data acquisition module; 14, data storage module; 2, influence factor calculation unit; 21, gas factor influence module; 22, water body factor influence module; 3, model prediction and analysis unit; 31, first prediction branch module; 32, second prediction branch module; 4, management decision unit. DETAILED DESCRIPTION

[0059] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0060] Embodiment: Please refer to Figure 1 As shown in the figure, a pollutant emission prediction and management system for livestock and poultry farms is provided, which comprises a data acquisition and storage unit 1, an influence factor calculation unit 2, a model prediction and analysis unit 3, and a management decision unit 4.

[0061] The data acquisition and storage unit 1 is used to acquire multi-source data of the farm in real time; the multi-source data includes production data, pollutant concentration data, and facility operation data.

[0062] The data acquisition and storage unit 1 comprises a production data acquisition module 11, a pollution concentration monitoring module 12, an operation data acquisition module 13, and a data storage module 14.

[0063] The production data acquisition module 11 is used to acquire data representing the livestock and poultry breeding scale and nutrient supply of the farm.

[0064] The livestock and poultry breeding scale data includes livestock and poultry species, stocking rate, current average age, and effective area of livestock and poultry house; the nutrient supply data includes feed formula and crude protein content in the feed formula.

[0065] The pollution concentration monitoring module 12 is used to monitor real-time water body index concentration data reflecting the water pollution degree and real-time gas index concentration data reflecting the gas pollution degree; the index data of the water pollution degree includes the concentrations of chemical oxygen demand, ammonia nitrogen, total nitrogen, and total phosphorus, which are acquired by a water quality online monitoring instrument at a wastewater discharge port; the index data of the gas pollution degree includes the concentrations of ammonia and hydrogen sulfide, which are acquired by gas sensors deployed in livestock and poultry houses, compost sheds, and sewage storage pools.

[0066] The operation data acquisition module 13 is used to acquire key operation parameters of water treatment facilities.

[0067] The key operation data of the facility includes the operation data of the aerator, sludge return pump, fan, and dissolved oxygen content of the aeration tank;

[0068] The data storage module 14 is configured to store real-time multi-source data and historical multi-source data in the form of a time series database, and also store preset reference parameters of the dynamic multi-dimensional pollutant emission prediction model, the reference parameters including a reference fecal pollutant generation rate and a reference maximum specific degradation rate of sewage;

[0069] The influencing factor calculation unit 2 is configured to obtain key influencing factors of water quality change and malodorous gas based on production data and facility operation data of the farm; the key influencing factors include a breeding density influencing factor, a feed composition influencing factor, and a process efficiency influencing factor.

[0070] The influencing factor calculation unit 2 includes a gas factor influencing module 21 and a water body factor influencing module 22.

[0071] The gas factor influencing module 21 is configured to obtain the breeding density influencing factor and the feed composition influencing factor based on the production data of the farm.

[0072] In the gas factor influencing module 21, the specific method for obtaining the breeding density influencing factor is as follows:

[0073] S21, convert the number of livestock and poultry in stock into standard livestock and poultry units according to the breed conversion coefficient from the preset breed coefficient table;

[0074]

[0075] In the formula, is the standard livestock and poultry unit; is the number of livestock and poultry in stock; is the breed conversion coefficient based on the metabolic energy requirement;

[0076] The conversion of the standard livestock and poultry unit eliminates the metabolic differences between different livestock and poultry species, and realizes unified quantitative evaluation across farms and across species; the system has a preset breed coefficient table built-in, and automatically matches the metabolic energy requirement conversion coefficient of the corresponding breed according to the current livestock and poultry species (such as pigs, chickens, and cows) of the farm. For example, 1 adult pig is converted into 1 standard unit, 100 chickens are converted into 1 standard unit, and 1 cow is converted into 5 standard units. The system multiplies the actual number of livestock and poultry in stock by the coefficient to obtain the number of standardized livestock and poultry units, thereby solving the evaluation inaccuracy problem caused by species differences in traditional breeding density calculation, making the density indicators of different farms comparable, and providing a unified benchmark for subsequent dynamic correction.

