AU2020102748A4

AU2020102748A4 - Intelligent feedback real-time control system and method for hydrodynamic circulation under external interference

Info

Publication number: AU2020102748A4
Application number: AU2020102748A
Authority: AU
Inventors: Xing Fan; Li He; Jing Li; Hongwei LU; Fangping Yin
Original assignee: Tianjin University
Current assignee: Tianjin University
Priority date: 2020-10-16
Filing date: 2020-10-16
Publication date: 2020-12-10
Anticipated expiration: 2028-10-16

Abstract

The invention provides a hydrodynamic circulation intelligent feedback real-time control system and method under external interference. The control method comprises the following steps: establishing a three-dimensional water-gas-heat-electricity multi-field coupled groundwater stochastic simulation model, and obtaining the concentration of pollutants in the remediation process by using the simulation model; Through the deep learning method of artificial intelligence, the fitting relationship between pollutant concentration and operation conditions of remediation system is established; The objective optimization model is established, and the fitting relationship between operating conditions and pollutant concentration is taken as the constraint condition to solve the optimal operating conditions of the remediation system under the dynamic change scenario. The feedback of groundwater index monitoring can control the optimal operating conditions of the remediation system in real time. According to that control system and the control method, the best applicable operate conditions can be determined under exogenous interference, different groundwater pollution condition and specific remediation targets, and the optimized remediation system operating conditions can be obtained in real time. 1 / 11 A three-dimensional water, gas, heat and power mut-fied sI coupled groundwater stochastic simulation model is established, and the pollutant concentration in the process of remediation is used by this simulation model Through the deep learning method of artificial intelligence, the -- fitting reationship between pollutant concentration and remediation system conditions is established The objective optimization model is established, and the fitting relationship between operating conditions and pollutant concentration is taken as a constraint condition to solve the optimal operating conditions of the remediation system under dynamic change scenarios. The feedback of groundwater index monitoring can control the optimal operating conditions of the remediation system in real time. --------------------------------------------------------- Fig.1I

Description

1 / 11

A three-dimensional water, gas, heat and power mut-fied sI coupled groundwater stochastic simulation model is established, and the pollutant concentration in the process of remediation is used by this simulation model

Through the deep learning method of artificial intelligence, the -- fitting reationship between pollutant concentration and remediation system conditions is established

The objective optimization model is established, and the fitting relationship between operating conditions and pollutant concentration is taken as a constraint condition to solve the optimal operating conditions of the remediation system under dynamic change scenarios. The feedback of groundwater index monitoring can control the optimal operating conditions of the remediation system in real time.

---------------------------------------------------------

Fig.1I

Intelligent feedback real-time control system and method for hydrodynamic circulation under external interference

TECHNICAL FIELD

[01] The invention belongs to the field of groundwater environment monitoring and control, and particularly relates to a hydrodynamic circulation intelligent feedback real-time control system and method under external interference.

BACKGROUND

[02] In the process of modernization, unreasonable exploitation and utilization of groundwater leads to an increasing shortage of groundwater resources, accompanied by various pollution problems. Considering the seriousness of groundwater pollution, countries around the world have already begun to study the remediation methods of groundwater pollution, which can be summarized into four aspects:

[03] water, using groundwater movement to remove in-situ or ectopic pollutants, mainly for heavy metal or organic pollution of groundwater with large pollution range and serious pollution degree, usually adopting osmotic reaction barrier technology, pumping treatment technology and in-situ natural attenuation, etc. Osmotic reaction barrier technology usually excavates trenches downstream of pollution sources, fills reaction medium and sets reaction walls. Pumping treatment technology is to use pumping wells to pump contaminated groundwater out of the ground, and use the surface treatment system to repair and control the pumped contaminated groundwater. When pumping/injecting water, the flow field of groundwater changes, and under certain hydraulic conditions, polluted water can be trapped, thus separating polluted water from clean water. Most of the treated groundwater up to standard is used for recharge. Recharged groundwater is mixed with the original groundwater, which dilutes the polluted groundwater and flushes the aquifer medium to a certain extent. At the same time, it also promotes the virtuous cycle of groundwater, accelerates the velocity of groundwater, thus increasing the removal rate of pollutants and reducing the treatment and repair time of groundwater. In-situ natural attenuation, injecting domesticated degradation bacteria into contaminated sites or using indigenous microorganisms to degrade and remove pollutants.

[04] gas, aiming at the removal of volatile and semi-volatile organic pollutants in groundwater, the main treatment technologies are soil gas extraction, air disturbance and steam stripping in wells, etc. Their common feature is that the pollutants are transferred to the gas phase by means of volatilization and vaporization, and then discharged with the gas.

