Deep denitrification continuous-flow aerobic granular sludge reactor and control method thereof

Abstract

The application discloses a kind of deep denitrification continuous flow aerobic granular sludge reactors and control method thereof, belong to wastewater treatment field.Reactor includes the anaerobic pool, aeration tank, anoxic tank, sludge separator and secondary sedimentation tank connected in sequence, and also includes intelligent control system.Control method includes: real-time monitoring state variable;According to the deviation of state variable and preset target value, the control input of last time and control barrier function, build performance function;Approximate long-term cost function using fractional kernel function;Calculate update error and update weight coefficient;Solve optimization problem to obtain control input;Execute control instruction.The application carries out intelligent optimization and closed-loop control to key operating parameters by safety reinforcement learning algorithm, maintains stable treatment effect when influent quality and quantity fluctuate, realizes efficient, stable, self-adapting deep denitrification operation, and has strong impact load capacity.

Description

Technical Field

[0001] This invention belongs to the field of wastewater treatment, and particularly relates to a deep denitrification continuous flow aerobic granular sludge reactor and its control method. Background Technology

[0002] The activated sludge process is one of the most widely used technologies in urban wastewater treatment. Since its invention by British scientists Arden and Lockett in 1914, it has become the mainstream process for treating urban wastewater. This technology degrades organic pollutants in wastewater through a microbial community (activated sludge). Its basic process includes zones with different dissolved oxygen concentrations to provide a suitable biochemical reaction environment, a secondary sedimentation tank for sludge-water separation, a sludge return system to maintain biomass balance, and a waste sludge discharge system. Over its more than 100-year development history, the activated sludge process has evolved from simple organic matter removal to highly efficient nitrogen and phosphorus removal to meet increasingly stringent water quality discharge standards. In recent years, the problem of excessive total nitrogen concentrations has frequently occurred, and traditional processes are gradually becoming unable to meet increasingly stringent discharge standards.

[0003] Aerobic granular sludge is a granular bioaggregate formed by microorganisms through self-aggregation under specific conditions. Compared with traditional flocculent activated sludge, aerobic granular sludge has a regular appearance, dense structure, and excellent settling performance. The granular sludge is composed of aerobic, facultative anaerobic, and anaerobic microorganisms distributed in layers. This structure provides an ideal microenvironment for simultaneous nitrification, denitrification, and phosphorus removal. Furthermore, continuous flow aerobic granular sludge technology is more suitable for in-situ upgrading and expansion of existing wastewater treatment plants, and has enormous application potential.

[0004] Currently, the application of aerobic granular sludge technology in practical engineering mainly relies on sequencing batch reactors (SBRs). SBRs create the selectivity required for aerobic granular sludge formation through time-series operational control. While SBRs have achieved success in cultivating aerobic granular sludge, their intermittent operation mode also has some inherent limitations. Furthermore, SBRs have low reactor volume utilization, and the discontinuous effluent characteristics require subsequent treatment units with large buffer capacities, making series connection with traditional continuous flow treatment processes difficult. These limitations have prompted researchers to explore technical pathways for cultivating and applying aerobic granular sludge in a continuous flow mode.

[0005] The continuous flow aerobic granular sludge treatment technology is still in the research and development stage. There are still problems such as the high concentration of dissolved oxygen in the internal reflux nitrification liquor affecting the denitrification effect in the anoxic zone. In particular, when the influent water quality and quantity fluctuate greatly, the system stability is easily affected, and the sludge reactor recovers slowly. Therefore, its resistance to shock loads is weak. These problems have led to the fact that the technology is not yet fully mature and there are very few engineering application cases. Summary of the Invention

[0006] This invention addresses the problems of poor operational stability and insufficient parameter control precision in existing continuous-flow aerobic granular sludge reactors for deep denitrification processes, and provides a deep denitrification continuous-flow aerobic granular sludge device. By introducing a safety reinforcement learning algorithm, key operating parameters such as sludge return ratio and dissolved oxygen are intelligently optimized and controlled in a closed loop. Even when there are significant changes in influent concentration or flow rate, or when the system faces shocks (such as drastic fluctuations in influent water quality (e.g., COD, ammonia nitrogen, total nitrogen concentration) or influent flow rate exceeding the system's normal operating range), the system can maintain stable treatment performance, keeping core indicators such as effluent quality, sludge concentration, and denitrification efficiency within safe boundaries, thus achieving efficient, stable, and adaptive deep denitrification operation.

[0007] This invention provides a deep denitrification continuous flow aerobic granular sludge reactor and its control method, the specific technical solution of which is as follows: A deep denitrification continuous flow aerobic granular sludge reactor includes: an anaerobic tank, an aeration tank, an anoxic tank, a sludge separator, and a secondary sedimentation tank connected in sequence, and also includes an intelligent control system; a first agitator is installed inside the anaerobic tank, an aeration system is installed inside the aeration tank, and an aeration pipe blower is connected to the aeration system and supplies air to it; the sludge separator is installed inside the anoxic tank, and the sludge separator includes a first sludge return system, the first sludge blower being connected to the first sludge return system to drive sludge back to the anaerobic tank, and controls... The system controls the amount of sludge returned to the first sludge return system; the secondary sedimentation tank includes a second sludge return system, and a second sludge return pump is installed on the second sludge return system to drive the sludge back to the anoxic tank and control the amount of sludge returned to the second sludge return system; the intelligent control system includes an online monitoring system and an automatic control system. The online monitoring system is used to monitor the state variables of each reaction zone, and the automatic control system is connected to the first sludge blower, the second sludge return pump, and the aeration pipe blower respectively to achieve precise parameter control and stable system operation.

[0008] Preferably, the first sludge return system includes a first sludge return pipe and a first sludge air pipe; one end of the first sludge return pipe is connected to the sludge hopper at the bottom of the sludge separator, and the other end is connected to the anaerobic tank; one end of the first sludge air pipe is connected to a first sludge blower, and the other end is connected to the underwater vertical pipe section of the first sludge return pipe.