[0077] S22, calculate the number of livestock per unit effective area to obtain the basic breeding density;

[0078]

[0079] wherein, is the basic breeding density; is the effective area of livestock and poultry house;

[0080] S23, calculating the age weight factor according to the current average age of livestock and poultry and the physiological mature age of the breed;

[0081]

[0082] wherein, is the age weight factor; is the current average age of livestock and poultry, is the physiological mature age of the breed of livestock and poultry, is the empirical growth factor, which is obtained by fitting historical data and is used to amplify the accelerating effect of metabolic waste production in the later growth stage. This step solves the problem of inaccurate density evaluation caused by ignoring the influence of the growth stage in the prior art.

[0083] This step characterizes the nonlinear influence of the growth stage of livestock and poultry on the production intensity of metabolic waste, and can obtain the current average age of livestock and poultry and the physiological mature age of the breed (such as the mature age of pigs being 180 days) in real time, and calculate the age weight factor in combination with the preset empirical growth factor (value obtained by fitting historical data, range 0.05-0.15). When the age of livestock and poultry reaches 60% of the mature age, the weight factor is 1+0.1×0.6=1.06. In the later growth stage of livestock and poultry (age close to maturity), the weight factor increases significantly, reflecting the characteristic of accelerated increase in metabolic waste production rate, making the density evaluation more in line with the actual physiological law and reducing the prediction deviation. S24, obtaining the breeding density influence factor based on the basic breeding density and the age weight factor;

[0084]

[0085]

[0086] wherein, is the breeding density influence factor; is the breeding density reference value;

[0087] In the gas factor influence module 21, the specific method for obtaining the feed ingredient influence factor is as follows:

[0088] S25, calculating the ratio of the crude protein content of the feed on the day to the standard crude protein reference content of the current growth stage of livestock and poultry based on the data of nutrient supply, to obtain the crude protein influence basic value;

[0089]

[0090] wherein, ​is the crude protein impact base value; is the crude protein content of the feed formula of the day; is the standard crude protein reference content of the growth stage;

[0091] S26, the ratio of the actual content of the key limiting amino acid in the feed formula of the day to the standard demand is obtained, and the amino acid balance degree coefficient is calculated;

[0092]

[0093] In the formula, is the amino acid balance degree coefficient; is the actual content of the key limiting amino acid; is the standard demand of the key limiting amino acid; is the type of the key limiting amino acid;

[0094] S27, based on the crude protein impact base value and the amino acid balance degree coefficient, the feed ingredient impact factor is obtained.

[0095]

[0096] In the formula, is the feed ingredient impact factor; is the weight coefficient;

[0097] This step introduces the concept of amino acid balance in nutrition, because unbalanced amino acid ratio will lead to decreased protein utilization and increased nitrogen emission, so as to more accurately predict the potential of nitrogen pollutants from the source.

[0098] The water body factor influencing module 22 is used to obtain a process efficiency impact factor based on key operation data of the facility;

[0099] In the water body factor influencing module 22, the implementation method of obtaining the process efficiency impact factor is as follows:

[0100] Based on the key operation data of the facility, the actual operation energy consumption, oxygen transfer efficiency and sludge reflux efficiency index are obtained, and the process efficiency impact factor is constructed according to the actual operation energy consumption, oxygen transfer efficiency and sludge reflux efficiency index.

[0101] Among them, the actual operation energy consumption includes aerator energy consumption, fan energy consumption and sludge reflux pump energy consumption; the oxygen transfer efficiency is evaluated by the fan air volume and the dissolved oxygen content of the aeration tank, and the sludge reflux efficiency index is calculated by combining the sludge reflux pump flow and the influent flow to obtain the reflux ratio, and the reflux efficiency index is defined by the reflux ratio.

[0102] If the dissolved oxygen content in the aeration tank is stable at 2-4 mg / L and the air volume is low, the aeration efficiency is high, if the dissolved oxygen content in the aeration tank fluctuates greatly or the air volume is high but the dissolved oxygen content is low, there is blockage or sludge aging, and the aeration efficiency decreases.