[05] Heat,refers to thermodynamic repair technology and thermal desorption repair technology. Thermodynamic remediation technology uses heat conduction, thermal blanket, thermal well or thermal wall, or thermal radiation and radio wave heating to realize remediation of contaminated soil. Thermal desorption remediation technology, which heats the soil polluted by organic matter above the boiling point of organic matter by heating, volatilizes the organic matter in adsorbed soil into gaseous state and then separates and treats it.

[06] Electricity, remove pollutants by electromigration, electroosmotic flow and electrophoresis. Electromigration is the movement of ions or complex ions to the opposite electrode, and the charged ions dissolved in groundwater are removed by this way; Electroosmosis is the directional movement of pore water in soil from one pole to the other in an electric field, and non-ionic pollutants are removed along with the movement of electroosmosis flow. Electrophoresis means that charged particles or colloids can be firmly adsorbed on movable particles under the action of DC electric field. Generally, an electrode well will be set, electrodes will be inserted, and an electric field will be formed after direct current is applied to treat pollutants.

[07] At present, the groundwater remediation methods are single, or only the combination of limited and limited remediation methods (mostly the combination of microbial remediation and other technologies), which has the problem of unsatisfactory treatment efficiency for pollution situations, especially compound pollution situations. Flexible adoption, reasonable planning and control of various restoration methods are worthy of attention.

[08] At the same time, the existing groundwater remediation technology is costly, time-consuming and inflexible in operation.

[09] Moreover, the groundwater remediation methods will change the flow movement of groundwater (external interference), and the best applicable remediation technology with good treatment effect and low cost under different groundwater pollution conditions needs a real-time control system and method for groundwater remediation under external interference, so as to obtain the optimized operation conditions of the remediation system in real time.

SUMMARY

[010] In order to solve the above problems in the prior art, the present invention provides an intelligent feedback real-time control system and method for hydrodynamic circulation under external interference.

[011] The purpose of the present invention is to provide the following technical scheme:

[012] (1) An intelligent feedback real-time control method of hydrodynamic circulation under external interference, which comprises:

[013] Si: establish a three-dimensional water-gas-heat-electricity multi-field coupled groundwater stochastic simulation model, and use the simulation model to obtain the concentration of pollutants in the remediation process;

[014] S2: Establish the fitting relationship between pollutant concentration and operating conditions of remediation system through deep learning method of artificial intelligence;

[015] S3, establishing an objective optimization model, taking the fitting relationship between operating conditions and pollutant concentration as a constraint condition, solving the optimal operating conditions of the remediation system under the dynamic change scenario, and controlling the optimal operating conditions of the remediation system in real time by feeding back parameters such as temperature, water level and conductivity monitored by groundwater indicators.

[016] Preferably, the three-dimensional water-gas-heat-electricity multi-field coupled groundwater stochastic simulation model is described by the following equation:

[017] n P (#Ck pJ% [I pk(C, i, -#OS,D, -V C, )] = Rk cOt /I1

[018] In the formula, k is the pollutant component; / represents mobile phase

such as water, gas and oil; Indicates soil porosity; k is the total concentration of

component k (volume of component k per unit pore volume); pk represents the density of component k [ML-3]; np represents the number of phases; Cki represents the

concentration (volume ratio) of component kin phase 1, Ulis darcy velocity of phase 1

[LT-1]; Si represents the saturation of phase / (volume of phase 1 per unit pore volume); Rk represents the total source/sink term of component k (volume of component k in unit

pore volume per unit time), D>, is diffusion tensor, a' is the phase 1 flow rate.

[019] Preferably, the objective function of the objective optimization model is cost minimization, and the specific formula is:

[020] objective function:

[021]

tk±P

MinJ=Min[ <p(x,(t),u(t),t)dt] u(t) u(t) tk

[022] constraint function:

[023]

FEq(x(t),u(t))= 0 s.t. IEq(x,(t),u(t))> 0

[024] In the formula,J is the cost; u(t)E 91 is the control factor,X,(t)E 9

output for the system, Eq,IEq is constrained by various equality or inequality, tk Is the

kth sampling time, P is the forecast period.

[025] (2) An intelligent feedback real-time control system for hydrodynamic circulation under external interference, which is used for implementing the control method described in (1) above, comprises:

[026] A three-dimensional water, gas, heat and power multi-field coupled groundwater stochastic simulation model is used to obtain the concentration of pollutants in the remediation process and establish the fitting relationship between the concentration of pollutants and the operating conditions of the remediation system;

[027] objective optimization model, which is used to solve the optimal operating conditions of the remediation system under dynamic change scenarios with the fitting relationship between operating conditions and pollutant concentrations as constraints.