[0009] Preferably, the aeration system includes an aeration pipe and an aeration disc. The aeration pipe is installed at the bottom of the aeration tank, and the aeration disc is installed on the aeration pipe. One end of the aeration pipe is connected to an aeration pipe blower, which drives the aeration pipe blower to supply air.

[0010] Preferably, the second sludge return system includes a second sludge return pipe, one end of which is connected to the sludge hopper at the bottom of the secondary sedimentation tank, and the other end is connected to the anoxic tank. The second sludge return pump is installed on the second sludge return pipe.

[0011] A control method for a deep denitrification continuous flow aerobic granular sludge reactor, used to control the aforementioned deep denitrification continuous flow aerobic granular sludge reactor, includes the following steps: S1: The sludge concentration (MLSS) and oxidation-reduction potential (ORP) in the anaerobic tank, the ammonia nitrogen and dissolved oxygen concentrations in the aeration tank, and the sludge concentration (MLSS) and nitrate nitrogen concentrations in the anoxic tank are used as state variables. The values ​​of these state variables are monitored in real-time by the intelligent control system and used as the current state variables. The desired values ​​of the state variables are set as preset target values. Lower and upper bounds for the state variables are set as safety boundaries. Control commands sent by the intelligent control system to the first sludge blower, the second sludge return pump, and the aeration pipe blower are used as control inputs. The control command values ​​sent by the intelligent control system at the previous moment are used as the control inputs at the previous moment. The allowable range of control input values ​​is set as the preset control input constraint range. S2: Construct the performance function for the current moment based on the deviation between the current state variable and the preset target value, the control input from the previous moment, and the control obstacle function; wherein, the control obstacle function is designed according to the safety boundary of the state variable and includes a min control obstacle function and a max control obstacle function; S3: The long-term cost function is approximated using a fractional kernel function to obtain the approximate long-term cost function at the current time; the long-term cost function is the time-discounted cumulative sum of the performance function; S4: Calculate the update error based on the difference between the approximate long-term cost function at the previous time step and the approximate long-term cost function at the current time step, as well as the efficiency function at the previous time step, and update the state variable weight coefficients and control input weight coefficients based on the update error; S5: Within the preset control input constraints, solve the optimization problem that minimizes the approximate long-term cost function to obtain the control input at the current moment; S6: The control input at the current moment is output to the intelligent control system, and the intelligent control system controls the operating frequency of the first sludge blower, the pumping frequency of the second sludge return pump, and the operating frequency of the aeration pipe blower; S7: Proceed to the next moment, return to step S4.

[0012] Preferably, the min control barrier function in step S2 is:

[0013] in, Let it be some real number; for The lower bound of the value, ; for Soft parameters; The initial min control barrier function is:

[0014] in, It is the natural logarithm function. The parameters are the initial min control barrier function; For the initial min control barrier function in The value at; let for The first derivative, for The second derivative of is:

[0015]

[0016] for The first derivative in The value at; for The second derivative in The value at that location.

[0017] The max control barrier function is:

[0018] in, Let it be some real number; for The upper bound of the value, And there are ; for Soft parameters; The initial max control barrier function:

[0019] in, It is the natural logarithm function. These are the parameters of the initial max control barrier function; For the initial max control barrier function in The value at; let for The first derivative, for The second derivative of is:

[0020]

[0021] for The first derivative in The value at; for The second derivative in The value at that location.

[0022] The control barrier function is:

[0023] in, The control barrier function; The min control barrier function; The max control barrier function.

[0024] Preferably, the performance function in S2 is:

[0025] in, for The efficiency function at time t, Indicates the current time. express Each component At the present moment The value of , For state variables monitored in real time by intelligent control systems, for 3D real space for dimensionality; express Each component At the present moment The value of , for The expected value; This represents the weighting coefficient for control error. ; express Each component At the present moment The value of , The control input calculated for the intelligent control system. for 3D real space for dimensionality; Indicates the weighting coefficient of the control input. ; This represents the weighting coefficients of the control barrier function. ; Represents the control barrier function exist The value at that location, .

[0026] Preferably, the fractional kernel function in S3 is:

[0027]

[0028] in, Indicates the current time. express Each component At the present moment The value of , For state variables monitored in real time by intelligent control systems, for 3D real space for dimensionality; express Each component At the present moment The value of , The control input calculated for the intelligent control system. for 3D real space for dimensionality; for The corresponding fractional kernel function, for Parameters, ; for The corresponding fractional kernel function, for Parameters, .

[0029] The long-term cost function described in S3 is:

[0030] in, for The long-term cost function at time t. Indicates the current time. Discount factor; for The long-run cost function at time t; for The efficiency function at time t; for The efficiency function at time t, .

[0031] The technical solution of the present invention has the following advantages: 1. In this invention, the sludge separator is built into the anoxic tank, and achieves efficient separation of heavy granular sludge and light flocculent sludge through hydraulic flow. The granular sludge is preferentially returned to the anaerobic tank to promote the purification and enrichment of granular sludge and maintain a high concentration of biomass.

[0032] 2. This invention introduces a safety reinforcement learning algorithm, constructs an efficiency function based on a control barrier function, and dynamically adjusts the first sludge return flow rate, aeration rate, and second sludge return flow rate through online learning to achieve system adaptive optimization.

[0033] 3. This invention combines the AOA mode with aerobic granular sludge to achieve simultaneous nitrification and denitrification. The sludge returned from the secondary sedimentation tank provides a carbon source for denitrification, achieving deep denitrification with total nitrogen below 10 mg / L without the need for an external carbon source.

[0034] 4. The high-concentration granular sludge of the present invention forms a buffer system. The intelligent control system ensures that the state variables are always within the safe boundary by controlling the barrier function. When the influent fluctuates, it recovers to stability within 2 hours, and the effluent quality does not exceed the standard.

