A sewage treatment process for the pulp and paper industry

Abstract

The present invention discloses a sewage treatment process for the pulp and paper industry, which relates to the technical field of sewage treatment. The present invention adopts a dynamic adjustment algorithm and a PID control mechanism to monitor the suspended solid concentration, flow rate and water temperature in real time. The chemical dosage can be automatically adjusted to stabilize the concentration near the target value, reducing the waste of chemical agents. At the same time, the stability and effect of sewage treatment are improved. At the same time, a reinforcement learning RL algorithm is introduced to adjust the aeration intensity and time in real time, control the ammonia nitrogen concentration and remove the organic load. The RL algorithm can continuously optimize the aeration strategy according to the changes in ammonia nitrogen concentration and organic load, so as to maximize the removal efficiency, avoid over-aeration and unnecessary energy consumption, and adopt predictive control based on the random forest algorithm to analyze the physical properties of sludge in real time, including moisture content, solid content and rheological properties, and dynamically adjust the concentration stirring intensity and concentration time.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage treatment, and particularly to a sewage treatment process for the pulp and paper industry. Background Art

[0002] In the sewage treatment of the pulp and paper industry, the dewatering treatment of sludge and the removal of ammonia nitrogen are key links. The sludge contains a large amount of organic matter, cellulose, and other difficult-to-degrade substances, making the sludge have a high water content, increasing the treatment difficulty. In addition, the ammonia nitrogen concentration in the pulp and paper wastewater fluctuates greatly, affecting the compliance of the effluent quality.

[0003] Most traditional sludge treatment methods adopt methods such as gravity thickening, mechanical dewatering, and chemical agent-assisted dewatering. In the thickening stage, most of the water is removed through gravity thickening or air flotation thickening to initially reduce the sludge volume. Subsequently, mechanical dewatering equipment such as belt filter presses, centrifuges, or screw filter presses is used to further remove water to achieve mud-water separation. Some treatment schemes also add chemical agents, such as coagulants or flocculants, to help strengthen the sedimentation and aggregation of solid particles in the sludge, thereby improving the dewatering efficiency and alleviating the problem of high sludge water content to a certain extent.

[0004] However, the operation cost of mechanical dewatering equipment is high, and the maintenance frequency is high. For the high fiber content in pulp and paper sludge, the equipment is easily damaged and the maintenance cost is high. Moreover, the effects of gravity thickening and chemical agent-assisted dewatering are also affected by the fluctuation of sludge composition, resulting in unstable treatment effects. In addition, the frequent replenishment of chemical agents not only increases the treatment cost but may also cause secondary pollution. For the removal of ammonia nitrogen, traditional methods mainly rely on the continuous aeration mode in the biological treatment system. However, due to the large fluctuation of ammonia nitrogen concentration, the fixed aeration intensity is difficult to meet the actual needs, resulting in high energy consumption, poor system stability, and unstable treatment effects. Therefore, there is an urgent need for a sewage treatment process for the pulp and paper industry to solve such problems. Summary of the Invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] The present invention provides a sewage treatment process for the pulp and paper industry to solve the problems that the operation cost of mechanical dewatering equipment is high, the maintenance frequency is high, for the high fiber content in pulp and paper sludge, the equipment is easily damaged and the maintenance cost is high, and the effects of gravity thickening and chemical agent-assisted dewatering are also affected by the fluctuation of sludge composition, resulting in unstable treatment effects. In addition, the frequent replenishment of chemical agents not only increases the treatment cost but may also cause secondary pollution.

[0007] To solve the above technical problems, the present invention provides the following technical solutions:

[0008] The present invention provides a sewage treatment process for the pulp and paper industry, which includes:

[0009] Step S1, sewage pretreatment

[0010] Adopt a combined process of air flotation and gravity sedimentation to remove large particle impurities, fibers and suspended solids in the sewage. Set up a water quality sensor to monitor the suspended solid concentration and flow rate of the sewage, and use a dynamic algorithm to adjust the dosage of flocculant and coagulant aid.

[0011] Step S2, staged aeration biological treatment

[0012] Perform biological treatment on the sewage pretreated in Step S1, and use the staged aeration method to remove high-concentration organic matter and ammonia nitrogen.

[0013] Step S3, sludge thickening predictive control

[0014] The sludge after biological treatment enters the thickening stage. Use predictive control based on the random forest algorithm to perform real-time analysis on the moisture content, solid content and rheological properties of the sludge, and intelligently adjust the thickening stirring intensity and thickening time.

[0015] Step S4, sludge dewatering adjustment control

[0016] The sludge thickened in Step S3 enters the dewatering link. Real-time monitor the moisture content and solid content of the sludge, and adjust the pressure, rotation speed and chemical dosage of the dewatering equipment in combination with the simulated annealing SA algorithm.

[0017] Step S5, sludge reduction and resource utilization

[0018] Perform reduction and resource treatment on the dewatered sludge, and recommend corresponding resource treatment methods according to the organic content, calorific value and heavy metal concentration of the sludge.