[028] Compared with the prior art, the intelligent feedback real-time control system and method for hydrodynamic circulation under external interference provided by the invention has the following beneficial effects:

[029] Establish a bio-enhanced three-dimensional stochastic simulation model of groundwater coupled with water, gas, heat and power fields, and support the in-situ remediation process of groundwater with the simulation model; Through the in-situ remediation test of groundwater, the groundwater remediation process after the threshold of organic matter leakage and overflow is simulated, and the parameters are calibrated and verified through observation data, and the three-dimensional water, gas, heat and power multi-field coupled groundwater stochastic simulation model is continuously improved, which improves the accuracy and reliability of the model;

[030] Making the object of remediation, such as,running the remediation system with minimum cost, as an object to establish nonlinear discrete optimization model, and the optimal operation conditions of the repair system under dynamic change scenarios can be solved by genetic algorithm, and intelligent feedback can be realized by automatic monitoring system to control the optimal operation conditions of the repair system in real time.

BRIEF DESCRIPTION OF THE FIGURES

[031] In order to explain the embodiments of the present invention more clearly, the following will briefly introduce the drawings used in the description of the embodiments.

[032] Fig. 1 shows the flow diagram of the intelligent feedback real-time control method of hydrodynamic cycle under external interference in the present invention;

[033] Fig. 2 shows the flow diagram of intelligent feedback real-time control in the present invention;

[034] Fig. 3 shows the pilot system for simulating the remediation site of petroleum groundwater pollution in the embodiment;

[035] Fig. 4 shows the well placement position in the pilot system;

[036] Fig. 5 shows a plan view of the well position in the pilot system;

[037] Fig. 6 shows the concentration distribution of benzene on the 57th day of the experiment;

[038] Fig. 7 shows the comparison of verification results between well 5 and well 6;

[039] Fig. 8 shows the optimal extraction/injection rate during repair;

[040] Fig. 9 shows benzene removal rate during remediation;

[041] Fig. 10 shows the predicted remediation results of benzene from day 2 to day 22, showing the benzene concentrations on days 2, 6, 10, 14, 18 and 22;

[042] Fig. 11 shows the benzene concentration of 9 virtual wells from day 2 to day 22.

DESCRIPTION OF THE INVENTION

[043] In order to make the object, technical scheme and advantages of the present invention clearer, the present invention will be further described in detail with reference to specific embodiments and drawings.

[044] In order to solve the problem that the existing remediation technology has a single remediation method under the condition of pollution, especially compound pollution, which leads to unsatisfactory remediation effect, and at the same time meet the target requirements of managers (such as controlling pollution remediation costs on the premise of meeting the remediation level), the inventor based on pumping treatment technology as the most widely used groundwater pollution control technology and its existing advantages, It is determined to implement groundwater pollution remediation based on the flexible combination of pumping treatment technology and other external disturbances such as electric remediation. By establishing a three-dimensional groundwater remediation simulation model and an objective optimization model, the real-time regulation of the operation conditions of the remediation system under dynamic conditions is implemented.

[045] In this invention, pumping treatment technology is flexibly combined with other remediation technologies to implement groundwater pollution remediation. The reasons are as follows: a. Pumping treatment technology needs to set up pumping wells and injection wells to extract polluted groundwater and recharge treated groundwater, while other remediation technologies, such as air disturbance and electric remediation, need to set up corresponding injection wells or electrode wells, combining various methods, and one well is multi-purpose, which is conducive to reducing installation and maintenance costs; B. The combination of various remediation technologies, real-time control and flexible selection of remediation technologies can improve the efficiency of pollutant treatment and have a wider application range for pollutants; C. allow a variety of remediation technologies to exist, which can meet the choice of remediation methods more suitable for groundwater treatment at different time periods, and at the same time, help meet the target requirements of managers.

[046] As shown in fig. 1, the intelligent feedback real-time control method of hydrodynamic circulation under external interference provided by the present invention includes:

[047] SI: Establish a three-dimensional water, gas, heat and power multi-field coupled groundwater stochastic simulation model, and use this simulation model to obtain the concentration of pollutants in the remediation process;

[048] S2: Establish the fitting relationship between pollutant concentration and operating conditions of remediation system by deep learning method of artificial intelligence;

[049] S3: Establish an objective optimization model, take the fitting relationship between operating conditions and pollutant concentration as constraint conditions, solve the optimal operating conditions of remediation system under dynamic change scenarios, and control the optimal operating conditions of remediation system in real time through feedback of parameters such as temperature, water level and conductivity monitored by groundwater indicators.

[050] In the present invention, foreign interference force refers to the operation of various repair methods of water, gas, heat and power, because the operation of various repair methods will have an impact on the water flow field.