[0035] 5. This invention can achieve precise adjustment of aeration and recirculation, avoid over-aeration and ineffective recirculation, and significantly reduce power consumption; the aerobic granular sludge has excellent settling performance, the secondary sedimentation tank has a small volume, and it can be modified in situ within a traditional biological treatment tank, making implementation easy. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram of the deep denitrification continuous flow aerobic granular sludge reactor structure of the present invention; Figure 2 This is a schematic diagram of the reactor control system of the present invention; Figure 3 This is a schematic diagram of the control method for the deep denitrification continuous flow aerobic granular sludge reactor of the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0039] Please see first. Figure 1 This invention discloses a deep denitrification continuous flow aerobic granular sludge reactor, which includes an anaerobic tank 1, an aeration tank 2, an anoxic tank 3, a secondary sedimentation tank 5, and an intelligent control system 6. A sludge separator 4 is installed within the anoxic tank.

[0040] Anaerobic tank 1 The anaerobic tank 1 is equipped with a first agitator 11. The anaerobic tank 1 is typically rectangular, but other shapes may be used depending on the application conditions. The flow pattern in the anaerobic tank 1 is a completely mixed flow. Pretreated wastewater enters the anaerobic tank 1 through the anaerobic tank inlet pipe 12. The anaerobic tank 1 and the aeration tank 2 share a common sidewall (i.e., anaerobic tank 1 and aeration tank 2 share a common wall), and a water passage 13 is provided on this common wall, through which the anaerobic tank 1 and the aeration tank 2 are connected.

[0041] The first agitator 11 can be selected from any type of agitator as needed, but a low-speed agitator is preferred. This provides the necessary mixing power for thorough mixing of the return sludge and the influent, while also preventing damage to the granular sludge within the anaerobic tank 1 during high-speed mixing. In practical use, the user can install one or more agitators depending on the tank type and capacity. This invention uses a single agitator as an example.

[0042] Aeration tank 2 The two sides of aeration tank 2 (or, depending on the wastewater flow direction, referred to as the front and rear sides of aeration tank 2) share walls with anaerobic tank 1 and anoxic tank 3, respectively. Aeration tank 2 is typically rectangular, but its shape can be adjusted as needed. The influent to aeration tank 2 enters through a water passage on its shared wall with the anaerobic tank, while the effluent from aeration tank 2 flows out through a water passage 13 located at the rear of the aeration tank, sharing a wall with the anoxic tank, and enters the anoxic tank 3.

[0043] An aeration system is installed in aeration tank 2. The flow pattern within the aeration tank is plug flow, enabling simultaneous removal of carbon, nitrogen, and phosphorus. The aeration system includes aeration pipes 21 and aeration discs 22. Aeration pipes 21 are installed at the bottom of aeration tank 2, and aeration discs 22 are installed on the aeration pipes 21. One end of the aeration pipe 21 is connected to an aeration pipe blower 23. During operation, the aeration rate of the aeration system is controlled by adjusting the operating frequency of the aeration pipe blower 23.

[0044] Anoxic Pool 3 The anoxic tank 3 shares a wall with the aeration tank 2 and is generally a rectangular tank. The influent of the anoxic tank 3 is introduced through the water passage 13 on the wall shared with the aeration tank 2, and the effluent of the anoxic tank 3 enters the sludge separator.

[0045] A second agitator 31 is installed in the anoxic tank 3. The second agitator 31 is generally a low-speed agitator, mainly to provide stirring power for the second return sludge from the secondary sedimentation tank 5 to the anoxic tank 3 to fully mix with the influent of the anoxic tank 3. At the same time, high-speed stirring should be avoided to prevent damage to the granular sludge. One or more agitators are installed according to the tank type and volume of the anoxic tank 3. In this invention, a single agitator is used as an example.

[0046] sludge separator 4 The sludge separator 4 includes a sludge separator body, a sludge separator inlet system, a sludge separator outlet system, and a first sludge return system.

[0047] The influent to sludge separator 4 is introduced from the anoxic tank 3, and the effluent from sludge separator 4 flows into the secondary sedimentation tank 5. The main function of sludge separator 4 is to screen and separate granular sludge from flocculent sludge to achieve the purification and enrichment of granular sludge. The mixed liquor flow direction in sludge separator 4 is that water enters from the short side of the sludge separator and exits from the effluent tank, presenting an overall longitudinal flow pattern from top to bottom and from bottom to top, with a flow velocity of 5-20 mm / s.

[0048] The sludge separator 4 is mainly located at the end of the anoxic tank 3. Depending on the scale of wastewater treatment, the number of sludge separators 4 can be one or more; this invention uses one as an example. The sludge separator 4 mainly consists of an open rectangular box at the top and a closed inverted triangular sludge hopper 48 at the bottom. The outer wall panel of the sludge separator separates it from the water in the anoxic tank. The sludge separator has a rectangular structure with a length-to-width ratio of 3:1 to 5:1. The influent flow rate of the sludge separator is controlled by adjusting the weir gate, and the surface loading of the sludge separator is 2-5 m³. 3 / m 2 ·h.

[0049] The sludge separator's water inlet system includes an inlet hole 41 and an inlet rectifier plate 42. The inlet hole 41 is located on the upper part of the short side plate of the sludge separator 4, and is generally submerged 0.2-0.5m underwater. The inlet rectifier plate 42 is located inside the sludge separator 4, parallel to the short side plate of the sludge separator 4 with the same width. The upper end of the inlet rectifier plate 42 is at the same height as the short side plate of the sludge separator, and the lower end is submerged 0.5-1.0m underwater.

[0050] The sludge separator effluent system includes an overflow weir, a collection trough 43, and a sludge separator effluent pipe 44. The collection trough 43 is located on the upper part of the sludge separator 4 and away from the sludge separator inlet system. Both ends of the collection trough 43 are fixed to the side wall plates of the sludge separator 4. One end of the sludge separator effluent pipe 44 is connected to the effluent trough, and the other end passes through the side wall of the sludge separator 4 and the anoxic tank 3 and connects to the secondary sedimentation tank.