[0019] Furthermore, in Step S1, sewage monitoring and data collection are carried out. Specifically:

[0020] Set up a water quality sensor to monitor the suspended solid concentration, flow rate and water temperature of the sewage in real time, and construct a dynamic adjustment model for the chemical dosage to calculate the chemical dosage:

[0021] Among them, F chem (t) represents the chemical dosage at the current moment t, C sus (t) represents the suspended solid concentration at the current moment t, Q in (t) represents the sewage flow rate at the current moment t, T(t) represents the water temperature at the current moment t, C targetDenote the target suspended solid concentration as the preset maximum allowable concentration. α, β, and γ represent adaptive adjustment parameters, which are adjusted according to the actual requirements during the sewage treatment process. It is the integral term of the historical error, representing the cumulative error of the suspended solid concentration from the start time to the current time t.

[0022] Furthermore, in step S1, dynamically adjust the chemical dosage:

[0023] Based on the sensor, obtain the suspended solid concentration, flow rate, and water temperature at the current time, set the target concentration, and calculate the historical error term When the suspended solid concentration fluctuates, it gradually approaches the target value to avoid large fluctuations; according to the real-time data and the historical error term, use the PID control algorithm to dynamically adjust the parameters (α, β, γ) of the chemical dosage.

[0024] Furthermore, in step S2, based on the real-time data obtained by the water quality sensor in step S1, combined with the reinforcement learning RL algorithm, continuously adjust the aeration intensity and time based on the changes in ammonia nitrogen concentration and organic load.

[0025] Furthermore, in step S2, the way of segmented aeration biological treatment is:

[0026] Based on the real-time water quality data monitored in step S1, use it as the input of the reinforcement learning algorithm. Divide the aeration process into pre-aeration, main aeration, and post-aeration. Aerate for different ammonia nitrogen concentrations and organic loads in each stage. The aeration intensity and time in each stage are defined as:

[0027] Among them, P aer (t) represents the aeration intensity at the current time t, represents the ammonia nitrogen concentration at the current time t, BOD(t) represents the organic load at the current time t, Q in (t) represents the sewage flow rate at the current time t, T(t) represents the water temperature at the current time t, α, β, γ represent empirical adjustment parameters, which control the influence of various factors on the aeration intensity and need to be adjusted according to the actual situation.

[0028] Optimize the aeration intensity P aer (t) and the aeration time T aer (t), adjust the ammonia nitrogen removal effect. The objective function of the reinforcement learning model is:

[0029] Among them, R total (t) represents the total reward value at the current time t, represents the ammonia nitrogen concentration at the current time t, BOD(t) represents the organic load at the current time t, C targetDenotes the target ammonia nitrogen concentration, which is the maximum allowable concentration, BOD tar get Denotes the target organic load, P aer (t) denotes the aeration intensity at the current time t, T aer (t) denotes the aeration time at the current time t. λ1, λ2, and λ3 denote weight coefficients. Based on the RL algorithm, the aeration intensity and time are adjusted according to the environmental state to ultimately maximize the reward R total (t), that is, minimize the degree of deviation of ammonia nitrogen and organic load from the target values.

[0030] Furthermore, in step S2, the method of segmented aeration biological treatment further includes:

[0031] Update the aeration strategy: Among them, Q(s t ,a t ) represents the Q value of the current state s t and action a t , representing the expected total reward for executing action a t under state s t , R total (t) represents the total reward value at the current time t. α represents the learning rate, which controls the response speed of the algorithm to new information. γ represents the discount factor, which represents the weight of future rewards. s t represents the system state at the current time t, including water quality data, a t represents the action taken at the current time t, that is, the adjusted aeration intensity and time. During the continuous interaction with the system environment, update Q(s t ,a t ) value to optimize the aeration operation;

[0032] According to the optimization of the RL strategy, the final aeration control formula is:

[0033] Among them, ΔQ(s t ,a t ) represents the adjustment amount calculated by the RL algorithm, representing the adjustment of the aeration intensity relative to the previous moment for the current decision.

[0034] Furthermore, in step S3, sludge thickening and predictive control are performed:

[0035] Build a random forest model to predict the sludge characteristics: Y(t) = f(X(t), θ) + ∈(t), where Y(t) represents the predicted results of the sludge physical characteristics at the current time t, including moisture content, solid content, and rheological characteristics, X(t) represents the set of operating parameters at the current time t, f(X(t), θ) represents the prediction function based on the operating parameters and the random forest algorithm, the learned model function, θ represents the parameters of the model, and ∈(t) represents the model error, that is, the prediction error;

[0036] Obtain the sludge physical characteristics and operating parameters in real time, use the random forest algorithm to predict the moisture content, solid content, and rheological characteristics in real time, and then based on the prediction results, adjust the concentration stirring intensity and concentration time. The control formula is: P stir (t) = α1·W(t) + α2·S(t) + α3·R(t) + β·Q sludge (t) + γ·T conc (t), T conc (t) = δ1·W(t) + δ2·S(t) + δ3·R(t) + ∈·Q sludge (t), where P stir (t) represents the stirring intensity at the current time t, T conc (t) represents the concentration time at the current time t, which is determined by the sludge physical characteristics and flow rate. W(t) represents the moisture content at the current time t, S(t) represents the solid content at the current time t, R(t) represents the rheological characteristics at the current time t, Q sludge (t) represents the sludge flow rate at the current time t, T conc (t) represents the concentration time. α1, α2, α3, β, γ, δ1, δ2, δ3 represent the adjustment coefficients of the model, which are used to adjust the influence of each parameter on the stirring intensity and concentration time.