[051] Groundwater exists in the soil or rock gap below the ground. Considering the underground environment and pollutant sources, pollutants exist not only in water phase, but also in gas phase and oil phase. The pollutant remediation process can be described by a 3D groundwater stochastic simulation model coupled with water, gas, heat and power fields, and the model supports the in-situ remediation process of groundwater. According to certain assumptions, the model is mainly described by the mass conservation equation in formula (1):

[052]

nP

(#Cp)+V [L p,(C,,ii, -- SD, - VC )]=Rk t =1 (1)

[053] In the formula, k is the pollutant component;/represents mobile phase such

as water, gas and oil; # Indicates soil porosity, Ck is the total concentration of component k (volume of component k per unit pore volume), represents the density of component k [ML-3 ]; np represents the number of phases; Ckirepresents the

concentration (volume ratio) of component kin phase 1, Uis darcy velocity of phase 1

[LT-1]; S represents the saturation of phase / (volume of phase L per unit pore volume); Rkrepresents the total source/sink term of component k (volume of component k in unit

pore volume per unit time),f>, is diffusion tensor. IIs the phase I flow rate, It is

closely related to the remediation effect and the operating conditions of the remediation

system, such as when L is water phase, UIis related to water flow, and then limits the flow rate of pumping groundwater and recharging groundwater , In the present

invention, U is can be ccalculated through the multiphase form of Darcy's theorem:

[054]

krk U, =- -(ri P - p 1gVz) i (2)

[055] In the formula, kri is the permeability of porous media relative to phase 1; k Is the intrinsic permeability tensor [L 2 ]; p, is the viscosity of phase 1 [ML-2 T-1]; pi is the density of phase / [ML-3 ]; g is the acceleration of gravity [LT-2 ]; z is the vertical distance, and it is defined as positivedown [L] ; Pi is the pressure of phase l[ML-T-2].

[056] The three-dimensional multi-field simulation model adopted in the invention considers the migration of pollutants in three directions, and compared with the two-dimensional simulation model in the prior art, the migration simulation of pollutants is more accurate and conforms to the actual migration law.

[057] In one embodiment, the application of bioaugmentation technology can be used to expand the soil porosity in underground polluted areas and increase the in-situ remediation radius of microorganisms; Efficient transportation of microorganisms and nutrients, effective use of degrading bacteria, avoiding the spread of pollution and deep degradation of organic pollutants; Shorten microbial remediation cycle and greatly improve remediation efficiency.Use preferred microbial remedication technology as a necessary technology for groundwater pollution remediation.

[058] In aquifer, biodegradable organic components can dissolve into water and become the matrix of free bacteria in water. Most of this matrix can be removed from water along with attached biomass through biodegradation reaction.

[059] The above formula and mass conservation equation are coupled into the three-dimensional simulation model to correct the groundwater pollutant level during remediation.

[060] Considering the high complexity and dynamic change process of the remediation system, some important information may be missed when establishing the simulation model. In this invention, the groundwater remediation process after the threshold of organic matter leakage and overflow is simulated through the small-scale experiment of groundwater in-situ remediation, and the parameters are calibrated and verified through the observed data, so as to continuously improve the 3D water, gas, heat and electricity multi-field coupled groundwater stochastic simulation model and improve the accuracy and reliability of the model. Among them, the parameters for calibration and verification are the concentration of leaked pollutants, chemical oxygen demand (COD), biochemical oxygen demand (BOD), dissolved organic carbon (DOC), etc. The values of these parameters have different choices according to different groundwater pollution conditions and remediation operation conditions.

[061] In the invention, in the small-scale experiment of in-situ groundwater remediation, a deep learning method of artificial intelligence is adopted to simulate the remediation effect of the groundwater remediation system under the dynamic change scenario. By analyzing a large number of simulation results, the relationship between pollutant concentration and remediation operation conditions was established. The simulation process is based on specific pollution condition,The equation X=G(U) is used to describe the concrete process of in-situ remediation of groundwater ,

X =(XIx 2, --- x n )..is fixed parameter, U=(uIu2 " --- U) is a variable parameter, U=U

The deep learning method of artificial intelligence takes the corresponding pollutant remediation concentration under different operating conditions as the dependent variable X, and the operating conditions and initial pollution conditions are the independent variable U, and the relationship is:

[062]

U (t f(X )) ( 9)

[063] The FCI simulation process is a deep learning method based on artificial intelligence, which is mainly used to simulate the remediation process of groundwater pollution under external interference, and can be expressed by the following equation:

[064]

X (t +1)= f (X (t),U (t)) (10)

[065] The above-mentioned operating conditions include the cooperative use mode and sequence of remediation technologies and the operating parameters of each remediation technology under the current pollution situation. For example, under the current groundwater pollution situation, select the combination of extraction treatment technology and in-situ microbial remediation, open and set the extraction wells injection wells in Group A, determine the pumping/injection rate and other parameters, and after repairing for B days, change to electric remediation technology, open and set the electrode wells in Group C, set the voltage intensity and other parameters, and days of repairing.