[0051] The first sludge return system includes a first sludge return pipe 45 and a first sludge air pipe 46. One end of the first sludge return pipe 45 is connected to the sludge hopper 48 at the bottom of the sludge separator 4, and the other end is connected to the anaerobic tank 1. One end of the first sludge air pipe 46 is connected to the first sludge blower 47, and the other end is connected to a 1.5-2.0m vertical section underwater in the first sludge return pipe 45. During operation, by controlling the operating frequency of the first sludge blower 47, the system controls whether air enters the first sludge air pipe 46 and the amount of air entering, ultimately controlling the return flow rate of the first sludge in the first sludge return pipe 45.

[0052] Secondary sedimentation tank 5 Secondary sedimentation tank 5 includes the secondary sedimentation tank body, secondary sedimentation tank outlet pipe 55, second sludge return system and excess sludge discharge system.

[0053] The secondary sedimentation tank (5) is cylindrical, with an annular effluent channel on the upper part of its inner wall and a sludge scraper in the middle. The inlet water to the secondary sedimentation tank is connected to the effluent pipe (44) of the sludge separator. The effluent from the secondary sedimentation tank is the supernatant discharged through the effluent pipe.

[0054] The second sludge return system includes a second sludge return pipe 51 and a second sludge return pump 52. One end of the second sludge return pipe 51 is connected to the sludge hopper at the bottom of the secondary settling tank 5, and the other end of the second sludge return pipe 51 is connected to the anoxic tank 3. The second sludge return pump 52 is installed on the second sludge return pipe 51 and is used to transport the second sludge from the secondary settling tank 5 to the anoxic tank 3.

[0055] The waste sludge discharge system includes a waste sludge discharge pump 53 and a waste sludge discharge pipe 54. One end of the waste sludge discharge pipe 54 is connected to the sludge hopper at the bottom of the secondary sedimentation tank 5, and the other end is connected to the sludge dewatering room. The waste sludge discharge pump 53 is installed on the waste sludge discharge pipe 54 and is used to transport the waste sludge from the secondary sedimentation tank 5 to the sludge dewatering room, so that the waste sludge is discharged outside the sludge reactor system.

[0056] Please see Figure 2 The intelligent control system 6 includes an online monitoring system 61 and an automatic control system 62.

[0057] The online monitoring system 61 includes: a sludge concentration meter (MLSS analyzer), an oxidation-reduction potential meter (ORP analyzer), a dissolved oxygen meter (DO meter), an ammonia nitrogen meter, and a nitrate nitrogen analyzer.

[0058] The anaerobic tank 1 is equipped with an ORP analyzer and an MLSS analyzer at the bottom to monitor the oxidation-reduction potential and sludge concentration in real time. The aeration tank 2 is equipped with a DO analyzer and an ammonia nitrogen analyzer at the bottom to monitor the dissolved oxygen and ammonia nitrogen concentrations in real time. The anoxic tank 3 is equipped with an MLSS analyzer and a nitrate nitrogen analyzer at the bottom to monitor the sludge and nitrate nitrogen concentrations in real time.

[0059] The automatic control system 62 is signal-connected to the online monitoring system to receive status data collected by the online monitoring system. The automatic control system is also signal-connected to the first sludge blower 47, the second sludge return pump 52, and the aeration pipe blower 23 to send control commands to these actuators.

[0060] The working principle of the deep denitrification continuous flow aerobic granular sludge reactor of the present invention is as follows: 1. After pretreatment, the influent enters the anaerobic tank through the anaerobic tank inlet pipe. The anaerobic tank influent is thoroughly mixed with the first sludge returned from the sludge separator. The first sludge is mostly granular sludge. The polyphosphate-accumulating bacteria in the first sludge can store carbon sources such as polyhydroxy fatty acids (PHAs) in the anaerobic stage, while fully releasing phosphorus.

[0061] 2. The effluent from the anaerobic tank enters the aeration tank. The main function of the aeration tank is to simultaneously perform nitrification / denitrification and phosphorus removal to remove ammonia nitrogen, total nitrogen, and phosphate from the water. AOB (ammonia oxidizing bacteria) and NOB (nitrite oxidizing bacteria) utilize dissolved oxygen to oxidize ammonia nitrogen in the wastewater into nitrate nitrogen. Simultaneously, due to the special stratified structure of the aerobic granular sludge, the denitrifying bacteria inside the aerobic granules utilize the intracellular carbon source stored in the anaerobic stage to perform denitrification, achieving preliminary nitrogen removal. Polyphosphate-accumulating bacteria utilize the intracellular carbon source previously stored in the anaerobic zone to transport extracellular phosphate into the cell to resynthesize polyphosphate, achieving phosphorus removal.

[0062] 3. The effluent from the aeration tank enters the anoxic tank. The main function of the anoxic tank is denitrification and nitrogen removal. It further utilizes the nitrate nitrogen flowing into the aeration tank and the intracellular carbon source in the sludge and the second sludge returned from the secondary sedimentation tank for the deep removal of actual nitrogen through denitrification.

[0063] 4. The effluent from the anoxic tank enters the sludge separator, whose main function is the purification and enrichment of granular sludge. Sludge with a high settling density, which settles to the bottom sludge hopper of the separator, is returned to the anaerobic tank via airlift. Lighter sludge, due to its lower density and slower settling velocity, settles to the upper part of the sludge hopper, then flows with the water into the effluent channel, and finally enters the secondary settling tank.

[0064] 5. After the mud-water mixture entering the secondary sedimentation tank is separated into mud and water, the supernatant flows into the deep treatment unit outside the sludge reactor through the annular effluent channel. Part of the sludge at the bottom of the secondary sedimentation tank is discharged from the sludge reactor as excess sludge, and the other part is returned to the anoxic tank as second sludge to continue to participate in deep denitrification.