[0103] Further, the reflux ratio is set to be in the ideal range of 50%-100%, if , the reflux efficiency index is 1; if , the reflux efficiency index is 0.8; if , the reflux efficiency index is 0.5;

[0104] The model prediction analysis unit 3 is used to generate a predicted value of pollutant emission based on real-time multi-source data by using a dynamic multi-dimensional pollutant emission prediction model, and key influencing factors are introduced for correction in the prediction process; the dynamic multi-dimensional pollutant emission prediction model realizes prediction of different pollutant types (malodorous gas and water pollutant) through two independent but cooperative branch modules, and the two branch modules are operated in parallel.

[0105] The model prediction analysis unit 3 includes a first prediction branch module 31 and a second prediction branch module 32.

[0106] The first prediction branch module 31 is used to obtain a predicted value of malodorous gas concentration by using a long short-term memory network based on a benchmark fecal pollutant generation rate, combined with a breeding density influencing factor, a feed composition influencing factor and real-time gas index concentration data.

[0107] The specific steps of the first prediction branch module 31 for obtaining the predicted value of malodorous gas concentration are as follows:

[0108] S311, based on the breeding density influencing factor and the feed composition influencing factor, the benchmark fecal pollutant generation rate is corrected by weighted deviation coupling to obtain a corrected fecal pollutant generation rate ;

[0109]

[0110] , wherein is the weight coefficient of breeding density; is the weight coefficient of feed composition; both are weight coefficients determined by historical data training, used to represent the contribution degree of different influencing factors to the generation rate; is the benchmark fecal pollutant generation rate, and the two influencing factors are coupled in the form of weighted deviation, which more accurately reflects the combined influence of breeding practice on the source intensity of pollutants.

[0111] S312, the corrected fecal pollutant generation rate , combined with the current inventory, the generation intensity of malodorous gas ;

[0112] The intensity reflects the potential amount of malodor gas generated per unit time under a certain stocking density and feed formula.

[0113] S313, input the generation intensity of malodor gas, real-time gas index concentration data sequence, and environmental data as input features into the trained long short-term memory network, which learns the complex nonlinear relationship between generation intensity, environmental factors, and gas concentration, and outputs the concentration prediction value of malodor gas in the farm (livestock and poultry house, compost shed, sewage storage pool, etc.) within a preset time in the future , The malodor gas includes ammonia NH3 and hydrogen sulfide;

[0114] The module dynamically corrects the baseline fecal pollutant generation rate by introducing a stocking density impact factor and a feed composition impact factor, enabling the model to accurately quantify the source generation intensity of malodor gas under different farming practices. It realizes the leap from static and average estimation to dynamic and field-specific prediction. The corrected parameters more realistically reflect the direct impact of high-density farming, high-protein daily ration, and other realistic production conditions on the production of fecal excretion and nitrogen and sulfur-containing precursors, thereby providing accurate source input variables that reflect the current management state for subsequent long short-term memory network-based time series prediction. This greatly improves the accuracy and reliability of malodor gas concentration prediction values and provides a scientific basis for pre-positioned and precise malodor control decisions.

[0115] The second prediction branch module 32 is used to obtain water body pollutant concentration prediction values based on the baseline maximum specific degradation rate of sewage, combined with process efficiency impact factors and real-time water body index concentration data, using water quality kinetics model;

[0116] The specific steps for the second prediction branch module 32 to obtain water body pollutant concentration prediction values are as follows:

[0117] S321, correct the baseline maximum specific degradation rate of sewage based on the process efficiency impact factor to obtain the corrected maximum specific degradation rate ;

[0118]

[0119] PEF is the process efficiency impact factor, which is used to reflect the real-time biological treatment efficiency of the current sewage treatment system. is the baseline maximum specific degradation rate of sewage;

[0120] S322, input the corrected maximum specific degradation rate As kinetic parameters, they are input into the water quality kinetic model (specifically, a simplified differential equation based on the activated sludge model ASM1), which takes pollutant load, real-time influent flow rate, and the corrected maximum specific degradation rate as inputs.

[0121] S323. Run the water quality dynamics model to simulate the degradation process of pollutants during treatment, and finally predict the concentration of wastewater at the effluent within a preset time period. , Wastewater includes chemical oxygen demand (COD), ammonia nitrogen (NH3-N), total nitrogen (TN), and total phosphorus (TP).