[0037] Furthermore, in step S3, according to the sludge physical characteristics predicted in real time, adjust the stirring intensity and concentration time:

[0038] J(t) = λ1·(W tar get - W(t)) 2 + λ2·(S tar get - S(t)) 2 + λ3·(R tar get - R(t)) 2 + λ4·P stir (t) + λ5·T conc (t), where J(t) represents the optimization objective function at the current time t, that is, the deviation between the target state and the actual state and the operating cost, W tar get , S tar get , R tar getDenote the target sludge characteristic value, and let λ1, λ2, λ3, λ4, λ5 represent the weight coefficients, balance the relationship between sludge characteristic optimization and operating cost, minimize the objective function J(t), and optimize the stirring intensity and thickening time;

[0039] According to the results in the optimization process, the final thickening process control formula is:

[0040] P stir (t) = α1·W(t) + α2·S(t) + α3·R(t) + β·Q sludge (t) + γ·T conc (t),

[0041] T conc (t) = δ1·W(t) + δ2·S(t) + δ3·R(t) + ∈·Q sludge (t).

[0042] Furthermore, in step S4, based on the real-time monitored sludge moisture content and solid content, the simulated annealing SA algorithm is used to adjust the parameters of the dehydration equipment:

[0043] Define the optimization objective function J(t) to describe the sludge dehydration effect. The goal is to minimize the moisture content of the sludge. The form of the objective function is: J(t) = w1·H(t) + w2·E(t) + w3·T(t), where H(t) represents the sludge moisture content at time point t, E(t) represents the energy consumption of the dehydration equipment at time point t, T(t) represents the total duration of the dehydration process, and w1, w2, w3 represent the weight coefficients;

[0044] Through the simulated annealing process, search for the optimal solution in the global solution space. The steps include:

[0045] Set the initial solutions P(0), N(0), C(0), and calculate the initial objective function value J(0). Based on the current solution, randomly adjust P(t), N(t), and C(t) to generate new solutions P′(t), N′(t), C′(t). According to the new operating parameters, calculate the new objective function value J′(t);

[0046] Decide whether to accept the new solution according to the change of the objective function value. The acceptance probability of the new solution is P accept :

[0047] where J′(t) represents the new objective function value, J(t) represents the current objective function value, T(t) represents the current temperature, which gradually decreases with the iteration. The "temperature" in simulated annealing represents the amplitude of exploring the new solution space;

[0048] According to P acceptThe value determines whether to accept the new solution. If J′(t) is smaller, or a worse solution is accepted with a certain probability, then update P(t), N(t), and C(t) to P′(t), N′(t), and C′(t);

[0049] With each iteration, the temperature T(t) gradually decreases, T(t) = T0·α t , where T0 represents the initial temperature, α represents the attenuation coefficient, and here t represents the current iteration number; when the temperature T(t) is less than the set threshold or the change in the objective function value is less than a certain threshold, terminate the algorithm and output the final parameter combinations P(t), N(t), and C(t).

[0050] Furthermore, in step S5, sludge reduction is carried out:

[0051] Define the comprehensive optimization objective function G(t). The goal is to minimize the environmental risk during the treatment process and maximize the resource recovery efficiency. The objective function is:

[0052] G(t) = w1·O(t) + w2·H(t) + w3·R(t) + w4·M(t), where O(t) represents the organic matter content of the sludge at time t, H(t) represents the calorific value of the sludge, R(t) represents the resource recovery rate, M(t) represents the heavy metal concentration, and w1, w2, w3, w4 represent the weight coefficients;

[0053] The goal is to minimize G(t). During the sludge resource utilization process, the recovery of organic matter and energy is given priority, and at the same time, the risk of heavy metals is controlled;

[0054] Estimate the relationship between the calorific value of the sludge and the recovery of organic matter, H(t) = α·O(t) + β, where α represents the correlation coefficient between the calorific value and the organic matter content, and β represents the basic calorific value of the sludge. Adjust the recovery of organic matter to optimize the calorific value increase;

[0055] Take the heavy metal concentration as an optimization constraint condition, and define the heavy metal risk control function M(t) as a limiting term in the treatment process. The heavy metal concentration does not exceed the environmental protection standard, that is, M(t) ≤ M limit where M(t) represents the heavy metal concentration of the sludge at time t, and M limit represents the environmental protection concentration standard of heavy metals;

[0056] Define the resource recovery rate: R(t) = γ·H(t) - δ·M(t), where γ represents the calorific value recovery efficiency coefficient and δ represents the heavy metal influence coefficient;

[0057] Based on the optimization objective function G(t) and the constraints, a multi-objective optimization algorithm is used to evaluate and select each sludge resource utilization plan. During each iteration process, the parameters of the treatment plan are adjusted. Finally, according to the above optimization process, the best resource utilization plan is selected.