[0065] 6. The sludge reactor of the present invention also has an intelligent control system, which includes an online monitoring system and an automatic control system. The online monitoring system monitors the real-time MLSS and ORP values ​​in the anaerobic tank, the ammonia nitrogen and dissolved oxygen values ​​in the aeration tank, and the MLSS and nitrate nitrogen values ​​in the anoxic tank. After acquiring the above values, the automatic control system sends control commands to the first sludge blower, the second sludge return pump, and the aeration pipe blower to regulate the first sludge return flow rate, the aeration rate of the aeration system in the aeration tank, and the second sludge return flow rate. This enhances the phosphorus release and carbon absorption effect of polyphosphate-accumulating bacteria in the anaerobic zone, ensures sufficient nitrification of ammonia nitrogen without over-aeration, and enhances the denitrification effect in the anoxic zone. Simultaneously, the system ensures that even when faced with sudden and significant fluctuations in influent water quality and volume, or when system stability is impacted (such as drastic fluctuations in influent water quality (e.g., COD, ammonia nitrogen, total nitrogen concentration) or influent water volume exceeding the system's normal operating range), the sludge reactor system can maintain stable treatment performance. This ensures that core indicators such as effluent quality, sludge concentration, and denitrification efficiency remain consistently within safe limits, achieving efficient, stable, and adaptive deep denitrification operation. The sludge reactor exhibits stronger resistance to shock loads.

[0066] Please see Figure 3 This invention provides a control method for a deep denitrification continuous flow aerobic granular sludge reactor. This method is applied to the deep denitrification continuous flow aerobic granular sludge reactor described in the foregoing embodiments and is used for precise and adaptive control of the wastewater treatment process.

[0067] The control method of a deep denitrification continuous flow aerobic granular sludge reactor of the present invention is as follows: S1: The sludge concentration (MLSS) and oxidation-reduction potential (ORP) in the anaerobic tank, the ammonia nitrogen and dissolved oxygen concentrations in the aeration tank, and the sludge concentration (MLSS) and nitrate nitrogen concentrations in the anoxic tank are used as state variables. The values ​​of these state variables are monitored in real-time by the intelligent control system and used as the current state variables. The desired values ​​of the state variables are set as preset target values. Lower and upper bounds for the state variables are set as safety boundaries. Control commands sent by the intelligent control system to the first sludge blower, the second sludge return pump, and the aeration pipe blower are used as control inputs. The control command values ​​sent by the intelligent control system at the previous moment are used as the control inputs at the previous moment. The allowable range of control input values ​​is set as the preset control input constraint range. S2: Construct the performance function for the current moment based on the deviation between the current state variable and the preset target value, the control input from the previous moment, and the control obstacle function; wherein, the control obstacle function is designed according to the safety boundary of the state variable and includes a min control obstacle function and a max control obstacle function; S3: The long-term cost function is approximated using a fractional kernel function to obtain the approximate long-term cost function at the current time; the long-term cost function is the time-discounted cumulative sum of the performance function; S4: Calculate the update error based on the difference between the approximate long-term cost function at the previous time step and the approximate long-term cost function at the current time step, as well as the efficiency function at the previous time step, and update the state variable weight coefficients and control input weight coefficients based on the update error; S5: Within the preset control input constraints, solve the optimization problem that minimizes the approximate long-term cost function to obtain the control input at the current moment; S6: The control input at the current moment is output to the intelligent control system, and the intelligent control system controls the operating frequency of the first sludge blower, the pumping frequency of the second sludge return pump, and the operating frequency of the aeration pipe blower; S7: Proceed to the next moment, return to step S4.

[0068] Specifically: S1: The sludge concentration (MLSS) and oxidation-reduction potential (ORP) in the anaerobic tank, the ammonia nitrogen and dissolved oxygen concentrations in the aeration tank, and the sludge concentration (MLSS) and nitrate nitrogen concentrations in the anoxic tank are used as state variables. The values ​​of these state variables are monitored in real-time by the intelligent control system and used as the current state variables. The desired values ​​of the state variables are set as preset target values. Lower and upper bounds for the state variables are set as safety boundaries. Control commands sent by the intelligent control system to the first sludge blower, the second sludge return pump, and the aeration pipe blower are used as control inputs. The control command values ​​sent by the intelligent control system at the previous moment are used as the control inputs at the previous moment. The allowable range of control input values ​​is set as the preset control input constraint range. The automatic control system acquires real-time status data through an online monitoring system and uses this data as the values ​​of the current status variables. Specifically, the online monitoring system collects the values ​​of the following status variables: sludge concentration (MLSS) and oxidation-reduction potential (ORP) in the anaerobic tank; ammonia nitrogen and dissolved oxygen concentrations in the aeration tank; and sludge concentration (MLSS) and nitrate nitrogen concentrations in the anoxic tank.

[0069] Furthermore, the expected values ​​of state variables, as well as the lower and upper bounds of their values, can be set according to actual needs and are not specifically limited here. The permissible range of control input values ​​can be set according to actual needs or the operating limitations of the equipment used and are not specifically limited here.

[0070] S2: Construct the performance function for the current moment based on the deviation between the current state variable and the preset target value, the control input from the previous moment, and the control obstacle function; wherein, the control obstacle function is designed according to the safety boundary of the state variable and includes a min control obstacle function and a max control obstacle function; This step constructs the performance function for the current moment based on the deviation between the current state variable and the preset target value, the control input from the previous moment, and the control obstacle function value.

[0071] 1. Design of the control barrier function Suppose for some real number ,expect and , and They are respectively The lower and upper bounds of the value are used as Safety boundary, , And there are .

[0072] 1.1 Design the initial min control barrier function : Regarding expectations Design the following initial min control barrier function. :

[0073] in, It is the natural logarithm function. These are the parameters of the initial min control barrier function. Calculate. The first and second derivatives have

[0074]

[0075] in, for The first derivative, for The second derivative of .

[0076] 1.2 Calculate the initial min control barrier function First and second derivatives at a specific point Pick for The soft parameters, according to The expressions for the first and second derivatives are given by... exist First and second derivatives at:

[0077] 1.3 Design the final min control barrier function :

[0078] in, for The first derivative in The value at; for The second derivative in The value at that location.

