An automatic control system and method for sludge treatment

Abstract

The present invention discloses an automatic control system and method for sludge treatment, which relates to the technical field of sludge treatment. The environmental state feature vector F is collected in real time through an integrated sensor, and the odor concentration prediction result vector Cpred at different time points in the future is calculated based on a dynamic odor diffusion prediction model, and then the decision vector D of the amount of additives is dynamically adjusted in combination with the predicted concentration trend, breaking through the limitation of the prior art that only relies on real-time odor concentration feedback for dosing control, and effectively reducing the dosing lag problem by combining the Gaussian plume diffusion model with the time series prediction method. At the same time, the benefit index E is evaluated by adjusting the concentration vector Cafter and the odor concentration change vector Gafter, so that the dosing strategy is not only optimized based on the concentration change trend, but also fully considers factors such as the cost of using the additive and the stability of odor diffusion, solving the defect that the dosage in the traditional method is difficult to balance the deodorization effect and cost.

Description

Technical Field

[0001] The present invention relates to the technical field of sludge treatment, and specifically provides an automatic control system and method for sludge treatment. Background Art

[0002] In the field of environmental protection and resource recycling, sewage treatment is one of the important links in urban and industrial development. Sewage treatment has become an important part of urban infrastructure construction. However, a large amount of sludge is generated during the sewage treatment process. These sludges usually contain organic matter, heavy metals, pathogenic microorganisms, etc., which pose potential threats to the environment and human health. Therefore, the safe disposal of sludge has become an important issue in the environmental protection industry.

[0003] In the existing automatic control system for sludge treatment, the deodorization strategy mainly relies on adding drugs for adjustment after real-time monitoring of the odor concentration, which belongs to a typical feedback control mode. Although this method can control the odor to a certain extent, there are multiple limitations. For example: reaction lag, untimely deodorization: due to the certain lag in the diffusion of odors, the odor value detected by the current concentration may already be the result after excessive diffusion. This means that even if drugs are added immediately, it is impossible to quickly suppress the odor, and the surrounding environment will still be polluted in a short time.

[0004] Drug waste, increased cost: due to the failure to predict the change trend of odors in advance, the system often uses a higher dosage of drugs to ensure the deodorization effect, resulting in waste of some chemical reagents, increasing the cost of sludge treatment, and at the same time increasing the possibility of secondary pollution of sewage.

[0005] Lack of comprehensive consideration of environmental factors: the current odor control is only based on concentration detection, without considering key factors affecting odor diffusion such as temperature, humidity, wind speed, and air pressure. For example, in a high-temperature and high-humidity environment, odors volatilize faster, while in a low-temperature and low-humidity environment, their diffusion is slower. If the drug dosage is not adjusted in combination with these factors, it is easy to cause problems such as insufficient or excessive drug addition.

[0006] Local control, difficult to optimize globally: the current deodorization strategy often controls a single odor monitoring point, without establishing a global diffusion model, and is unable to predict the odor diffusion trend in different areas of the plant area, resulting in a relatively high odor concentration in some areas, while the drug dosage in other areas is excessive. Summary of the Invention

[0007] In view of the deficiencies of the prior art, the present invention provides an automatic control system and method for sludge treatment, which solves the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention is realized through the following technical solutions: an automatic control method for sludge treatment, including the following steps:

[0009] S1. Collect environmental information and odor concentration information during the sludge treatment process in real time through integrated sensors, and form an environmental state feature vector F;

[0010] S2. Based on the obtained environmental state feature vector F, construct a dynamic odor diffusion prediction model, calculate the odor concentration change trend at different time points, and obtain an odor concentration prediction result vector CP;

[0011] S3. Calculate based on the obtained odor concentration prediction result vector CP to obtain a decision vector D for the dosage of additives;

[0012] S4. Simulate the change in odor concentration after applying the decision vector D to obtain an odor concentration change vector Gqs calculated by simulating the odor concentration and an adjusted concentration vector Cafter;

[0013] S5. Based on the obtained adjusted concentration vector Cafter, decision vector D, and odor concentration change vector Gqs, conduct a comprehensive evaluation to obtain a benefit index E, and adjust the dosage of additives according to the benefit index E to obtain an execution decision vector D1, and adaptively adjust the sludge treatment process according to the execution decision vector D1;

[0014] S6. Record the actual odor concentration Cfinal after executing the decision vector D1, calculate the deviation △C from the preset target odor concentration, and trigger an iterative optimization mechanism according to the deviation △C.

[0015] Preferably, the S1 includes S11 and S12;

[0016] S11. Collect environmental information affecting odor diffusion during the sludge treatment process in real time through a variety of integrated sensors, and form an environmental information vector Fel. The environmental information vector Fel includes temperature T, humidity H, wind speed W, wind direction θ, air pressure P, and estimated odor source intensity S;

[0017] Among them, the sludge treatment process includes a sludge storage process, a conditioning process, a dehydration process, and a stacking process;

[0018] Among them, the variety of sensors include a temperature sensor, a humidity sensor, an ultrasonic anemometer, a wind direction sensor, and an air pressure sensor;

[0019] The estimated odor source intensity S is obtained by estimating the volatilization of organic matter in the sludge. After measuring the organic matter content Corg in the sludge, the estimated odor source intensity S is obtained by using an empirical estimation formula in combination with temperature T and humidity H;

[0020] The estimated odor source intensity S is obtained through the following empirical estimation formula:

[0021] ;

[0022] In the formula, β1, β2 and β3 represent empirical coefficients. Specifically, β1 represents the empirical coefficient of the organic matter content Corg in the sludge, β2 represents the empirical coefficient of the humidity H, and β3 represents the empirical coefficient of the temperature T.

[0023] Preferably, S12: The odor concentration information is collected in real time by odor sensors arranged at multiple locations in the treatment plant. After processing the odor concentration information obtained at multiple locations, the average odor concentration Cobs of the treatment plant is obtained, and combined with the environmental information vector Fel to form the environmental state feature vector F, and the time stamp information is marked and stored at fixed intervals as historical data.

[0024] The environmental state feature vector F is specifically F = {Fel, Cobs};

[0025] Among them, the odor sensor includes using a gas detection sensor to measure the odor concentration in the air and obtaining the odor concentration Cobs(x, y) at the location (x, y) of the treatment plant.