[0058] The beneficial effects of the present invention are as follows:

[0059] In the present invention, a dynamic adjustment algorithm and a PID control mechanism are adopted to monitor the suspended solid concentration, flow rate and water temperature in real time. The chemical dosage can be automatically adjusted to stabilize the concentration near the target value, effectively avoiding the phenomenon of frequent adjustment or excessive dosing of chemicals in the traditional method, reducing the waste of chemical agents, and at the same time improving the stability and effect of sewage treatment.

[0060] In the present invention, a reinforcement learning RL algorithm is introduced to adjust the aeration intensity and time in real time, control the ammonia nitrogen concentration, and remove the organic load. The RL algorithm can continuously optimize the aeration strategy according to the changes in ammonia nitrogen concentration and organic load, so as to maximize the removal efficiency, while avoiding over-aeration and unnecessary energy consumption.

[0061] In the present invention, predictive control based on a random forest algorithm is adopted to analyze the physical properties of sludge in real time, including moisture content, solid content and rheological properties, and dynamically adjust the concentration stirring intensity and concentration time, overcoming the challenges brought by the changes in sludge properties.

[0062] In the present invention, a simulated annealing algorithm SA is introduced to search for the optimal solution in the global solution space and optimize the working parameters of the dewatering equipment, including pressure, rotation speed and chemical dosage, significantly improving the dewatering efficiency. At the same time, a comprehensive optimization objective function is introduced, combined with the weight coefficients of organic matter content, calorific value and heavy metal concentration, and different resource utilization treatment plans are evaluated based on a multi-objective optimization algorithm, effectively balancing the organic matter recovery, calorific value improvement and heavy metal concentration control, reducing waste emissions, and improving resource utilization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0064] Figure 1 It is a schematic diagram of the sewage treatment process flow for the pulp and paper industry of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0065] In order to make the above objects, features and advantages of the present invention more obvious and understandable, the following will make a detailed description of the specific embodiments of the present invention with reference to the drawings of the specification.

[0066] In the following description, many specific details are set forth in order to provide a thorough understanding of the present invention. However, the present invention may be practiced in other ways than those specifically described herein. Those skilled in the art can make similar extensions without departing from the spirit of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0067] Secondly, as used herein, "an embodiment" or "embodiments" refer to specific features, structures, or characteristics that may be included in at least one implementation of the present invention. The phrase "in an embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments.

[0068] Example 1, referring to Figure 1 , this embodiment provides a sewage treatment process for the pulp and paper industry, including the following steps:

[0069] Step S1, sewage pretreatment,

[0070] Adopt a combined process of air flotation and gravity sedimentation to remove large particle impurities, fibers, and suspended solids in the sewage. Set up a water quality sensor to monitor the suspended solid concentration and flow rate of the sewage, and use a dynamic algorithm to adjust the dosage of flocculants and coagulants;

[0071] In step S1, sewage monitoring and data collection are carried out. Specifically:

[0072] Set up a water quality sensor to continuously monitor the suspended solid concentration, flow rate, and water temperature of the sewage, and construct a dynamic adjustment model for the dosage of chemicals to calculate the dosage of chemicals:

[0073] Among them, F chem (t) represents the dosage of chemicals at the current moment t, C sus (t) represents the suspended solid concentration at the current moment t, Q in (t) represents the sewage flow rate at the current moment t, T(t) represents the water temperature at the current moment t, C target represents the target suspended solid concentration, is the preset maximum allowable concentration, α, β, γ represent adaptive adjustment parameters, and are adjusted according to the actual needs in the sewage treatment process, is the integral term of the historical error, representing the cumulative suspended solid concentration error from the start time to the current moment t;

[0074] In step S1, dynamically adjust the dosage of chemicals:

[0075] Based on the sensor to obtain the suspended solid concentration, flow rate, and water temperature at the current moment, set the target concentration, and calculate the historical error term When the suspended solid concentration fluctuates, it gradually approaches the target value to avoid large fluctuations; according to real-time data and historical error terms, the parameters (α, β, γ) of the chemical dosing amount are dynamically adjusted using the PID control algorithm.

[0076] Specifically, in step S1, the suspended solid concentration, flow rate, and water temperature in the sewage are monitored in real time, and the dosing amounts of the flocculant and coagulant aid are dynamically adjusted. Combining with the PID control algorithm, the suspended solid concentration in the sewage can be stabilized within the target range. Through PID control, large fluctuations in the dosing amount are avoided, ensuring the stability and efficiency of the treatment process.