[0079] 1.4 Design of the initial max control barrier function : Regarding expectations Design the following initial max control barrier function. :

[0080] in, It is the natural logarithm function. These are the parameters for the initial max control barrier function. Calculation. The first and second derivatives have

[0081]

[0082] in, for The first derivative, for The second derivative of .

[0083] 1.5 Calculate the initial max control barrier function First and second derivatives at a specific point Pick for The soft parameters, according to The expressions for the first and second derivatives are given by... exist First and second derivatives at:

[0084]

[0085] 1.6 Design the final max control barrier function :

[0086] in, for The first derivative in The value at; for The second derivative in The value at that location.

[0087] 1.7 Based on the final min control barrier function and the final max control barrier function, design the final control barrier function:

[0088] in, The control barrier function; The min control barrier function; The max control barrier function.

[0089] 2. Determine the control objectives of the intelligent control system. Let the state variable monitored in real time by the intelligent control system be... , for 3D real space for The dimension of the intelligent control system is the control input calculated by the intelligent control system. , for 3D real space for dimensionality and Relationship satisfaction

[0090] in, For system functions, , Indicates the current time. Indicates the next moment, State variable representing the next time step The value of , Indicates the control input at the current moment. The possible values ​​of ; They represent the past 1 to 12 respectively. At that moment; For the past 1 to The state variable at time 1 Values; They represent the past 1 to 12 respectively. At that moment, For the past 1 to Control input at each moment Values, and The system order is denoted by .

[0091] because , can Written as ,in for Each component, set The expected value is The control objective of an intelligent control system is to achieve the desired result in the system function. System order and Design under the premise of unknown conditions , making Gradually approaching over time At the same time, always satisfy and .in, , , , for Each component The preset target value, and satisfies and ; For setting Each component The lower bound of the value, For setting Each component The upper bound of the value is used as The safety boundary, and has ; express , express .

[0092] Specifically, state variables The values ​​of sludge concentration (MLSS) and oxidation-reduction potential (ORP) in the anaerobic tank, ammonia nitrogen concentration and dissolved oxygen concentration in the aeration tank, and sludge concentration (MLSS) and nitrate nitrogen concentration in the anoxic tank are given. .

[0093] The MLSS value of sludge concentration in the anaerobic tank is ,set up The preset target value is The lower bound of the value is The upper bound of the value is ; The oxidation-reduction potential (ORP) value in the anaerobic tank is ,set up The preset target value is The lower bound of the value is The upper bound of the value is ; The ammonia nitrogen concentration in the aeration tank is ,set up The preset target value is The lower bound of the value is The upper bound of the value is ; The dissolved oxygen concentration in the aeration tank is ,set up The preset target value is The lower bound of the value is The upper bound of the value is ; The MLSS value of sludge concentration in the anoxic tank is ,set up The preset target value is The lower bound of the value is The upper bound of the value is ; The nitrate nitrogen concentration in the anoxic tank is ,set up The preset target value is The lower bound of the value is The upper bound of the value is .

[0094] 3. Set the allowable range of control input values ​​as the preset control input constraint range. because , can Written as ,in for Each component corresponds to a control command for an actuator. Based on the characteristics of each actuator, the following settings are configured: Each component The permissible range of values ​​is and As The range of constraints, where for Each component The lower bound of allowed values, for Each component The upper bound on the allowed values.

[0095] Specifically, control input The operating frequency of the first sludge blower, the pump frequency of the second sludge return pump, and the operating frequency of the aeration pipe blower are... .

[0096] The operating frequency of the first sludge blower is Set the lower bound of the allowed value to be . The upper bound for the allowed value is 0. ; The pump frequency of the second sludge return pump is Set the lower bound of the allowed value to be . The upper bound for the allowed value is 0. ; The operating frequency of the aeration pipe blower is: Set the lower bound of the allowed value to be . The upper bound for the allowed value is 0. .

[0097] 4. Construct the performance function Based on 1 and 2, construct the following performance function.

[0098] in, for The efficiency function at time t, Indicates the current time. express Each component At the present moment The value of , For state variables monitored in real time by intelligent control systems, for 3D real space for dimensionality; express Each component At the present moment The value of , for The expected value; This represents the weighting coefficient for control error. ; express Each component At the present moment The value of , The control input calculated for the intelligent control system. for 3D real space for dimensionality; Indicates the weighting coefficient of the control input. ; This represents the weighting coefficients of the control barrier function. ; Represents the control barrier function exist The value at that location, .

[0099] S3: The long-term cost function is approximated using a fractional kernel function to obtain the approximate long-term cost function at the current time; the long-term cost function is the time-discounted cumulative sum of the performance function; according to Performance function at time step Design the following long-term cost function.

[0100] in, for The long-term cost function at time t. Indicates the current time. Discount factor; for The long-run cost function at time t; for The efficiency function at time t; for The efficiency function at time t, Using fractional kernel functions to... By approximation, we obtain

[0101] in, for Approximation, Indicates the current time. express Each component At the present moment The value of , For state variables monitored in real time by intelligent control systems, for 3D real space for dimensionality; Weight coefficients for state variables At the present moment The value of , for The corresponding fractional kernel function, ; for The corresponding fractional kernel function, express Each component At the present moment The value of , The control input calculated for the intelligent control system. for 3D real space for dimensionality; To control the input weighting coefficients At the present moment The value of , .

[0102] Corresponding fractional kernel function The specific expression is

[0103] in, for Parameters, .

[0104] Corresponding fractional kernel function The specific expression is

[0105] in, for Parameters, .

[0106] S4: Calculate the update error based on the difference between the approximate long-term cost function at the previous time step and the approximate long-term cost function at the current time step, as well as the efficiency function at the previous time step, and update the state variable weight coefficients and control input weight coefficients based on the update error; Define update error

[0107] in, Indicates the current time. Indicates the previous moment, This represents the value of the performance function at the previous time step. This represents the approximate long-term cost function value at the previous time step. Furthermore, after obtaining the current state value... and the previous moment Control input value Then, the weight coefficients of the state variables are updated in the following manner. and control input weight coefficients

[0108] ,

[0109] ,

[0110] in, Weight coefficients for state variables Update step size, ; To control the input weighting coefficients The update step size.