[066] In the present invention, the fitting relationship between pollutant concentration and operating conditions of remediation system is established through the deep learning method of artificial intelligence, and the influence of operating conditions of remediation system on pollutant concentration at key positions under dynamic change scenarios is designed to simulate the in-situ groundwater remediation process under external interference. Through the above operations, it is conducive to the selection of operating conditions under the actual dynamic change scenario; And when various operating conditions meet the pollution control requirements, select the operating conditions that better meet the needs of operators (such as the lowest cost or the shortest time, etc).

[067] In the process of restoration, the restoration system needs to meet the changing hydrogeological conditions and human factors. Therefore, it is necessary to develop an intelligent feedback real-time control system to ensure that the concentration of pollutants in groundwater meets the prescribed groundwater environmental quality standards. The control system must satisfy a series of constraints. For example, the suction pump must maintain a certain effective suction head, the storage tank should not overflow or empty, and the maximum extraction rate of the pump must be used at its rated power. The above requirements indicate that it is necessary to realize the control objectives with the help of manual participation (designers and operators) and reasonable configuration of instruments and equipment (sensors, regulating valves, controllers and computers) in the repair process. Generally speaking, the basic requirements that the control system should meet include: restraining the influence of external disturbance, ensuring the stability in the process of technical operation, and optimizing the working conditions in the repair process in real time.

[068] As shown in Figure 2, X(t), the initially set pollutant remediation concentration, serves as the input condition of 3D water, gas, heat and power multi field coupling, and U(t), the initial condition of contaminated site, plays a negative role in regulating multi-field coupling remediation, and Xr (t+1) plays a positive role in

regulating 3D coupling site for the input of optimal control of remediation system. U'(t) is the pollution producing area after multi-field coupling remediation, and X(t) is used as the input condition of in-situ bioaugmentation remediation simulation predictor and remediation regression analysis for groundwater remediation under the dynamic change situation, and the output of regression analysis is used for feedback adjustment of remediation system. Xp(t+l) is finally output for deep learning of artificial intelligence.

[069] In the present invention,X, X, xand x have the same meanings. In the

present invention, the control system is realized by the objective optimization model. When the control demand is to minimize the operation cost of the repair system, J is the cost, and the nonlinear discrete optimization model of formula (11) is established, and the optimal operation condition of the repair system under the dynamic change scenario is solved by genetic algorithm. The specific formula is:

[070] objective function:

[071]

tk±P

MinJ=Min[ f(xn(t),u(t),t)dt] ( 1) u(t) u(t) tk

[072] Constraint function:

[073]

FEq(x(t),u(t)) 0(12) s.t-{IEq(x,(t),u(t))>0

[074] In the formula, u(t)e 91 is the control factor, Xn(t)E output for the

system, Eq,IEq is constrained by various equality or inequality,tk isthe kth sampling

time , P is the forecast period.The objective function can take different forms according to different control requirements.

[075] If the constraint is linear, the nonlinear optimization in this model can be transformed into a quadratic optimization problem (QP). Specifically, the above QP problem can be transformed into:

[076]

N2 N

J= [x,(t + i)-x(t + i)]2 + A(i)[Au(t + i)] (13) i=Ni i=O

[077]

umin < u(t + i) <Umax (i = 0, ... NP ,I xm- i (t + i) Xma (i =N 1 ,..., N2 )

Au(t + i) 0Auma (i= 0N,-- N) 14)

Au(t +i)=-0 (i> N, - 1)

[078] Wherein, X,(t+i) is the standard reference value, xt+i) is the model

predicted value, and Au(t + i) is the control increment, u(t + i) - u(t + i - 1) is

defined as the input weight factor, NI, N 2 is the minimum and maximum prediction

N period, N is the control period, and u(t+i) is the expected output.To solve the above models, heuristic modem optimization algorithms, such as genetic algorithms, should be applied.

[079] Because the groundwater pollution changes with the remediation process, the automatic monitoring system (that is, the collection of equipment used to detect groundwater) can monitor and intelligently feedback the groundwater situation, and the objective optimization model can be used to control the optimal operating conditions of the remediation system in real time.

[080] Another aspect of the present invention is to provide an intelligent feedback real-time control system for hydrodynamic circulation under external interference, which is used to implement the intelligent feedback real-time control method for hydrodynamic circulation under external interference.The control system includes:

[081] A three-dimensional water, gas, heat and power multi-field coupled groundwater stochastic simulation model is used to obtain the concentration of pollutants in the remediation process and establish the fitting relationship between the concentration of pollutants and the operating conditions of the remediation system;

[082] objective optimization model, which is used to solve the optimal operating conditions of the remediation system under dynamic change scenarios with the fitting relationship between operating conditions and pollutant concentrations as constraints.