[0077] Step S2, stepwise aeration biological treatment

[0078] The sewage pretreated in step S1 is subjected to biological treatment, and the stepwise aeration method is used to remove high-concentration organic matter and ammonia nitrogen.

[0079] In step S2, based on the real-time data obtained by the water quality sensor in step S1, combined with the reinforcement learning RL algorithm, the aeration intensity and time are continuously adjusted based on the changes in ammonia nitrogen concentration and organic load.

[0080] In step S2, the method of stepwise aeration biological treatment is as follows:

[0081] Based on the real-time water quality data monitored in step S1, which is used as the input of the reinforcement learning algorithm, the aeration process is divided into pre-aeration, main aeration, and post-aeration. Aeration is carried out for different ammonia nitrogen concentrations and organic loads in each stage. The aeration intensity and time in each stage are defined as:

[0082] Among them, P aer (t) represents the aeration intensity at the current time t, represents the ammonia nitrogen concentration at the current time t, BOD(t) represents the organic load at the current time t, Q in (t) represents the sewage flow rate at the current time t, T(t) represents the water temperature at the current time t, and α, β, γ represent empirical adjustment parameters to control the influence of various factors on the aeration intensity, which need to be adjusted according to the actual situation.

[0083] Optimize the aeration intensity P aer (t) and the aeration time T aer (t) based on real-time feedback to adjust the ammonia nitrogen removal effect. The objective function of the reinforcement learning model is:

[0084] Among them, R total (t) represents the total reward value at the current time t, represents the ammonia nitrogen concentration at the current time t, BOD(t) represents the organic load at the current time t, Ctar get represents the target ammonia nitrogen concentration, which is the maximum allowable concentration, BOD tar get represents the target organic load, P aer I(t) represents the aeration intensity at the current time t, T aer t(t) represents the aeration time at the current time t, and λ1, λ2, λ3 represent the weight coefficients. Based on the RL algorithm, the aeration intensity and time are adjusted according to the environmental state to finally maximize the reward R total I(t), that is, minimize the degree of deviation of ammonia nitrogen and organic load from the target values;

[0085] In step S2, the way of performing segmented aeration biological treatment also includes:

[0086] Update the aeration strategy: Among them, Q(s t ,a t ) represents the Q value of the current state s t and the action a t , representing the expected total reward when performing the action a t under the state s t , R total R(t) represents the total reward value at the current time t, α represents the learning rate, which controls the response speed of the algorithm to new information, and γ represents the discount factor, representing the weight of future rewards, s t represents the system state at the current time t, including water quality data, a t represents the action taken at the current time t, that is, the adjusted aeration intensity and time. During the continuous interaction with the system environment, the Q(s t ,a t ) value is updated to optimize the aeration operation;

[0087] According to the optimization of the RL strategy, the final aeration control formula is:

[0088] Among them, ΔQ(s t ,a t ) represents the adjustment amount calculated by the RL algorithm, representing the adjustment of the aeration intensity relative to the previous moment;

[0089] Specifically, in step S2, the segmented aeration method is adopted and combined with the reinforcement learning algorithm. According to the real-time water quality data, the aeration intensity and aeration time are adjusted, significantly improving the removal efficiency of ammonia nitrogen and organic load in the sewage. The reinforcement learning model can continuously optimize the aeration process according to the changes in ammonia nitrogen concentration and organic load, not only improving the treatment efficiency but also avoiding energy waste caused by excessive aeration; by continuously updating the aeration strategy, the removal effect of ammonia nitrogen and organic load is maximized.

[0090] Step S3, sludge thickening predictive control,

[0091] The biologically treated sludge enters the thickening stage. Predictive control based on the random forest algorithm is used to perform real-time analysis on the moisture content, solid content, and rheological properties of the sludge, and intelligently adjust the thickening stirring intensity and thickening time.

[0092] In step S3, sludge thickening and predictive control are carried out:

[0093] A random forest model is constructed to predict the sludge characteristics: Y(t) = f(X(t), θ) + ∈(t), where Y(t) represents the predicted result of the sludge physical characteristics at the current time t, including moisture content, solid content, and rheological properties, X(t) represents the set of operating parameters at the current time t, f(X(t), θ) represents the prediction function based on the operating parameters and the random forest algorithm, the learned model function, θ represents the parameters of the model, and ∈(t) represents the model error, that is, the prediction error.

[0094] The sludge physical characteristics and operating parameters are obtained in real time. The random forest algorithm is used to predict the moisture content, solid content, and rheological properties in real time. Then, based on the prediction results, the thickening stirring intensity and thickening time are adjusted. The control formula is: P stir (t) = α1·W(t) + α2·S(t) + α3·R(t) + β·Q shudge (t) + γ·T conc (t), T conc (t) = δ1·W(t) + δ2·S(t) + δ3·R(t) + ∈·Q sludge (t), where P stir (t) represents the stirring intensity at the current time t, T conc (t) represents the thickening time at the current time t, which is determined by the sludge physical characteristics and flow rate. W(t) represents the moisture content at the current time t, S(t) represents the solid content at the current time t, R(t) represents the rheological properties at the current time t, Q sludge (t) represents the sludge flow rate at the current time t, T conc (t) represents the thickening time, and α1, α2, α3, β, γ, δ1, δ2, δ3 represent the adjustment coefficients of the model, which are used to adjust the influence of each parameter on the stirring intensity and thickening time.