[0026] Preferably, the S2 includes S21 and S22;

[0027] S21: Based on the obtained environmental state feature vector F and the principle of aerodynamics, a dynamic odor diffusion prediction model is established. The dynamic odor diffusion prediction model includes using the Gaussian plume diffusion model for establishment to simulate the diffusion process of odor in the air.

[0028] By using the established dynamic odor diffusion prediction model to simulate the diffusion of odor in the air affected by the wind speed W, wind direction θ and the estimated intensity S of the odor source, the model predicted odor concentration Cmodel(x, y, t) at the location (x, y) at different times t is calculated, and the set YCmodel of the model predicted concentrations at multiple monitoring point locations is obtained.

[0029] The predicted odor concentration Cmodel(x, y, t) is obtained through the following calculation formula:

[0030] ;

[0031] In the formula, π represents the mathematical constant, with a value of 3.14159, and represents the diffusion coefficient of the odor, specifically representing the degree of diffusion in the horizontal y direction and the vertical z direction. exp represents the exponential function, and G represents the pollutant release height, which is set according to the geometric characteristics of the sludge tank.

[0032] Preferably, in S22, by extracting the historically stored environmental state feature vectors F, which are marked as the historical odor concentration dataset Hc, and performing short-term prediction with the environmental information vector Fel and the model prediction concentration set YCmodel. The short-term prediction is performed using an autoregressive integrated moving average model to obtain the predicted concentration Cpred(t + △t) after a period of △t. Then, in combination with the model predicted odor concentration Cmodel(x, y, t), the predicted concentration Cpred(t + △t) is dynamically corrected to obtain the predicted concentration Cpred(x, y, t + △t) at the position (x, y) at the future time t + △t. And by integrating multiple time points of future predictions, an odor concentration prediction result vector CP is formed;

[0033] The odor concentration prediction result vector CP is specifically CP = {Cpred(x, y, t + 1), Cpred(x, y, t + 2), ……, Cpred(x, y, t + △t)}

[0034] The predicted concentration Cpred(x, y, t + △t) at the future time t + △t at the position (x, y) is obtained through the following dynamic correction formula:

[0035] ;

[0036] In the formula, λ represents the model weight coefficient, and its value range is 0 < λ < 1, represents the odor concentration at the position (x, y) at the current time t, m represents the maximum time step of backtracking, γ(k) represents the weight coefficient of the backtracking time k, and Cobs(x, y, t - k) represents the odor concentration at the position (x, y) at the backtracking time t - k.

[0037] Preferably, the S3 includes S31;

[0038] S31、By extracting the historical auxiliary agent dosing record data, which is marked as the historical main auxiliary agent dosing record HD, and performing calculations with the obtained odor concentration prediction result vector CP to obtain the simulated concentration change vector L and the decision vector D of the auxiliary agent dosing amount;

[0039] The simulated concentration change vector L is obtained through the following calculation formula:

[0040] ;

[0041] In the formula, L(t + △t) represents the odor concentration change vector at the future time t + △t, and △t represents the future time step;

[0042] The decision vector D is obtained through the following calculation formula:

[0043] ;

[0044] In the formula, D(t) represents the decision vector at the current time t, and d1, d2, and d3 represent adjustment coefficients. Specifically, d1 represents the adjustment coefficient of the simulated concentration change vector L(t + △t) at the future time t + △t, d2 represents the adjustment coefficient of the predicted concentration Cpred(x, y, t + △t) at the future time t + △t at the position (x, y), and d3 represents the adjustment coefficient of the wind speed W.

[0045] Preferably, the S4 includes S41;

[0046] S41. After applying the corresponding decision vector D, the odor concentration state in the air will change. Based on the odor concentration prediction result vector CP and the decision vector D, a simulation is performed to obtain the adjusted concentration vector Cafter, the adjusted odor concentration change vector Gqs, and the additive removal rate vector R;

[0047] The adjusted concentration vector Cafter is obtained through the following calculation formula:

[0048] ;

[0049] In the formula, Cafter(t + △t) represents the adjusted concentration vector at the future time t + △t, and D(t + △t) represents the applied decision vector at the future time t + △t. represents the additive removal efficiency coefficient;

[0050] The adjusted odor concentration change vector Gqs is obtained through the following calculation formula:

[0051] ;

[0052] In the formula, Gqs(t + △t) represents the adjusted odor concentration change vector at the future time t + △t;

[0053] The additive removal rate vector R is obtained through the following calculation formula:

[0054] ;

[0055] In the formula, R(t + △t) represents the additive removal rate vector at the future time t + △t.

[0056] Preferably, the S5 includes S51;

[0057] S51. Based on the obtained adjusted concentration vector Cafter, decision vector D, and odor concentration change vector Gqs, conduct a comprehensive evaluation to obtain the benefit index E, and adjust the dosage of the regulating agent according to the benefit index E to obtain the execution decision vector D1. Then, adaptively regulate the sludge treatment process according to the execution decision vector D1, and set boundary constraints for the execution decision vector D1 at the same time.

[0058] Among them, the execution decision vector D1 adaptively regulates the sludge treatment process by controlling the operation of the regulating agent dosing pump and dosing device to achieve adaptive control.

[0059] The benefit index E is obtained through the following calculation formula:

[0060] ;

[0061] In the formula, Ctarget represents the preset target odor concentration, and e1, e2, and e3 represent the equity weight coefficients.

[0062] The execution decision vector D1 is obtained through the following calculation formula:

[0063] ;

[0064] In the formula, D1(t + △t) represents the execution decision vector at the future time t + △t. represents the step size factor, which is specifically used to control the adjustment speed. represents the gradient of the benefit index with respect to the dosage of the regulating agent, which specifically represents adjusting the dosage of the regulating agent to minimize the benefit index E.

[0065] The boundary constraints are as follows:

[0066] Dmin ≤ D1(t + △t) ≤ Dmax, where Dmin represents the minimum dosage of the regulating agent, and Dmax represents the maximum dosage of the regulating agent.

[0067] Preferably, the S6 includes S61.