[0111] S5: Within the preset control input constraints, solve the optimization problem that minimizes the approximate long-term cost function to obtain the control input at the current moment; Based on the allowable range of control input values ​​set in S2.3, the following calculation of the control input for the intelligent control system is performed. The optimization problem is to obtain the control input at the current moment.

[0112]

[0113] in, Indicates the current time. for Each component At the present moment The value of , The control input calculated for the intelligent control system. for 3D real space for The dimension of . This can be obtained by solving the above optimization problem. That is, we obtained The solution methods include, but are not limited to, heuristic optimization algorithms and sequential quadratic programming algorithms.

[0114] S6: The control input at the current moment is output to the intelligent control system, and the intelligent control system controls the operating frequency of the first sludge blower, the pumping frequency of the second sludge return pump, and the operating frequency of the aeration pipe blower; The control input at the current moment obtained from the solution The output is sent to the intelligent control system, which then controls the corresponding actuators to perform actions based on the current control input. The operating frequency of the first sludge blower is controlled, which in turn controls the air intake of the first sludge air pipe, thereby controlling the return flow rate of the first sludge in the first sludge return pipe; Control the pump frequency of the second sludge return pump and adjust the return flow rate of the second sludge in the second sludge return pipe; Control the operating frequency of the blower in the aeration pipe to adjust the aeration volume of the aeration system.

[0115] S7: Proceed to the next moment, return to step S4.

[0116] After execution, wait for the next control cycle and proceed to the next moment. Return to step S4, update the state variable weight coefficients and control input weight coefficients to form closed-loop control.

[0117] The intelligent control method described in this embodiment can achieve the following technical effects: 1. Precise control: Through the dual constraints of the min and max control barrier functions, the key state variables (MLSS and ORP in the anaerobic tank, ammonia nitrogen and dissolved oxygen in the aeration tank, and MLSS and nitrate nitrogen in the anoxic tank) are always kept within the preset safe and efficient range, avoiding system instability and excessive effluent.

[0118] 2. Adaptive optimization: By updating the weight coefficients of state variables and control inputs online, the automatic control system can autonomously learn the changing patterns of influent water quality and quantity, dynamically adjust the control strategy, and achieve personalized optimization.

[0119] 3. Energy saving and consumption reduction: By precisely adjusting the operating frequency of the first sludge blower, the second sludge return pump and the aeration pipe blower, over-aeration and ineffective return are avoided. Compared with traditional manual control or PID control, power consumption can be significantly reduced.

[0120] 4. Deep denitrification: By synergistically controlling the first and second sludge return flows and optimizing carbon source distribution, a deep denitrification effect of less than 10 mg / L of total nitrogen can be achieved without adding external carbon sources.

[0121] 5. Enhanced resistance to shock loads: When the influent water quality and quantity fluctuate significantly, the intelligent control system can respond quickly by adjusting the first sludge return flow rate, aeration rate, and second sludge return flow rate to ensure that the system maintains a stable treatment effect, keeping core indicators such as effluent water quality, sludge concentration, and denitrification efficiency consistently within safe limits.

[0122] Example To test the actual operating effect of the deep denitrification continuous flow aerobic granular sludge reactor and its control method of the present invention, the following embodiments are provided.

[0123] The test was conducted at a municipal wastewater treatment plant. The reactor had a treatment capacity of 1600 tons / day and consisted of sequentially connected anaerobic, aeration, anoxic, and secondary sedimentation tanks. A sludge separator was installed in the anoxic tank. The main process parameters are as follows:

[0124] This embodiment employs an intelligent control system to control the sludge reactor. An online monitoring system collects real-time data on MLSS and ORP in the anaerobic tank, ammonia nitrogen and dissolved oxygen in the aeration tank, and MLSS and nitrate nitrogen in the anoxic tank as state variables. The automatic control system calculates the optimal control inputs based on these state variables, including the frequency of the first sludge blower, the frequency of the second sludge return pump, and the frequency of the aeration pipe blower.

[0125] The sludge reactor and its control method of this invention were used for treatment. After 45 days of actual operation of the sludge reactor, the test results are as follows: MLSS in anaerobic ponds 4200 mg / L 6050 mg / L 6000 mg / L 4000 mg / L 7000 mg / L Anaerobic tank ORP -180 mV -198 mV -200 mV -220 mV -170 mV Ammonia nitrogen in aeration tank 2.5 mg / L 0.95 mg / L 1.0 mg / L 0.8 mg / L 3 mg / L Dissolved oxygen in aeration tank 1.1 mg / L 0.78 mg / L 0.8 mg / L 0.6 mg / L 1.2 mg / L MLSS in anoxic pool 4800 mg / L 6950 mg / L 7000 mg / L 4500 mg / L 7500 mg / L Nitrate nitrogen in anoxic pond 9.5 mg / L 5.2 mg / L 5.0 mg / L 3.0 mg / L 12.0 mg / L Effluent water quality: COD: 15.4 mg / L, ammonia nitrogen: 0.32 mg / L, total nitrogen: 5.4 mg / L. The effluent water quality meets the Class III standard of the "Surface Water Environmental Quality Standard" (GB3838-2002).

[0126] During the operation of the sludge reactor, the values ​​of each state variable always remained within the set lower and upper limits. After 45 days of operation, the particle size of the sludge reactor reached 21.3%, and the granular sludge had a dense structure and excellent settling performance.

[0127] During operation, the influent water quality was simulated to fluctuate (COD increased from 300 mg / L to 450 mg / L). The sludge reactor responded quickly, adjusted the control input in real time, and the system returned to stable operation. The effluent water quality did not exceed the standard.

[0128] The above operational results demonstrate that the sludge reactor of this invention operates with high efficiency.