[0095] In step S3, according to the real-time predicted sludge physical characteristics, the stirring intensity and thickening time are adjusted:

[0096] J(t) = λ1·(W tar get - W(t)) 2 + λ2·(S tar get - S(t)) 2 + λ3·(R tar get - R(t)) 2+λ4·P stir (t)+λ5·T conc (t), where J(t) represents the optimization objective function at the current time t, that is, the deviation between the target state and the actual state and the operation cost, W tar get , S tar get , R tar get represents the target sludge characteristic value, and λ1, λ2, λ3, λ4, λ5 represent the weight coefficients, balancing the relationship between sludge characteristic optimization and operation cost, minimizing the objective function J(t), and optimizing the stirring intensity and concentration time;

[0097] According to the results in the optimization process, the final formula for controlling the concentration process is:

[0098] P stir (t) = α1·W(t) + α2·S(t) + α3·R(t) + β·Q sludge (t) + γ·T conc (t),

[0099] T conc (t) = δ1·W(t) + δ2·S(t) + δ3·R(t) + ∈·Q sludge (t);

[0100] Specifically, in step S3, through predictive control based on the random forest algorithm, the physical characteristics of the sludge are analyzed and predicted in real time. According to the prediction results, the concentration stirring intensity and concentration time are dynamically adjusted, effectively optimizing the sludge concentration process; timely responding to changes in sludge characteristics, reducing the uncertainty in the operation process, and improving the stability and efficiency of sludge concentration.

[0101] Step S4, sludge dewatering adjustment control,

[0102] The sludge concentrated in step S3 enters the dewatering link, and the moisture content and solid content of the sludge are monitored in real time, and the pressure, rotation speed, and chemical dosage of the dewatering equipment are adjusted in combination with the simulated annealing SA algorithm;

[0103] In step S4, based on the moisture content and solid content of the sludge monitored in real time, the simulated annealing SA algorithm is used to adjust the parameters of the dewatering equipment:

[0104] Define the optimization objective function J(t) to describe the sludge dewatering effect. The goal is to minimize the moisture content of the sludge. The form of the objective function is: J(t) = w1·H(t) + w2·E(t) + w3·T(t), where H(t) represents the moisture content of the sludge at time point t, E(t) represents the energy consumption of the dewatering equipment at time point t, T(t) represents the total duration of the dewatering process, and w1, w2, w3 represent the weight coefficients;

[0105] Search for the optimal solution in the global solution space through the simulated annealing process. The steps include:

[0106] Set the initial solutions P(0), N(0), C(0), and calculate the initial objective function value J(0). Based on the current solutions, randomly adjust P(t), N(t), and C(t) to generate new solutions P′(t), N′(t), C′(t), and calculate the new objective function value J′(t) according to the new operating parameters;

[0107] Determine whether to accept the new solution based on the change in the objective function value. The probability of accepting the new solution is P accept :

[0108] where J′(t) represents the new objective function value, J(t) represents the current objective function value, T(t) represents the current temperature, which gradually decreases as the iteration progresses. The "temperature" in simulated annealing represents the amplitude of exploring the new solution space;

[0109] According to P accept value, determine whether to accept the new solution. If J′(t) is smaller, or accept a worse solution with a certain probability, then update P(t), N(t), and C(t) to P′(t), N′(t), and C′(t);

[0110] With each iteration, the temperature T(t) gradually decreases, T(t) = T0·α t , where T0 represents the initial temperature, α represents the attenuation coefficient, and here t represents the current iteration number; when the temperature T(t) is less than the set threshold or the change in the objective function value is less than a certain threshold, terminate the algorithm and output the final parameter combination P(t), N(t), and C(t);

[0111] Specifically, in step S4, the moisture content and solid content of the sludge are monitored in real time, and the pressure, rotation speed, and chemical dosage of the dewatering equipment are adjusted in combination with the simulated annealing algorithm, significantly improving the efficiency and effect of sludge dewatering. Introducing the simulated annealing algorithm enables the dewatering process to find the optimal solution in the global solution space.