[0068] S61. Record the actual odor concentration Cfinal after the execution decision vector D1. The actual odor concentration Cfinal is obtained by extracting the odor concentration Cobs(x, y) at the position (x, y), calculate the deviation △C from the preset target odor concentration, and compare the deviation △C with the preset error threshold V to trigger the iterative optimization mechanism.

[0069] The deviation △C is obtained through the calculation formula, where △C(t) represents the deviation at the current time t.

[0070] The triggering method of the iterative optimization mechanism is as follows:

[0071] When the deviation ΔC(t) at the current time t > the error threshold V, it indicates that the iterative optimization mechanism is triggered, and the adjustment of the auxiliary agent dosing strategy is executed, including adjusting the proportion and executing the decision vector D1 to execute the adjustment of the auxiliary agent dosing strategy;

[0072] When the deviation ΔC(t) at the current time t ≤ the error threshold V, it indicates that the iterative optimization mechanism is not triggered.

[0073] An automatic control system for sludge treatment includes an information collection module, a trend analysis module, a trend decision module, a simulation prediction module, an optimization module, and an iterative trigger module;

[0074] The information collection module collects the environmental information and odor concentration information in the sludge treatment process in real time through the integrated sensors, and forms the environmental state feature vector F;

[0075] The trend analysis module constructs a dynamic odor diffusion prediction model based on the obtained environmental state feature vector F, calculates the odor concentration change trend at different time points, and obtains the odor concentration prediction result vector CP;

[0076] The trend decision module calculates based on the obtained odor concentration prediction result vector CP to obtain the decision vector D of the auxiliary agent dosing amount;

[0077] The simulation prediction module simulates the odor concentration change after applying the decision vector D, and obtains the odor concentration change vector Gqs and the adjusted concentration vector Cafter calculated from the odor concentration;

[0078] The optimization module conducts a comprehensive evaluation based on the obtained adjusted concentration vector Cafter, decision vector D, and odor concentration change vector Gqs, obtains the benefit index E, and adjusts the auxiliary agent dosing amount according to the benefit index E to obtain the execution decision vector D1, and adaptively controls the sludge treatment process according to the execution decision vector D1;

[0079] The iterative trigger module records the actual odor concentration Cfinal after executing the decision vector D1, calculates the deviation ΔC from the preset target odor concentration, and triggers the iterative optimization mechanism according to the deviation ΔC.

[0080] The present invention provides an automatic control system and method for sludge treatment, having the following beneficial effects:

[0081] (1) The environmental state feature vector F is collected in real time through an integrated sensor, and the predicted odor concentration result vector Cpred at different future time points is calculated based on a dynamic odor diffusion prediction model. Then, in combination with the predicted concentration trend, the decision vector D for the dosing amount of the additive is dynamically adjusted. This method breaks through the limitation of the existing technology that only relies on real-time odor concentration feedback for dosing control. By combining the Gaussian plume diffusion model with the time series prediction method, it improves the foresight and accuracy of the dosing decision, effectively reducing the dosing lag problem. At the same time, the benefit index E is evaluated through the adjusted concentration vector Cafter and the odor concentration change vector Gafter, enabling the dosing strategy to be optimized not only based on the concentration change trend but also fully considering factors such as the cost of additive use and the stability of odor diffusion, solving the defect in the traditional method that it is difficult to balance the deodorization effect and cost in terms of the dosing amount.

[0082] (2) By extracting the historical main additive dosing records HD and combining with the predicted odor concentration result vector Cpred, the simulated concentration change vector L and the decision vector D for the additive dosing amount at future time t + △t are calculated, realizing the intelligent decision optimization of the additive dosing strategy. Compared with the traditional method that solely relies on real-time odor concentration feedback for dosing the additive, this method can combine historical data with the predicted trend, ensuring that the dosing decision not only considers the future odor concentration change trend but also can dynamically adapt to environmental parameters such as the wind speed W, thus avoiding the situation of over-dosing or under-dosing.

[0083] (3) By calculating the benefit index E based on the adjusted concentration vector Cafter, the decision vector D, and the adjusted odor concentration change vector Gqs, the global optimization of the additive dosing strategy is realized. Different from the traditional method that only focuses on a single index for dosing amount adjustment, this method quantifies factors such as the odor removal effect, the additive dosing cost, and the smoothness of the odor concentration change through the benefit index E, and combines the boundary constraints of the execution decision vector D1, enabling the additive dosing amount to not only meet the requirement of the odor concentration standard but also minimize the use amount of the additive, reducing resource waste. It can also adapt to different environmental conditions, effectively improving the long-term operation efficiency of the sludge treatment system and ensuring that the system always maintains the optimal balance point of deodorization and energy conservation and consumption reduction. Description of the Drawings

[0084] Figure 1 It is a block diagram flow schematic diagram of an automatic control method for sludge treatment according to the present invention;

[0085] Figure 2 It is a step schematic diagram of an automatic control system for sludge treatment according to the present invention. Detailed Embodiments

[0086] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0087] Embodiment 1

[0088] The present invention provides an automatic control method for sludge treatment. Please refer to Figure 1 , which includes the following steps:

[0089] S1. Real-time collect the environmental information and odor concentration information during the sludge treatment process through the integrated sensors, and form an environmental state feature vector F;

[0090] S2. Based on the obtained environmental state feature vector F, construct a dynamic odor diffusion prediction model, calculate the odor concentration change trend at different time points, and obtain an odor concentration prediction result vector CP;

[0091] S3. Calculate based on the obtained odor concentration prediction result vector CP to obtain a decision vector D for the additive dosage;

[0092] S4. Simulate the odor concentration change after applying the decision vector D to obtain an odor concentration change vector Gqs and an adjusted concentration vector Cafter calculated from the odor concentration;

[0093] S5. Based on the obtained adjusted concentration vector Cafter, decision vector D, and odor concentration change vector Gqs, conduct a comprehensive evaluation to obtain a benefit index E, and adjust the additive dosage according to the benefit index E to obtain an execution decision vector D1, and adaptively adjust the sludge treatment process according to the execution decision vector D1;

[0094] S6. Record the actual odor concentration Cfinal after executing the decision vector D1, calculate the deviation △C from the preset target odor concentration, and trigger an iterative optimization mechanism according to the deviation △C.