[0122] This module calibrates the maximum specific degradation rate of baseline wastewater in real time using a process efficiency influencing factor, enabling the built-in water quality kinetic model to adaptively simulate the actual treatment efficiency of the wastewater treatment system under real-world operating conditions. This overcomes the inherent limitations of traditional mechanistic models, where key kinetic parameters remain fixed and cannot respond to changes in system state (such as fluctuations in sludge activity and water quality load shocks). Through this correction, the model is no longer merely a theoretical deduction tool under ideal conditions, but a dynamic digital twin capable of sensing and reflecting the "health status" of the treatment facility. This allows for more accurate prediction of pollutant concentrations at the effluent under different influent loads and operating conditions, providing crucial support for the refined management and optimized control of the wastewater treatment process.

[0123] The management decision-making unit 4 is used to obtain the ratio of the predicted emission value to the emission standard based on the predicted value of odor gas concentration and the predicted value of water pollutant concentration, determine the warning level according to the ratio, and generate control instructions according to the warning level.

[0124] In management decision-making unit 4, the specific steps for generating control instructions are as follows:

[0125] S41. Obtain emission standard limits. Emission standard limits are determined based on local emission standards. Calculate the instantaneous exceedance ratio of water body indicators based on the emission standard limits. The ratio of instantaneous exceedance of gas standards At the same time, the instantaneous exceedance ratio of water body indicators was selected. The ratio of instantaneous exceedance of gas standards The maximum value in the range is used as the dominant exceedance ratio. ;

[0126]

[0127]

[0128] In the formula, These are the limits for water pollution discharge standards; The emission limits for malodorous gases;

[0129] S42, count the dominant over-standard ratio the length of time that continuously exceeds the preset threshold value in the future prediction period , and convert the length of time into a time duration factor ;

[0130]

[0131] wherein, is a proportional coefficient. The introduction of the time duration factor allows the risk of short-term instantaneous over-standard and long-term persistent over-standard to be distinguished, the latter of which means that the problem is more serious and requires more decisive intervention.

[0132] S43, based on the dominant over-standard ratio and the time duration factor , obtain the final early warning level value ;

[0133]

[0134] wherein, is a weight coefficient of the time duration factor;

[0135] S44, define the optimization threshold value (value 0.8), the standard limit threshold value (value 1.0), and the emergency response threshold value (value 1.5), if , it is determined as a blue early warning, triggering a feed ratio fine-tuning instruction; under the premise of meeting the nutritional needs of livestock and poultry, the crude protein content is slightly reduced (reduced by 0.5%), and essential amino acids such as lysine are supplemented at the same time to maintain balance. This instruction aims to preventively reduce the intake and excretion of nitrogen from the source, rather than remedying afterwards.

[0136] if , trigger the manure cleaning control instruction and the feed ratio adjustment instruction;

[0137] The manure cleaning control instruction: the system predicts that there is a risk of rising to over-standard of odor or water concentration, and sends an instruction to the manure cleaning control system in advance to increase the manure cleaning frequency (from 2 times a day to 3 times a day), which aims to reduce the residence time of manure in the shed and cut off the process of odor generation and water pollution load;

[0138] The feed ratio adjustment instruction: execute the instruction of the blue early warning, but the adjustment range is larger (reduced by 1%).

[0139] if , trigger the sewage treatment process parameter adjustment instruction, the deodorization device control instruction, and the manure cleaning control instruction;

[0140] Among them, the sewage treatment instruction: the system confirms that the water body pollutant prediction value has exceeded or will exceed the standard, immediately sends an instruction to the sewage treatment PLC controller, the instruction contains increasing the sludge return flow (10%) and improving the aeration intensity (15%), if the total nitrogen prediction value is too high, it will also add the instruction of adding external carbon source (sodium acetate), forming a linkage adjustment of the treatment process;

[0141] Deodorization device control instruction: send the start instruction to the spray deodorization system, and dynamically set the spray frequency and reagent dosage according to the predicted malodor concentration curve, realize accurate deodorization on demand, and avoid waste caused by blind operation;

[0142] Cleaning control instruction: continue to execute, the frequency is increased to the highest range.