[0112] Step S5, sludge reduction and resource utilization

[0113] Perform sludge reduction and resource treatment on the dewatered sludge, and recommend corresponding resource treatment methods according to the organic content, calorific value, and heavy metal concentration of the sludge;

[0114] Sludge reduction is carried out in step S5:

[0115] Define the comprehensive optimization objective function G(t). The goal is to minimize the environmental risk in the treatment process and maximize the resource recovery efficiency. The objective function is:

[0116] G(t) = w1·O(t) + w2·H(t) + W3·R(t) + W4·M(t), where O(t) represents the organic matter content of the sludge at time t, H(t) represents the calorific value of the sludge, R(t) represents the resource recovery rate, M(t) represents the heavy metal concentration, and w1, w2, w3, w4 represent the weight coefficients;

[0117] The goal is to minimize G(t). During the sludge resource utilization process, the recovery of organic matter and energy is given priority, and at the same time, the risk of heavy metals is controlled;

[0118] Estimate the relationship between the calorific value of the sludge and the recovery of organic matter, H(t) = α·O(t) + β, where α represents the correlation coefficient between the calorific value and the organic matter content, and β represents the basic calorific value of the sludge. Adjust the recovery of organic matter to optimize the calorific value increase;

[0119] Take the heavy metal concentration as an optimization constraint condition, define the heavy metal risk control function M(t) as a limiting term in the treatment process, and the heavy metal concentration does not exceed the environmental protection standard, that is, M(t) ≤ M limit , where M(t) represents the heavy metal concentration of the sludge at time t, and M limit represents the environmental protection concentration standard of heavy metals;

[0120] Define the resource recovery rate: R(t) = γ·H(t) - δ·M(t), where γ represents the calorific value recovery efficiency coefficient and δ represents the heavy metal influence coefficient;

[0121] Based on the optimization objective function G(t) and the constraint conditions, use the multi-objective optimization algorithm to evaluate and select each sludge resource utilization plan. Adjust the treatment plan parameters during each iteration process. Finally, according to the above optimization process, select the best resource utilization plan.

[0122] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A sewage treatment process for the pulp and paper industry, characterized in that: including Step S1, sewage pretreatment Adopt the combined process of air flotation and gravity sedimentation to remove large particle impurities, fibers and suspended solids in the sewage. Set up a water quality sensor to monitor the suspended solid concentration and flow rate of the sewage, and use a dynamic algorithm to adjust the dosage of flocculant and coagulant aid Step S2, staged aeration biological treatment Biologically treat the sewage pretreated in Step S1, and use the staged aeration method to remove high-concentration organic matter and ammonia nitrogen Step S3, sludge thickening predictive control The sludge after biological treatment enters the thickening stage. Adopt predictive control based on the random forest algorithm to analyze the moisture content, solid content and rheological properties of the sludge in real time, and intelligently adjust the thickening stirring intensity and thickening time Step S4, sludge dewatering adjustment control The sludge thickened in Step S3 enters the dewatering link. Real-time monitor the moisture content and solid content of the sludge, and adjust the pressure, rotation speed and chemical dosage of the dewatering equipment in combination with the simulated annealing SA algorithm Step S5, sludge reduction and resource utilization Carry out sludge reduction and resource treatment on the dewatered sludge, and recommend corresponding resource treatment methods according to the organic content, calorific value and heavy metal concentration of the sludge In Step S1, sewage monitoring and data collection are carried out. Specifically Set up a water quality sensor to monitor the suspended solid concentration, flow rate and water temperature in the sewage in real time, and construct a dynamic adjustment model for the dosage of chemical agents to calculate the dosage of chemical agents , where represents the chemical dosage at the current moment , represents the suspended solid concentration at the current moment , represents the sewage flow rate at the current moment , represents the water temperature at the current moment , represents the target suspended solid concentration, which is the preset maximum allowable concentration represents the adaptive adjustment parameter, which is adjusted according to the actual requirements during the sewage treatment process is the integral term of the historical error, representing the cumulative suspended solid concentration error from the start time to the current moment , In Step S1, dynamically adjust the dosage of chemical agents Based on the sensor, obtain the suspended solid concentration, flow rate, and water temperature at the current moment, set the target concentration, and calculate the historical error term , and gradually approach the target value when the suspended solid concentration fluctuates; according to the real-time data and historical error term, use the PID control algorithm to dynamically adjust the parameters of the chemical dosage ; In Step S2, based on the real-time data obtained by the water quality sensor in Step S1, combined with the reinforcement learning RL algorithm, continuously adjust the aeration intensity and time based on the changes in ammonia nitrogen concentration and organic load Carry out sludge reduction in Step S5 Define the comprehensive optimization objective function , with the goal of minimizing the environmental risk during the processing and maximizing the resource recovery efficiency. The objective function is as follows: , where represents the organic matter content of the sludge at time , represents the calorific value of the sludge, represents the resource recovery rate, represents the heavy metal concentration, represents the weight coefficient; The goal is to minimize ; Estimate the relationship between the calorific value of sludge and organic matter recovery, , where, represents the correlation coefficient between the calorific value and the organic matter content, represents the basic calorific value of the sludge; Taking the heavy metal concentration as an optimization constraint condition, a heavy metal risk control function is defined As a limiting item in the treatment process, the heavy metal concentration does not exceed the environmental protection standard, that is , where represents time the heavy metal concentration in the sludge at time represents the environmental protection concentration standard of heavy metals; Define the resource recovery rate: , where represents the calorific value recovery efficiency coefficient, represents the heavy metal influence coefficient; Based on the optimization objective function and constraint conditions, use the multi-objective optimization algorithm to evaluate and select each sludge resource utilization plan, and adjust the treatment plan parameters during each iteration process.