[0095] In this embodiment, the environmental state feature vector F is collected in real time through the integrated sensors, and the predicted result vector Cpred of the odor concentration at different future time points is calculated based on the dynamic odor diffusion prediction model. Then, in combination with the predicted concentration trend, the decision vector D of the additive dosage is dynamically adjusted. This method breaks through the limitation of the existing technology that only relies on the real-time odor concentration feedback for dosing control. By combining the Gaussian plume diffusion model with the time series prediction method, it improves the foresight and accuracy of the dosing decision, and effectively reduces the dosing lag problem. At the same time, the benefit index E is evaluated through the adjusted concentration vector Cafter and the odor concentration change vector Gafter, so that the dosing strategy is optimized not only based on the concentration change trend, but also fully considers factors such as the cost of additive use and the stability of odor diffusion, solving the defect that it is difficult to balance the deodorization effect and cost in the traditional method. In addition, by recording the actual odor concentration Cfinal after executing the decision vector D1 in real time and calculating the odor concentration deviation vector △C, an iterative optimization mechanism is formed, enabling the additive dosing to dynamically adapt to different environmental changes, achieving precise control, and avoiding the problem of excessive or insufficient dosing caused by environmental condition fluctuations in the traditional method. Therefore, this method significantly improves the intelligence level of the sludge treatment system, the stability of the deodorization effect, and the utilization efficiency of the additive, realizing the comprehensive optimization of automation, predictive control, and energy conservation and consumption reduction, and greatly enhancing the environmental protection and economy of the sludge treatment process.

[0096] Example 2

[0097] This embodiment is an explanatory description based on Embodiment 1. Please refer to Figure 1 , specifically: S1 includes S11 and S12;

[0098] S11. The environmental information affecting odor diffusion during the sludge treatment process is collected in real time through a variety of integrated sensors, and an environmental information vector Fel is formed. The environmental information vector Fel includes temperature T, humidity H, wind speed W, wind direction θ, air pressure P, and the estimated intensity S of the odor source;

[0099] Among them, the sludge treatment process includes the sludge storage process, conditioning process, dehydration process, and stacking process;

[0100] Among them, the variety of sensors include a temperature sensor, a humidity sensor, an ultrasonic anemometer, a wind direction sensor, and an air pressure sensor;

[0101] The estimated intensity S of the odor source is obtained by estimating the volatilization of organic matter in the sludge. After measuring the organic matter content Corg in the sludge, the estimated intensity S of the odor source is obtained by using an empirical estimation formula in combination with temperature T and humidity H;

[0102] The estimated intensity S of the odor source is obtained through the following empirical estimation formula:

[0103] ;

[0104] In the formula, β1, β2, and β3 represent empirical coefficients. Specifically, β1 represents the empirical coefficient of the organic matter content Corg in the sludge, β2 represents the empirical coefficient of the humidity H, and β3 represents the empirical coefficient of the temperature T.

[0105] S12. Real-time collect the odor concentration information through odor sensors deployed at multiple locations in the treatment plant. After processing the obtained odor concentration information at multiple locations, obtain the average odor concentration Cobs of the treatment plant, and combine it with the environmental information vector Fel to form the environmental state feature vector F, and mark the timestamp information for fixed-period storage as historical data;

[0106] The environmental state feature vector F is specifically F = {Fel, Cobs};

[0107] Among them, the odor sensor includes using a gas detection sensor to measure the odor concentration in the air and obtain the odor concentration Cobs(x, y) at the location (x, y) of the treatment plant.

[0108] In this embodiment, through a variety of integrated sensors, real-time collect the environmental information affecting odor diffusion to form the environmental information vector Fel, and combine the odor sensors arranged at multiple points to obtain the odor concentration information Cobs. Finally, construct the environmental state feature vector F including the timestamp for fixed-period storage. Compared with the traditional mode of single-time-point sampling or periodic detection in the sludge treatment process, this method can continuously and real-time monitor environmental changes. The estimated intensity S of the odor source calculated by the empirical formula can not only accurately reflect the volatilization of organic matter in the sludge, but also provide supplementary data when the actual monitoring data is insufficient or the sensor is abnormal, thus avoiding the problem of misjudgment of the odor concentration caused by abnormal data of a single sensor. Combining the spatial layout strategy of multiple odor sensors provides more stable input data for the subsequent dynamic odor diffusion prediction model, improving the robustness and environmental adaptability of the prediction.

[0109] Embodiment 3

[0110] This embodiment is an explanatory description based on Embodiment 1. Please refer to Figure 1 , specifically: The S2 includes S21 and S22;

[0111] S21. Based on the obtained environmental state feature vector F and the principle of aerodynamics, establish a dynamic odor diffusion prediction model. The dynamic odor diffusion prediction model includes using the Gaussian plume diffusion model for establishment to simulate the diffusion process of odor in the air;

[0112] By simulating the diffusion of odor in the air through the established dynamic odor diffusion prediction model, the diffusion of odor is affected by the wind speed W, wind direction θ, and the estimated intensity S of the odor source. Calculate the predicted odor concentration Cmodel(x, y, t) at the position (x, y) at different times t, and obtain the set YCmodel of predicted concentrations at multiple monitoring point positions;

[0113] The predicted odor concentration Cmodel(x, y, t) is obtained through the following calculation formula:

[0114] ;

[0115] In the formula, π represents the mathematical constant, with a value of 3.14159, and represent the diffusion coefficients of the odor, specifically representing the degree of diffusion in the horizontal y direction and the vertical z direction. exp represents the exponential function, and G represents the pollutant release height, which is set through the geometric characteristics of the sludge tank; is used to calculate the basic concentration of the odor. The ratio of the estimated intensity S of the odor source to the wind speed W determines the influence of the wind speed on the spread of the odor, and represent the diffusion range of the odor. The larger the diffusion range, the lower the odor concentration per unit volume, represents the influence of calculating the diffusion in the horizontal y direction. y² represents the offset from the horizontal y-axis. The farther away from the y-axis, the lower the concentration; represents the influence of calculating the vertical z direction, represents the offset between the height z of a certain point and the source height G. The farther away from the height G, the lower the concentration.