[0143] If Then trigger the highest level sewage treatment instruction, the highest level deodorization instruction, the production limit and suspension suggestion and the artificial emergency intervention alarm.

[0144] Among them, the sewage treatment instruction: instruct all sewage treatment facilities to run at the maximum design capacity (aeration intensity 100%, maximum carbon source dosage); and start the emergency pool or backup treatment;

[0145] Deodorization instruction: instruct all deodorization devices (spray, biological filter fan, etc.) to run at maximum power;

[0146] Production intervention suggestion: from the end treatment to the production end, generate a suggestive instruction: "suggest to suspend new batches of livestock and poultry into the barn" or "suggest to temporarily remove the livestock and poultry in a certain shed to reduce the density", directly reduce the pollution load from the source;

[0147] Manual alarm: immediately send the highest level alarm to the management personnel's mobile phone and monitoring center, prompting the need for manual intervention for inspection and treatment.

[0148] The above shows and describes the basic principles, main features and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only preferred examples of the present application and are not intended to limit the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application.

Claims

1. A pollutant emission prediction and management system for livestock farms, characterized in that, The application relates to a dynamic multi-dimensional pollutant emission prediction model for aquaculture farms. The application comprises: a data acquisition and storage unit (1) for acquiring multi-source data of an aquaculture farm in real time; the multi-source data comprises production data, pollutant concentration data and facility operation data; an influencing factor calculation unit (2) for acquiring key influencing factors of water quality change and malodorous gas based on the production data and the facility operation data of the aquaculture farm; a model prediction and analysis unit (3) for generating a predicted value of pollutant emission by using a dynamic multi-dimensional pollutant emission prediction model based on real-time multi-source data, and introducing key influencing factors for correction in the prediction process; the model prediction and analysis unit (3) comprises a first prediction branch module (31) and a second prediction branch module (32); the first prediction branch module (31) is used for acquiring a predicted value of malodorous gas concentration by using a long short-term memory network based on a benchmark manure pollutant generation rate, in combination with a breeding density influencing factor, a feed composition influencing factor and real-time gas index concentration data; the second prediction branch module (32) is used for acquiring a predicted value of water body pollutant concentration by using a water quality dynamics model based on a benchmark maximum sewage degradation rate, in combination with a process efficiency influencing factor and real-time water body index concentration data; S311、based on the breeding density influence factor and the feed composition influence factor, the reference fecal pollutant production rate is corrected by weighted deviation coupling to obtain the corrected fecal pollutant production rate ; S312、corrected fecal pollutant generation rate , and in combination with the current inventory, obtain the generation intensity of malodorous gas ; S313, input the generation intensity of the malodorous gas, the real-time gas index concentration data sequence, and the environmental data as input features into the trained long short-term memory network, and output the concentration prediction value of the malodorous gas in the farm in the future preset time ; the specific steps of acquiring the predicted value of malodorous gas concentration by the first prediction branch module (31) are as follows: S321、based on the process efficiency influence factor, correct the baseline sewage maximum degradation rate to obtain a corrected maximum degradation rate ; S322, The corrected maximum specific degradation rate As kinetic parameters, they are input into the water quality kinetic model, which takes pollutant load, real-time influent flow rate and corrected maximum specific degradation rate as inputs. S323, running the water quality kinetics model to simulate the degradation process of the pollutants in the treatment, and finally predicting the concentration value of the effluent at the outlet within a preset time in the future ; the specific steps of acquiring the predicted value of water body pollutant concentration by the second prediction branch module (32) are as follows: a management decision unit (4) for acquiring a ratio of a predicted emission value to an emission standard based on the predicted value of malodorous gas concentration and the predicted value of water body pollutant concentration, determining an early warning level according to the ratio, and generating a control instruction according to the early warning level; S41, obtain the emission standard limit value, calculate the instantaneous exceeding ratio of the water body index according to the emission standard limit value and the instantaneous exceeding ratio of the gas , select the maximum value in the instantaneous exceeding ratio of the water body index and the instantaneous exceeding ratio of the gas as the dominant exceeding ratio ; S42, count the leading over-standard ratio the length of time continuously exceeding the preset threshold in the future prediction period and convert the length of time into a time duration factor ; S43, based on the dominant over-standard ratio value and the time persistence factor to obtain the final early warning level value ; S44, define optimization threshold , standard limit threshold and emergency response threshold , if , determine blue early warning, trigger feed ratio fine tuning instruction; If , the manure cleaning control instruction and the feed ratio adjustment instruction are triggered. If , a sewage treatment process parameter adjustment instruction, a deodorization device control instruction, and a manure cleaning control instruction are triggered. If then a highest level sewage treatment order, a highest level deodorization order, production limitation and suspension recommendations and an artificial emergency intervention alarm are triggered.