[083] Among them, the 3D groundwater stochastic simulation model with multi field coupling of water, gas, heat and electricity is described by the following equation:

[084]

(#Cp)+V [ZL p,(Cuii, -- SD -VC )] = R at 1=1

[085] In the formula, k is the pollutant component; 1 stands for mobile phase such

as water, gas and oil, < indicates soil porosity, C, is the total concentration of

component k (volume of component k per unit pore volume); pk represents the density of component k [ML-3 ]; np represents the number of phases; Cl represents the

concentration (volume ratio) of component k in phase 1,uris darcy velocity of phase 1

[LT-1],Si represents the saturation of phase / (volume of phase I per unit pore volume); Rkrepresents the total source/sink term of component k (volume of component k in unit

pore volume per unit time); D,, is diffusion tensor, and is flow velocity of phase 1.

[086] the objective function of the objective optimization model is cost minimization, and the specific formula is:

[087] objective function:

[088]

tk±P

MinJ= Min[ f(P(x,(t),u(t),t)dt] u(t) u(t) tk

[089] constraint function:

[090]

Eq(x.(t),u(t)= 0 s.t. IEq(x,(t),u(t)) 0

[091] In the formula, J is the cost; U(t)E 91 is the control factor, X,(t)E 91

kth sampling time, P is the forecast period.

[092] That is to say, the intelligent feedback real-time control system can dynamically adjust the remediation conditions according to the basic parameters of groundwater, such as groundwater level, conductivity, soil porosity and human factors.

[093] Embodiment

[094] The above method was applied to the remediation site of petroleum groundwater pollution in an oilfield. In order to simulate the underground aquifer under natural conditions, a pilot-scale system (reactor) is proposed. The system is cuboid, and its size is LxWxH=3.6x1.2x1.0 m 3 (see Figure 3). The reactor is divided into four sections, each section of which is equipped with a certain number of sampling holes (monitoring wells) arranged in a square array. Sampler enters the reactor from the sampling hole to obtain soil and water samples at different positions and depths. Sampler is made of stainless steel and used to collect soil and underground water samples. The sampler can move vertically, so that samples with different depths and positions can be obtained. Samplers are divided into two types, which are used to collect soil samples and water samples respectively. Organic matter concentration is determined by Varian CP-3800 gas chromatograph (GC), which is controlled by microcomputer and used to analyze the content of organic pollutants in gas phase (in soil unsaturated zone) and liquid phase (in water sample).

[095] According to the characteristics of soil profile, migration of organic pollutants and movement of pollutant plume, the concentration of benzene in six wells (No.5, No.7, No.8, No.10, No.11 and No.12) is selected to represent groundwater 0 00 0 0 0 pollution (the concentration is expressed as x, x , x3, x, x, fnx 6 ). In order to reflect

as much pollution as possible, the concentration range of benzene is set to vary greatly, with the maximum concentration of 30 mg/L and the minimum concentration of Omg/L.. Within this range, 50 concentration levels are randomly generated for each related well (monitoring well), and then 50 pollution scenarios are set.

[096] Fig. 4 shows the location of well arrangement (according to the measurement of pollutants in the contaminated site and the actual geological conditions), (and the location of the hypothetical well in the simulation process); Fig. 5 shows the well location (plan view) (the wells monitored in the third and fourth monitoring stages during layered monitoring). Circulating nutrients and oxygen in polluted aquifer are transported by underground water pump system. The process involves two parts: introducing aerated water and water rich in nutrients and biomass into the polluted area through two injection wells; The water with descending gradient is recovered through two extraction wells. Circulate through polluted areas to ensure the mixing and close contact between oxygen, nutrients, pollutants and microorganisms. Therefore, pumping/injection rate directly affects pollutant removal efficiency and system operation cost. The pumping/injection rate of the selected well is determined as the main control condition. Based on the soil porosity and permeability in the remediation system, the scope of pumping/injection rate is determined by verifying the developed bioaugmentation groundwater remediation model. The maximum flow rate is set to 40L/d, while the minimum flow rate is set to 1OL/d .. The biomass, oxygen and nutrient concentrations in the injected fluid are respectively 20, 8 and 1500(mg/L). There are 50 operating conditions randomly generated, and the related control variables are represented as ui (injection rate of well I, L/d), U2 (injection rate of well ii, L/d), U3 (extraction rate of well III, L/d) and U4 (extraction rate of well IV, L/d).