[0116] S22. By extracting the historical stored environmental state feature vector F, marked as the historical odor concentration dataset Hc, and performing short-term prediction with the environmental information vector Fel and the set YCmodel of model predicted concentrations. The short-term prediction is carried out using the autoregressive integrated moving average model to obtain the predicted concentration Cpred(t + △t) after a period of time △t. Then, combined with the model predicted odor concentration Cmodel(x, y, t), dynamically correct the predicted concentration Cpred(t + △t) to obtain the predicted concentration Cpred(x, y, t + △t) at the position (x, y) at the future time t + △t, and integrate multiple time points of future predictions to form the odor concentration prediction result vector CP;

[0117] The odor concentration prediction result vector CP is specifically CP = {Cpred(x, y, t + 1), Cpred(x, y, t + 2), ……, Cpred(x, y, t + △t)}

[0118] The predicted concentration Cpred(x, y, t + Δt) at the future time t + Δt at the position (x, y) is obtained through the following dynamic correction formula:

[0119] ;

[0120] In the formula, λ represents the model weight coefficient, and its value range is 0 < λ < 1. Specifically, it is used to control the influence of the model to predict the odor concentration Cmodel(x, y, t). represents the odor concentration at the position (x, y) at the current time t. m represents the maximum time step of backtracking. γ(k) represents the weight coefficient of the backtracking time k, which is specifically used to adjust the influence of different historical data. Cobs(x, y, t - k) represents the odor concentration at the position (x, y) at the backtracking time t - k. represents analyzing the influence of historical observation data on future predictions;

[0121] In this embodiment, through a dynamic odor diffusion prediction model established based on the environmental state feature vector F and the aerodynamic principle, the diffusion process of odor in the air is accurately simulated, and the Gaussian plume diffusion model is used to calculate the set YCmodel of model predicted concentrations at different time points and different spatial positions, significantly improving the understanding and prediction accuracy of odor diffusion behavior. Compared with the traditional prediction method that only relies on single-point monitoring and static data modeling, the calculation of the predicted concentration Cpred(x, y, t + Δt) at a specific future time t + Δt and a specific position (x, y) is realized. At the same time, by extracting the historical environmental state feature vector F to form the historical odor concentration dataset HC and combining it with the autoregressive integrated moving average model, this method not only predicts odor diffusion based on a mathematical model, but also can dynamically fuse historical monitoring data to form the odor concentration prediction result vector Cpred, avoiding the problem of cumulative prediction errors caused by environmental factor fluctuations in the traditional method. In addition, by adjusting the model predicted concentration Cmodel(x, y, t) through the dynamic correction formula, the predicted value can better match the actual situation, breaking through the problem that odor concentration prediction in the prior art cannot adapt to different working conditions. Therefore, this method not only significantly improves the prediction accuracy of odor concentration changes in the sludge treatment process, but also maintains strong adaptability and robustness under complex environmental conditions, providing more scientific data support for the optimization of additive dosing, thereby realizing more accurate intelligent control of sludge treatment.

[0122] Example 4

[0123] This embodiment is an explanatory description carried out in Example 1. Please refer to Figure 1 , specifically: The S3 includes S31;

[0124] S31. By extracting the historical data of auxiliary agent dosing records, marking it as the historical main auxiliary agent dosing record HD, and calculating it with the obtained odor concentration prediction result vector CP, the simulated concentration change vector L and the decision vector D of the auxiliary agent dosing amount are obtained;

[0125] The simulated concentration change vector L is obtained through the following calculation formula:

[0126] ;

[0127] In the formula, L(t + △t) represents the odor concentration change vector at the future time t + △t, and △t represents the future time step;

[0128] The decision vector D is obtained through the following calculation formula:

[0129] ;

[0130] In the formula, D(t) represents the decision vector at the current time t, and d1, d2, and d3 represent adjustment coefficients. Specifically, d1 represents the adjustment coefficient of the simulated concentration change vector L(t + △t) at the future time t + △t, d2 represents the adjustment coefficient of the predicted concentration Cpred(x, y, t + △t) at the position (x, y) at the future time t + △t, and d3 represents the adjustment coefficient of the wind speed W.

[0131] The said S4 includes S41;

[0132] S41. After applying the decision vector D, the odor concentration state in the air will change. Based on the odor concentration prediction result vector CP and the decision vector D, simulation is carried out to obtain the adjusted concentration vector Cafter, the adjusted odor concentration change vector Gqs, and the auxiliary agent removal rate vector R;

[0133] The adjusted concentration vector Cafter is obtained through the following calculation formula:

[0134] ;

[0135] In the formula, Cafter(t + △t) represents the adjusted concentration vector at the future time t + △t, D(t + △t) represents the applied decision vector at the future time t + △t, represents the auxiliary agent removal efficiency coefficient, specifically representing the odor concentration that can be reduced by 1L of auxiliary agent, and at the same time used to describe the effect of the auxiliary agent on the pollutants in the odor. Its function is to measure how much odor concentration can be removed by a unit volume of auxiliary agent;

[0136] The adjusted odor concentration change vector Gqs is obtained through the following calculation formula:

[0137] ;

[0138] Wherein, Gqs(t + △t) represents the adjusted odor concentration change vector at the future time t + △t;

[0139] The auxiliary agent removal rate vector R is obtained through the following calculation formula:

[0140] ;

[0141] Wherein, R(t + △t) represents the auxiliary agent removal rate vector at the future time t + △t.

[0142] In this embodiment, by extracting the historical main auxiliary agent dosing record HD and combining it with the odor concentration prediction result vector Cpred, the simulated concentration change vector L and the decision vector D of the auxiliary agent dosing amount at the future time t + △t are calculated, realizing the intelligent decision optimization of the auxiliary agent dosing strategy. Compared with the traditional method that solely relies on the real-time odor concentration feedback to dose the auxiliary agent, this method can combine historical data with prediction trends, ensuring that the dosing decision not only considers the future odor concentration change trend but also can dynamically adapt to environmental parameters such as the wind speed W, thereby avoiding the situation of over-dosing or under-dosing. Further, by simulating the change of the odor concentration state in the air after applying the decision vector D, this method can obtain the adjusted concentration vector Cafter, the adjusted odor concentration change vector Gqs, and the auxiliary agent removal rate vector R, providing a quantitative basis for the effect evaluation after the auxiliary agent is dosed. Especially the introduction of the auxiliary agent removal rate vector R enables the auxiliary agent dosing strategy to not only focus on the degree of odor concentration reduction but also measure the deodorization efficiency of the unit auxiliary agent, thereby optimizing the usage amount of the auxiliary agent and avoiding resource waste. In addition, by calculating the future odor diffusion trend through the adjusted odor concentration change vector Gqs, the system can predictively adjust the dosing strategy, breaking through the limitation of the traditional method that can only rely on passive feedback to adjust the dosing amount. Therefore, this method not only realizes the intelligent dosing decision based on prediction but also can dynamically balance among the auxiliary agent utilization rate, deodorization effect, and environmental adaptability, making the automatic control of the sludge treatment system more accurate, efficient, and energy-saving.