[0129] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A deep denitrification continuous flow aerobic granular sludge reactor, characterized in that... It includes: an anaerobic tank, an aeration tank, an anoxic tank, a sludge separator, and a secondary sedimentation tank connected in sequence, and also includes an intelligent control system; The anaerobic tank is equipped with a first agitator, and the aeration tank is equipped with an aeration system. An aeration pipe blower is connected to the aeration system and supplies air to it. The sludge separator is located inside the anoxic tank. The sludge separator includes a first sludge return system, with the first sludge blower connected to the first sludge return system to drive sludge back to the anaerobic tank and control the amount of sludge returned to the first sludge return system. The secondary sedimentation tank includes a second sludge return system, with a second sludge return pump installed on the second sludge return system to drive sludge back to the anoxic tank and control the amount of sludge returned to the second sludge return system. The intelligent control system includes an online monitoring system and an automatic control system. The online monitoring system monitors the state variables of each reaction zone, and the automatic control system is connected to the first sludge blower, the second sludge return pump, and the aeration pipe blower to achieve precise parameter control and stable system operation.

2. The sludge reactor as described in claim 1, characterized in that... The first sludge return system includes a first sludge return pipe and a first sludge air pipe; one end of the first sludge return pipe is connected to the sludge hopper at the bottom of the sludge separator, and the other end is connected to the anaerobic tank; one end of the first sludge air pipe is connected to the first sludge blower, and the other end is connected to the underwater vertical pipe section of the first sludge return pipe.

3. The sludge reactor as described in claim 1, characterized in that... The aeration system includes an aeration pipe and an aeration disc. The aeration pipe is installed at the bottom of the aeration tank, and the aeration disc is installed on the aeration pipe. One end of the aeration pipe is connected to an aeration pipe blower, which drives the aeration pipe blower to supply air.

4. The sludge reactor as described in claim 1, characterized in that... The second sludge return system includes a second sludge return pipe, one end of which is connected to the sludge hopper at the bottom of the secondary sedimentation tank, and the other end is connected to the anoxic tank. The second sludge return pump is installed on the second sludge return pipe.

5. A control method for a deep denitrification continuous flow aerobic granular sludge reactor, used to control the deep denitrification continuous flow aerobic granular sludge reactor as described in claim 1, characterized in that... It includes the following steps: S1: The sludge concentration (MLSS) and oxidation-reduction potential (ORP) in the anaerobic tank, the ammonia nitrogen and dissolved oxygen concentrations in the aeration tank, and the sludge concentration (MLSS) and nitrate nitrogen concentrations in the anoxic tank are used as state variables. The values ​​of these state variables are monitored in real-time by the intelligent control system and used as the current state variables. The desired values ​​of the state variables are set as preset target values. Lower and upper bounds for the state variables are set as safety boundaries. Control commands sent by the intelligent control system to the first sludge blower, the second sludge return pump, and the aeration pipe blower are used as control inputs. The control command values ​​sent by the intelligent control system at the previous moment are used as the control inputs at the previous moment. The allowable range of control input values ​​is set as the preset control input constraint range. S2: Construct the performance function for the current moment based on the deviation between the current state variable and the preset target value, the control input from the previous moment, and the control obstacle function; wherein, the control obstacle function is designed according to the safety boundary of the state variable and includes a min control obstacle function and a max control obstacle function; S3: The long-term cost function is approximated using a fractional kernel function to obtain the approximate long-term cost function at the current time; the long-term cost function is the time-discounted cumulative sum of the performance function; S4: Calculate the update error based on the difference between the approximate long-term cost function at the previous time step and the approximate long-term cost function at the current time step, as well as the efficiency function at the previous time step, and update the state variable weight coefficients and control input weight coefficients based on the update error; S5: Within the preset control input constraints, solve the optimization problem that minimizes the approximate long-term cost function to obtain the control input at the current moment; S6: The control input at the current moment is output to the intelligent control system, and the intelligent control system controls the operating frequency of the first sludge blower, the pumping frequency of the second sludge return pump, and the operating frequency of the aeration pipe blower; S7: Proceed to the next moment, return to step S4.

6. The control method as described in claim 5, characterized in that... The min control obstacle function mentioned in step S2 is: in, Let it be some real number; for The lower bound of the value, ; for Soft parameters; The initial min control barrier function is: in, It is the natural logarithm function. The parameters are the initial min control barrier function; For the initial min control barrier function in The value at; let for The first derivative, for The second derivative of is: for The first derivative in The value at; for The second derivative in The value at; The max control barrier function is: in, Let it be some real number; for The upper bound of the value, And there are ; for Soft parameters; The initial max control barrier function: in, It is the natural logarithm function. These are the parameters of the initial max control barrier function; For the initial max control barrier function in The value at; let for The first derivative, for The second derivative of is: for The first derivative in The value at; for The second derivative in The value at that location.

7. The control method as described in claim 6, characterized in that... , The control barrier function is: in, The control barrier function; The min control barrier function; The max control barrier function.

8. The control method as described in claim 7, characterized in that... The performance function in S2 is: in, for The efficiency function at time t, Indicates the current time. express Each component At the present moment The value of , For state variables monitored in real time by intelligent control systems, for 3D real space for dimensionality; express Each component At the present moment The value of , for The expected value; This represents the weighting coefficient for control error. ; express Each component At the present moment The value of , The control input calculated for the intelligent control system. for 3D real space for dimensionality; Indicates the weighting coefficient of the control input. ; This represents the weighting coefficients of the control barrier function. ; Represents the control barrier function exist The value at that location, .

9. The control method as described in claim 5, characterized in that... The fractional kernel function described in S3 is: in, Indicates the current time. express Each component At the present moment The value of , For state variables monitored in real time by intelligent control systems, for 3D real space for dimensionality; express Each component At the present moment The value of , The control input calculated for the intelligent control system. for 3D real space for dimensionality; for The corresponding fractional kernel function, for Parameters, ; for The corresponding fractional kernel function, for Parameters, .

10. The control method as described in claim 5, characterized in that... The long-run cost function described in S3 is: in, for The long-term cost function at time t. Indicates the current time. Discount factor; for The long-run cost function at time t; for The efficiency function at time t; for The efficiency function at time t, .

Images