[097] The combination of 50 pollution level scenarios and 50 operating condition scenarios produced 2500 scenarios. Accordingly, the bio-enhanced groundwater remediation model of organic pollution generated 2500 dynamic scenarios. The results showed that the benzene concentration in groundwater decreased significantly 18 days after remediation. The repair time was set at 22 days and divided into 112-day cycles.

For each pollution level scenario (X X ' 'X' X4 5 3 6 ), 50 sets of data about benzene concentration removal percentage (9) at specific locations and operating conditions of bio-enhanced in-situ remediation of groundwater can be obtained from

the simulation operation (U, u2, U 3 , RU 4 ). In the regression analysis system, through the deep learning method of artificial intelligence, the bio-enhanced groundwater in situ remediation model is calibrated and verified based on the experimental data. The absolute error between simulated concentration and observed concentration ranged from 0.08 to 0.85 mg/L, with an average value of 0.36 mg/L.. The mean square error is 0.47 mg/L, and the correlation coefficient is 0.93. Figure 6 shows the verification results on the 57th day. Fig. 7 shows the verification results of benzene concentration changes with time in wells 5 and 6. After calibration and verification, this simulation model can be used to study the effect of different bioremediation on benzene concentration. The initial pollutant concentration distribution on the 57th day is taken as the initial condition of the system. Fig. 8 shows the best operating conditions in 11 time periods. Fig. 9 shows the removal rate of benzene during remediation. It was found that there were two plateau periods in the restoration process (12th to 16th day, 18th to 22nd day).

[098] By the end of the monitoring period, the remediation site has reached the cleaning target, which means that the benzene concentration anywhere in the simulation domain has dropped below 300pg/L, and the removal rate is 93%. Fig. 10 shows the predicted remediation results of DPC system (groundwater system in benzene polluted area) from day 2 to day 22. The results show that the pollution level has been significantly reduced during the remediation process. In order to reflect the efficiency of the remediation system, nine virtual wells (HW) were selected from the simulation domain. Fig. 11 shows the benzene concentration of 9 virtual wells from the 2nd day to the 22nd day. The analysis of the predicted data shows that the benzene concentration decreases slowly or even increases in some places 10 days before remediation, and the pumping/injection rate also increases correspondingly. The signal of increasing pollutant concentration triggers the necessary adjustment of process control operation. After 10 days of operation, the concentration of pollutants in most locations decreased, and the pumping/injection rate correspondingly decreased from the 12th day.

[099] Although the invention has been described with reference to specific examples, it will be appreciated by those skilled in the art that the invention may be embodied in many other forms, in keeping with the broad principles and the spirit of the invention described herein.

[0100] The present invention and the described embodiments specifically include the best method known to the applicant of performing the invention. The present invention and the described preferred embodiments specifically include at least one feature that is industrially applicable

Claims

THE CLAIMS DEFINING THE INVENTION ARE AS FOLLOWS:

1. An intelligent feedback real-time control method of hydrodynamic circulation under external interference, characterized by comprising:

SI: Establish a three-dimensional water, gas, heat and electricity multi-field coupled groundwater stochastic simulation model, and use this simulation model to obtain the concentration of pollutants in the remediation process;

S2: Establish the fitting relationship between pollutant concentration and operation conditions of remediation system through deep learning method of artificial intelligence;

S3, establishing an objective optimization model, taking the fitting relationship between operating conditions and pollutant concentration as a constraint condition, solving the optimal operating conditions of the remediation system under the dynamic change scenario, and controlling the optimal operating conditions of the remediation system in real time by feeding back parameters such as temperature, water level and conductivity monitored by groundwater indicators.

2. The method for intelligent feedback real-time control of hydrodynamic circulation under external disturbance according to claim 1, which is characterized in that in step Sl, the three-dimensional stochastic simulation model of groundwater coupled with water, gas, heat and power fields is described by the following equation:

(#C p)V - [ pk(Cld, -#S1Dk1 -V Cl ] = Rk

In the formula,k is the pollutant component; 1 represents mobile phase such as

water, gas and oil; # indicates soil porosity, Ckis the total concentration of component k (volume of component k per unit pore volume), represents the density of component k [ML-3 ]; rp represents the number of phases; Ckirepresents the concentration (volume

ratio) of component kin phase 1, ilis darcy velocity of phase 1 [LT-1]; Si represents the saturation of phasel (volume of phase 1 per unit pore volume); Rkrepresents the total source/sink term of component k (volume of component k in unit pore volume per unit time), >, is diffusion tensor, Is the phase 1 flow rate.