[0143] Example 5

[0144] This embodiment is an explanatory description based on Example 4. Please refer to Figure 1 , specifically: The S5 includes S51;

[0145] S51. Based on the obtained adjusted concentration vector Cafter, decision vector D, and odor concentration change vector Gqs, perform a comprehensive evaluation to obtain the benefit index E, and adjust the dosing amount of the auxiliary agent according to the benefit index E to obtain the execution decision vector D1, and adaptively adjust the sludge treatment process according to the execution decision vector D1, and at the same time set boundary constraints for the execution decision vector D1;

[0146] Among them, the execution decision vector D1 adaptively regulates the sludge treatment process to achieve adaptive control by controlling the operation of the auxiliary agent dosing pump and the dosing device; at the same time, during the sludge conditioning stage, by controlling the intensity and time of the stirring device, the auxiliary agent and the sludge are fully mixed to improve the deodorization efficiency;

[0147] The benefit index E is obtained through the following calculation formula:

[0148] ;

[0149] In the formula, Ctarget represents the preset target odor concentration, and e1, e2, and e3 represent the equity weight coefficients. Specifically, e1 controls the impact of odor removal effect on the benefit, e2 controls the impact of the dosing amount of the auxiliary agent on the benefit, and e3 controls the smoothness of the change in odor concentration;

[0150] By obtaining the benefit index E, the odor concentration is accurately controlled to be close to the target odor concentration Ctarget. At the same time, by calculating the deviation between the current odor concentration Cafte and the target odor concentration Ctarget, it is ensured that the dosing strategy can effectively reduce the odor concentration and prevent resource waste caused by excessive deodorization;

[0151] The application scenarios include:

[0152] During the sludge conditioning stage (adjusting the dosing amount of the auxiliary agent and optimizing the mixing efficiency):

[0153] In the sludge conditioning tank, by calculating the benefit index E to evaluate the current mixing effect of the auxiliary agent, it is ensured that the odor control is effective and the auxiliary agent is not wasted, and then the stirring time or the concentration of the auxiliary agent is adjusted;

[0154] During the sludge dewatering stage (optimizing the ratio of the auxiliary agent to the dehydrating agent):

[0155] By calculating the benefit index E, the dosing amount of the auxiliary agent is controlled to ensure the minimization of the release of volatile gases from the sludge after dewatering, and at the same time, the waste of chemical agents caused by excessive dosing is reduced;

[0156] The execution decision vector D1 is obtained through the following calculation formula:

[0157] ;

[0158] In the formula, D1(t + △t) represents the execution decision vector at the future time t + △t, represents the step size factor, which is specifically used to control the adjustment speed, represents the gradient of the benefit index with respect to the dosing amount of the auxiliary agent, which specifically represents adjusting the dosing amount of the auxiliary agent to minimize the benefit index E;

[0159] The following are the boundary constraints:

[0160] Dmin ≤ D1(t + △t) ≤ Dmax, where Dmin represents the minimum dosage of the additive, used to prevent insufficient dosage, and Dmax represents the maximum dosage of the additive, used to prevent waste caused by excessive dosage.

[0161] The said S6 includes S61;

[0162] S61, record the actual odor concentration Cfinal after executing the decision vector D1. The actual odor concentration Cfinal is obtained by extracting the odor concentration Cobs(x, y) at the position (x, y), calculate the deviation △C from the preset target odor concentration, and compare the deviation △C with the preset error threshold V to trigger the iterative optimization mechanism;

[0163] The said deviation △C is obtained through the calculation formula. In the formula, △C(t) represents the deviation at the current time t;

[0164] The triggering method of the said iterative optimization mechanism is as follows:

[0165] When the deviation △C(t) at the current time t > the error threshold V, it means triggering the iterative optimization mechanism, and execute the adjustment of the additive dosage strategy, including adjusting the proportion to execute the decision vector D1 to execute the adjustment of the additive dosage strategy;

[0166] When the deviation △C(t) at the current time t ≤ the error threshold V, it means not triggering the iterative optimization mechanism.

[0167] In this embodiment, the global optimization of the auxiliary agent dosing strategy is achieved by calculating the benefit index E based on the adjusted concentration vector Cafter, the decision vector D, and the adjusted odor concentration change vector Gqs. Different from the traditional method that only focuses on a single index for dosing adjustment, this method quantifies factors such as the odor removal effect, the cost of auxiliary agent dosing, and the smoothness of odor concentration change through the benefit index E, and combines the boundary constraints of the execution decision vector D1, so that the dosing amount of the auxiliary agent not only meets the requirement of odor concentration compliance, but also minimizes the usage amount of the auxiliary agent and reduces resource waste. At the same time, this method calculates the actual odor concentration Cfinal after executing the decision vector D1, and calculates the odor concentration deviation △C based on the preset target odor concentration Ctarget, so that the system can continuously monitor the effectiveness of the dosing decision. In particular, an error threshold V is introduced as the trigger condition for the iterative optimization mechanism to ensure that adjustments are only made when the dosing strategy deviation exceeds the acceptable range, preventing frequent adjustments from causing system instability or mis-dosing. In addition, the optimized auxiliary agent dosing strategy adjusts the execution decision vector D1 proportionally, making the dosing optimization process smoother and avoiding system fluctuations caused by abrupt adjustments in traditional control strategies. Therefore, this method can not only dynamically optimize the auxiliary agent dosing strategy, improve the accuracy and stability of the dosing decision, but also adapt to different environmental conditions, effectively improve the long-term operation efficiency of the sludge treatment system, and ensure that the system always maintains the optimal balance between deodorization and energy conservation and consumption reduction.