2. The sewage treatment process for the pulp and paper industry according to claim 1, characterized in that, In Step S2, the method of staged aeration biological treatment is Based on the real-time water quality data monitored in Step S1, use it as the input of the reinforcement learning algorithm. Divide the aeration process into pre-aeration, main aeration and post-aeration. Aerate for different ammonia nitrogen concentrations and organic loads in each stage. The aeration intensity and time in each stage are defined as , where represents the aeration intensity at the current moment , represents the ammonia nitrogen concentration at the current moment , represents the organic load at the current moment , represents the sewage flow rate at the current moment , represents the water temperature at the current moment , represents the empirical adjustment parameter, controlling the influence of various factors on the aeration intensity, and needs to be adjusted according to the actual situation Optimize Aeration Intensity Based on Real-Time Feedback and Aeration Time , adjust the ammonia nitrogen removal effect, and the objective function of the reinforcement learning model is: , where represents the total reward value at the current moment , represents the ammonia nitrogen concentration at the current moment , represents the organic load at the current moment , represents the target ammonia nitrogen concentration, which is the maximum allowable concentration represents the target organic load represents the aeration intensity at the current moment , represents the aeration time at the current moment , represents the weight coefficient.

3. The sewage treatment process for the pulp and paper industry according to claim 2, characterized in that, In Step S2, the method of staged aeration biological treatment also includes Updated Aeration Strategy: , where represents the current state and the action The Q-value of, indicating the expected total reward for executing the action in the state . represents the total reward value at the current moment . represents the learning rate, which controls the response speed of the algorithm to new information represents the discount factor, indicating the weight of future rewards represents the system state at the current moment , including water quality data represents the action taken at the current moment , that is, the adjusted aeration intensity and time. During the continuous interaction with the system environment, update value to optimize the aeration operation; According to the optimization of the RL strategy, the final aeration control formula is , where represents the adjustment amount calculated by the RL algorithm.

4. The sewage treatment process for the pulp and paper industry according to claim 3, wherein In Step S3, carry out sludge thickening and predictive control Build a random forest model to predict the sludge characteristics: , where represents the prediction result of the sludge physical characteristics at the current moment , including moisture content, solid content and rheological properties represents the set of operating parameters at the current moment represents the model function learned by the prediction function based on the operating parameters and the random forest algorithm represents the parameters of the model represents the model error, that is, the prediction error​ Obtain the physical properties and operating parameters of the sludge in real time, and use the random forest algorithm to predict the moisture content, solid content and rheological properties in real time. Then, based on the prediction results, adjust the concentration stirring intensity and concentration time. The control formula is: , , where represents the stirring intensity at the current moment , represents the thickening time at the current moment , which is determined by the physical properties and flow rate of the sludge represents the moisture content at the current moment , represents the solid content at the current moment , represents the rheological properties at the current moment , represents the sludge flow rate at the current moment , represents the thickening time represents the adjustment coefficient of the model, which is used to adjust the influence of each parameter on the stirring intensity and thickening time.

5. The sewage treatment process for the pulp and paper industry according to claim 4, characterized in that, In Step S3, adjust the stirring intensity and thickening time according to the real-time predicted physical properties of the sludge , where represents the current moment of the optimized objective function, that is, the deviation between the target state and the actual state and the operation cost, represents the target sludge characteristic value, represents the weight coefficient, balancing the relationship between sludge characteristic optimization and operation cost, and minimizing the objective function , optimizing the stirring intensity and the concentration time; According to the results in the optimization process, the final thickening process control formula is ， 。 6. The sewage treatment process for the pulp and paper industry according to claim 5, wherein In Step S4, based on the real-time monitored moisture content and solid content of the sludge, use the simulated annealing SA algorithm to adjust the parameters of the dewatering equipment Define the optimization objective function , which describes the sludge dewatering effect. The goal is to minimize the moisture content of the sludge, and the form of the objective function is: , where represents the moisture content of the sludge at time point , represents the energy consumption of the dewatering equipment at time point , represents the total duration of the dewatering process, represents the weight coefficient; Search for the optimal solution in the global solution space through the simulated annealing process. The steps include Set the initial solution , and calculate the initial objective function value . Based on the current solution, randomly adjust and to generate a new solution . According to the new operation parameters, calculate the new objective function value ; Determine whether to accept the new solution according to the change of the objective function value, and the probability of accepting the new solution is :[[]]END]] , where represents the new objective function value, represents the current objective function value, represents the current temperature, which gradually decreases as the iteration progresses; Determine whether to accept a new solution according to the value. If is smaller, or accept a worse solution with a certain probability, then update and to ; With each iteration, the temperature gradually decreases, , where represents the initial temperature, represents the attenuation coefficient, and here represents the current iteration number; when the temperature is less than the set threshold or the change in the objective function value is less than a certain threshold, the algorithm is terminated and the final parameter combination and is output.

Images