2. The pollutant emission prediction and management system for livestock and poultry farms as claimed in claim 1 wherein: the specific steps of generating the control instruction in the management decision unit (4) are as follows: the data acquisition and storage unit (1) comprises a production data acquisition module (11), a pollution concentration monitoring module (12), an operation data acquisition module (13) and a data storage module (14); the production data acquisition module (11) is used for acquiring data representing the scale of livestock and poultry breeding and the supply of nutrients in the aquaculture farm; the scale of livestock and poultry breeding data comprises livestock and poultry types, the number of livestock and poultry, the current average age and the effective area of livestock and poultry sheds; the supply of nutrients data comprises a feed formula and the crude protein content in the feed formula; the pollution concentration monitoring module (12) is used for monitoring real-time water body index concentration data reflecting the degree of water body pollution and real-time gas index concentration data reflecting the degree of gas pollution; the operation data acquisition module (13) is used for acquiring key operation parameters of a water treatment facility; the key operation data of the facility comprise operation data of aerators, sludge reflux pumps and fans and the dissolved oxygen content of an aeration tank; the data storage module (14) is used for storing real-time multi-source data and historical multi-source data in the form of a time series database, and also storing benchmark parameters preset in the dynamic multi-dimensional pollutant emission prediction model, the benchmark parameters comprising a benchmark manure pollutant generation rate and a benchmark maximum sewage degradation rate.

3. The pollutant emission prediction and management system for livestock and poultry farms as claimed in claim 2 wherein: The influence factor calculation unit (2) comprises a gas factor influence module (21) and a water body factor influence module (22); The gas factor influence module (21) is configured to obtain a breeding density influence factor and a feed composition influence factor based on production data of the farm; The water body factor influence module (22) is configured to obtain a process efficiency influence factor based on key operation data of the facility.

4. The pollutant emission prediction and management system for livestock and poultry farms as claimed in claim 3 wherein: In the gas factor influence module (21), the specific method for obtaining the breeding density influence factor is as follows: S21, converting the number of livestock and poultry in stock into standard livestock and poultry units according to the breed conversion coefficient obtained from the pre-set breed coefficient table according to the type of livestock and poultry; S22, calculating the number of livestock and poultry per unit of effective area to obtain the basic breeding density; S23, calculating the age weight factor according to the current average daily age of the livestock and poultry and the physiological mature age of the breed; S24, obtaining the breeding density influence factor based on the basic breeding density and the age weight factor.

5. The pollutant emission prediction and management system for livestock and poultry farms as claimed in claim 4 wherein: In the gas factor influence module (21), the specific method for obtaining the feed composition influence factor is as follows: S25, calculating the ratio of the crude protein content of the feed on the day to the standard crude protein reference content of the current livestock and poultry growth stage based on the data of the nutrient supply to obtain the crude protein influence basic value; S26, obtaining the ratio of the actual content of the key limiting amino acid in the feed formula on the day to the standard demand amount, and calculating the amino acid balance degree coefficient; S27, obtaining the feed composition influence factor based on the crude protein influence basic value and the amino acid balance degree coefficient.

6. The pollutant emission prediction and management system for livestock and poultry farms as claimed in claim 5 wherein: In the water body factor influence module (22), the implementation method for obtaining the process efficiency influence factor is as follows: Based on the key operation data of the facility, the actual operation energy consumption, the oxygen transfer efficiency and the sludge reflux efficiency index are obtained, and the process efficiency influence factor is constructed according to the actual operation energy consumption, the oxygen transfer efficiency and the sludge reflux efficiency index.

Images