[0168] Example 6

[0169] An automatic control system for sludge treatment, please refer to Figure 2 , specifically: it includes an information collection module, a trend analysis module, a trend decision module, a simulation prediction module, an optimization module, and an iterative trigger module;

[0170] The information collection module collects environmental information and odor concentration information during the sludge treatment process in real time through integrated sensors, and forms an environmental state feature vector F;

[0171] The trend analysis module constructs a dynamic odor diffusion prediction model based on the obtained environmental state feature vector F, calculates the odor concentration change trend at different time points, and obtains an odor concentration prediction result vector CP;

[0172] The trend decision module calculates based on the obtained odor concentration prediction result vector CP to obtain a decision vector D for the dosing amount of the auxiliary agent;

[0173] The simulation prediction module simulates the odor concentration change after applying the decision vector D, and obtains an odor concentration change vector Gqs and an adjusted concentration vector Cafter calculated by odor concentration simulation;

[0174] The optimization module conducts a comprehensive evaluation based on the obtained adjusted concentration vector Cafter, decision vector D, and odor concentration change vector Gqs, obtains the benefit index E, and adjusts the dosage of the regulating agent according to the benefit index E to obtain the execution decision vector D1, and adaptively regulates the sludge treatment process according to the execution decision vector D1;

[0175] The iterative trigger module records the actual odor concentration Cfinal after the execution decision vector D1, calculates the deviation △C from the preset target odor concentration, and triggers the iterative optimization mechanism according to the deviation △C.

[0176] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A sludge treatment automation control method, characterized in that: The following steps are involved: S1, collect environmental information and odor concentration information in the sludge treatment process in real time through integrated sensors to form an environmental state feature vector F; S2. Based on the acquired environmental state feature vector F, a dynamic odor diffusion prediction model is constructed to calculate the odor concentration change trend at different time points and obtain the odor concentration prediction result vector CP; S3, performing calculation based on the obtained odor concentration prediction result vector CP to obtain the decision vector D of the additive dosage; S4, simulating the change of odor concentration after applying the decision vector D, and obtaining the odor concentration change vector Gqs and the adjusted concentration vector Cafter calculated by the odor concentration simulation; S5. Perform a comprehensive evaluation based on the obtained adjusted concentration vector Cafter, decision vector D and odor concentration change vector Gqs, obtain the benefit index E, and adjust the dosage of the additive according to the benefit index E, obtain the execution decision vector D1, and adaptively adjust the sludge treatment process according to the execution decision vector D1; S6. Record the actual odor concentration Cfinal after executing the decision vector D1, calculate the deviation △C from the preset target odor concentration, and trigger the iterative optimization mechanism according to the deviation △C.

2. The sludge treatment automation control method according to claim 1, characterized in that: Said S1 includes S11 and S12; S11, collecting environmental information that affects odor diffusion during sludge treatment in real time through integrated multiple sensors, and forming an environmental information vector Fel, wherein the environmental information vector Fel includes temperature T, humidity H, wind speed W, wind direction θ, air pressure P and odor source estimated intensity S; Among them, the sludge treatment process includes sludge storage process, conditioning process, dehydration process and stacking process; Among them, the various sensors include temperature sensors, humidity sensors, ultrasonic anemometers, wind direction sensors, and air pressure sensors; The odor source estimation intensity S is obtained by estimating the volatilization of organic matter in the sludge, and after measuring the sludge organic matter content Corg, combining the temperature T and the humidity H using an empirical estimation formula to obtain the odor source estimation intensity S; The estimated intensity S of the odor source is obtained by the following empirical estimation formula: ； In the formula, β1, β2 and β3 represent empirical coefficients. Specifically, β1 represents the empirical coefficient of sludge organic matter content Corg, β2 represents the empirical coefficient of humidity H, and β3 represents the empirical coefficient of temperature T.

3. The sludge treatment automation control method according to claim 2 is characterized in that: S12, collect odor concentration information in real time by odor sensors installed at multiple locations in the treatment plant, obtain the average odor concentration Cobs of the treatment plant after processing the odor concentration information obtained at multiple locations, and combine it with the environmental information vector Fel to form an environmental state feature vector F, and mark the timestamp information for fixed-period storage as historical data; The environmental state feature vector F is specifically F={Fel, Cobs}; The odor sensor includes using a gas detection sensor to measure the odor concentration in the air, and obtaining the odor concentration Cobs (x, y) at the location (x, y) of the treatment plant.

4. The sludge treatment automation control method according to claim 3 is characterized in that: The S2 includes S21 and S22; S21, based on the acquired environmental state characteristic vector F and the principle of aerodynamics, establishing a dynamic odor diffusion prediction model, wherein the dynamic odor diffusion prediction model is established by using a Gaussian plume diffusion model to simulate the diffusion process of the odor in the air; The dynamic odor diffusion prediction model established is used to simulate the diffusion of odor in the air, which is affected by wind speed W, wind direction θ and estimated odor source intensity S. The model-predicted odor concentration Cmodel (x, y, t) at the position (x, y) at different times t is calculated to obtain the model-predicted concentration set YCmodel at multiple monitoring points. The predicted odor concentration Cmodel (x, y, t) is obtained by the following calculation formula: ； In the formula, π is a mathematical constant with a value of 3.14159. and It represents the diffusion coefficient of odor, specifically the degree of diffusion in the horizontal y direction and the vertical z direction. exp represents the exponential function. G represents the pollutant release height, which is set by the geometric characteristics of the sludge tank.