3. The intelligent feedback real-time control method of hydrodynamic circulation

under external disturbance according to claim 2 is characterized in that, 1 can be calculated by the following equation:

_kk

U,=- -(VP-pgVz) il

In the formula,kri is the permeability of porous media relative to phase 1; k Is

the intrinsic permeability tensor [L2 ]; p, is the viscosity of phase 1 [ML- 2 T- 1]; pi is the density of phase 1 [ML-3 ]; g is the acceleration of gravity [LT-2 ]; z is the vertical distance, and it is defined as positive down [L] ; Pi is the pressure of phase 1 [ML-T-2].

4. The intelligent feedback real-time controlmethod of hydrodynamic circulation under external disturbance according to claim 1, characterized in that in step S2, the deep learning method of artificial intelligence takes the corresponding pollutant remediation concentration under different operating conditions as the dependent variable, i.e. X, and the operating conditions and initial pollution conditions as the independent variable, i.e. U, and establishes a fitting relationship between the pollutant concentration and the operating conditions of the remediation system, which is as follows:

U(t)= f(X(t))

The FCI simulation process is as follows

x(t+)=f(x(t),U(t))

5. The method for intelligent feedback real-time control of hydrodynamic circulation under external disturbance according to claim 1, which is characterized in that step S2 further comprises: simulating the groundwater remediation process after the pollutant leakage and overflow threshold through an in-situ groundwater remediation test, and calibrating and verifying parameters in a three-dimensional water, gas, heat and power multi-field coupled groundwater stochastic simulation model through observation data.

6. The intelligent feedback real-time control method of hydrodynamic cycle under external disturbance according to claim 1, characterized in that in step S3, the objective function of the objective optimization model is cost minimization, and the specific formula is:

objective function:

tk±P

MinJ=Min[ go(xn(t),u(t),t)dt] u(t) u(t) tk

constraint function:

FEq(x(t),u(t))= 0 S.t.IEq(x, (t),u(t)) > 0

In the formula,J is the cost; U(t)e 91 is the control factor, Xn (t) G output

for the system, Eq,IEq is constrained by various equality or inequality,tk Is the kth

sampling time, P is the forecast period.

7. The intelligent feedback real-time control method of hydrodynamic cycle under external disturbance according to claim 6, characterized in that when the limit is linear, the formula in the objective optimization model is as follows:

N2 N, =) (0(i)[Au(t i)] i=N, i=O

Umin u(t + i) Umax (i= 0,..., N

) Xmin X (t +i ) XmX (i =N1 ,..., N 2

) s.t. |Au(t +i) Aum (i 0,..., N, Au(t +i) =0 (i> N, - 1)

Wherein, X,(t+i) is the standard reference value, x,t+i) is the model

predicted value, and Au(t + i) is the control increment, u(t + i) - u(t + i - 1) , A

is defined as the input weight factor, NI, N 2 is the minimum and maximum prediction

N period, N is the control period, and u(t+i) is the expected output.

8. An intelligent feedback real-time control system for hydrodynamic circulation under external disturbance, which is used to implement the control method of claim 1 7, characterized in that the control system comprises:

A three-dimensional water, gas, heat and power multi-field coupled groundwater stochastic simulation model is used to obtain the concentration of pollutants in the remediation process and establish the fitting relationship between the concentration of pollutants and the operating conditions of the remediation system;

The objective optimization model is used to solve the optimal operating conditions of the remediation system under dynamic change scenarios with the fitting relationship between operating conditions and pollutant concentrations as constraints.

9. The intelligent feedback real-time control system of hydrodynamic circulation under external disturbance according to claim 8, which is characterized in that the three dimensional water-gas-heat-electricity multi-field coupled groundwater stochastic simulation model is described by the following equation:

n _:

at (#Ckpk [±V pk(Cu 1=1 -#SDd -V Cld)]= R

In the formula, k is the pollutant component; 1 stands for mobile phase such as

water, gas and oil, < indicates soil porosity, k is the total concentration of component

k (volume of component k per unit pore volume); pk represents the density of component k [ML-3]; Np represents the number of phases; Cirepresents the concentration (volume

ratio) of component k in phase 1; i 1is darcy velocity of phase 1 [LT-1]; Si represents

the saturation of phase 1 (volume of phase I per unit pore volume); Rkrepresents the total source/sink term of component k (volume of component k in unit pore volume per

unit time); >, is diffusion tensor, and Uis flow velocity of phase 1.

1 / 11

Fig. 1

2 / 11 16 Oct 2020 2020102748

Fig. 2 Design flow of nonlinear predictive controller based on simulation, prediction and multivariate analysis

Fig. 3

3 / 11

Fig. 5 Fig. 4

4 / 11

Fig. 6

5 / 11

Fig. 7

6 / 11

Fig. 9 Fig. 8

7 / 11 16 Oct 2020 2020102748

Direction of water

Fig. 10

8 / 11

9 / 11

10 / 11

11 / 11

Fig. 11