5. A sludge treatment automation control method according to claim 4, characterized in that: S22, extract the historically stored environmental state feature vector F, mark it as the historical odor concentration data set Hc, and perform short-term prediction with the environmental information vector Fel and the model predicted concentration set YCmodel, the short-term prediction is performed by using the autoregressive integral sliding average model, obtain the predicted concentration Cpred(t+△t) after a period of time △t, and then dynamically correct the predicted concentration Cpred(t+△t) in combination with the model predicted odor concentration Cmodel (x, y, t), obtain the predicted concentration Cpred(x, y, t+△t) at the position (x, y) at the future time t+△t, and integrate multiple time points of future prediction to form the odor concentration prediction result vector CP; The odor concentration prediction result vector CP is specifically CP={Cpred(x, y, t+1), Cpred(x, y, t+2), ..., Cpred(x, y, t+△t)} The predicted concentration Cpred(x, y, t+△t) at the position (x, y) at the future time t+△t is obtained by the following dynamic correction formula: ； In the formula, λ represents the model weight coefficient, and its value range is 0<λ<1. represents the odor concentration at the position (x, y) at the current time t, m represents the maximum time step of the backtracking, γ(k) represents the weight coefficient of the backtracking time k, and Cobs(x, y, tk) represents the odor concentration at the position (x, y) at the backtracking time tk.

6. A sludge treatment automation control method according to claim 5, characterized in that: The S3 includes S31; S31, extracting historical additive dosing record data, marking it as historical main additive dosing record HD, and calculating it with the obtained odor concentration prediction result vector CP to obtain the simulated concentration change vector L and the additive dosing amount decision vector D; The simulated concentration change vector L is obtained by the following calculation formula: ； In the formula, L(t+△t) represents the odor concentration change vector at the future time t+△t, and △t represents the future time step; The decision vector D is obtained by the following calculation formula: ； Where D(t) represents the decision vector at the current time t, d1, d2 and d3 represent the adjustment coefficients. Specifically, d1 represents the adjustment coefficient of the simulated concentration change vector L(t+△t) at the future time t+△t, d2 represents the adjustment coefficient of the predicted concentration Cpred(x, y, t+△t) at the position (x, y) at the future time t+△t, and d3 represents the adjustment coefficient of the wind speed W.

7. The sludge treatment automation control method according to claim 6, characterized in that: The S4 includes S41; S41, after applying the decision vector D, the odor concentration state in the air will change, and simulation is performed based on the odor concentration prediction result vector CP and the decision vector D to obtain the adjusted concentration vector Cafter, the adjusted odor concentration change vector Gqs and the additive removal rate vector R; The adjusted concentration vector Cafter is obtained by the following calculation formula: ; Where Cafter(t+△t) represents the adjusted concentration vector at future time t+△t, and D(t+△t) represents the decision vector at future time t+△t. It represents the additive removal efficiency coefficient; The adjusted odor concentration change vector Gqs is obtained by the following calculation formula: ； In the formula, Gqs (t+△t) represents the adjusted odor concentration change vector at future time t+△t; The additive removal rate vector R is obtained by the following calculation formula: ； Where R(t+△t) represents the additive removal rate vector at the future time t+△t.

8. The sludge treatment automation control method according to claim 7, characterized in that: The S5 includes S51; S51, based on the obtained adjusted concentration vector Cafter, decision vector D and odor concentration change vector Gqs, a comprehensive evaluation is performed to obtain the benefit index E, and the dosage of the additive is adjusted according to the benefit index E, and the execution decision vector D1 is obtained, and the sludge treatment process is adaptively adjusted according to the execution decision vector D1, and a boundary constraint is set for the execution decision vector D1; Among them, the execution decision vector D1 adaptively regulates the sludge treatment process to achieve adaptive control by controlling the operation of the additive dosing pump and the dosing device; The benefit index E is obtained by the following calculation formula: ； In the formula, Ctarget represents the preset target odor concentration, e1, e2 and e3 represent the equity weight coefficients; The execution decision vector D1 is obtained by the following calculation formula: ； Where D1(t+△t) represents the execution decision vector at future time t+△t. Represents the step size factor, which is used to control the adjustment speed. It represents the gradient of benefit index to additive dosage, specifically, it means adjusting additive dosage to minimize benefit index E; The boundary constraints are as follows: Dmin≤D1(t+△t)≤Dmax, where Dmin represents the minimum additive dosage and Dmax represents the maximum additive dosage.

9. The sludge treatment automation control method according to claim 8, characterized in that: The S6 includes S61; S61, recording the actual odor concentration Cfinal after executing the decision vector D1, the actual odor concentration Cfinal is obtained by extracting the odor concentration Cobs(x, y) at the position (x, y), and calculating the deviation △C from the preset target odor concentration, and comparing the deviation △C with the preset error threshold V to trigger the iterative optimization mechanism; The deviation △C is The calculation formula is obtained, where △C (t) represents the deviation at the current time t; The iterative optimization mechanism is triggered as follows: When the deviation △C(t) at the current time t is greater than the error threshold V, it indicates that the iterative optimization mechanism is triggered and the additive dosing strategy adjustment is executed, including proportionally adjusting the execution decision vector D1 to execute the additive dosing strategy adjustment; When the deviation △C(t) at the current time t ≤ the error threshold V, it means that the iterative optimization mechanism is not triggered.

10. A sludge treatment automation control system, applied to a sludge treatment automation control method according to any one of claims 1 to 9, characterized in that: It includes information collection module, trend analysis module, trend decision module, simulation prediction module, optimization module and iteration trigger module; The information collection module collects environmental information and odor concentration information in the sludge treatment process in real time through integrated sensors to form an environmental state feature vector F; The trend analysis module constructs a dynamic odor diffusion prediction model based on the acquired environmental state feature vector F, calculates the odor concentration change trend at different time points, and obtains the odor concentration prediction result vector CP; The trend decision module performs calculation based on the obtained odor concentration prediction result vector CP to obtain the decision vector D of the additive dosage; The simulation prediction module simulates the change of odor concentration after applying the decision vector D, and obtains the odor concentration change vector Gqs and the adjusted concentration vector Cafter calculated by the odor concentration simulation; The optimization module performs a comprehensive evaluation based on the obtained adjusted concentration vector Cafter, the decision vector D and the odor concentration change vector Gqs, obtains the benefit index E, and adjusts the dosage of the additive according to the benefit index E, obtains the execution decision vector D1, and adaptively adjusts the sludge treatment process according to the execution decision vector D1; The iterative trigger module records the actual odor concentration Cfinal after executing the decision vector D1, calculates the deviation △C from the preset target odor concentration, and triggers the iterative optimization mechanism according to the deviation △C.

Images