Control system and method for optimizing deep well injection parameters of high-salinity mine water

Abstract

The invention discloses a control system and method for optimizing high-salinity mine water deep well injection parameters. The system comprises a data acquisition unit, a control unit, an execution unit and an environment monitoring module. According to the method, injection parameters are monitored and dynamically adjusted in real time through data acquisition, data processing and analysis, optimization and calculation, execution and adjustment, environment monitoring and protection, remote monitoring and management and the like. According to the invention, real-time data analysis, dynamic optimization and remote monitoring functions are combined, the efficiency and safety of mine water treatment are remarkably improved, and negative effects on the environment are reduced.

Description

Technical Field

[0001] The present invention belongs to the field of water resource management, and relates to a control system and method for optimizing deep well water injection parameters in high-salinity mines, and specifically to an adaptive control system and method for optimizing deep well water injection parameters in high-salinity mines and alleviating environmental impacts. Background Art

[0002] With the increase in mining activities, the discharge of high-salinity mine water has become an increasingly serious environmental problem. These mine waters not only contain a large amount of salt, but may also contain other harmful substances. Direct discharge will cause serious pollution to surface water and soil. Traditional mine water treatment methods such as evaporation ponds and surface discharge have problems such as low treatment efficiency, large footprint and high environmental risks. Deep well injection, as an effective mine water treatment method, has the advantages of small footprint, large treatment volume and small impact on the environment. However, how to optimize the parameters of deep well injection to ensure the efficient operation of the system and reduce environmental impact is still a problem that needs to be solved. Summary of the invention

[0003] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a control system and method for optimizing deep well injection parameters of high-salinity mine water. By real-time monitoring and dynamic adjustment of injection parameters, the efficiency of deep well injection can be effectively improved and potential harm to the environment can be reduced.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solutions to achieve the above problems:

[0005] A control system for optimizing injection parameters of deep well water in high-salinity mines, comprising:

[0006] Data acquisition unit, including multiple sensors and detection equipment, used to monitor various parameters of mine water in real time;

[0007] The control unit can receive real-time data from the data acquisition unit, analyze and process the data, and calculate the best injection parameters according to the preset optimization algorithm;

[0008] The execution unit, including a variety of actuators and control devices, can automatically adjust the operating status of the injection equipment according to the instructions of the control unit to ensure that the mine water is injected into the ground in the optimal state;

[0009] The environmental monitoring module, including groundwater quality monitoring instruments and data transmission equipment, can monitor the groundwater quality in the injection area and promptly detect and prevent possible pollution problems.

[0010] The present invention also includes the following technical features:

[0011] Specifically, the multiple sensors and detection equipment include a salinity sensor, a temperature sensor, a pressure sensor and a flow meter;

[0012] The various actuators and control devices include variable frequency pumps, regulating valves and injection equipment.

[0013] A method for optimizing water injection parameters in a high-salinity mine deep well, the method is implemented based on the control system for optimizing water injection parameters in a high-salinity mine deep well; the method comprises the following steps:

[0014] Step 1: Data collection:

[0015] The salinity, temperature, pressure and flow data of the mine water are collected in real time through the salinity sensor, temperature sensor, pressure sensor and flow meter, and the collected data are transmitted to the control unit;

[0016] Step 2: Data processing and analysis:

[0017] Data preprocessing: After receiving the data, the control unit performs data cleaning and normalization to ensure data accuracy; Real-time data analysis: Perform multi-dimensional analysis of the data to identify key factors that affect the injection effect;

[0018] Step 3: Optimization and calculation:

[0019] Based on the key variables affecting the current injection effect obtained in step 2, enter the optimization stage and dynamically adjust the injection parameters to ensure that the system can achieve a balance between injection efficiency, environmental protection and safety;

[0020] Step 4: Execution and Adjustment:

[0021] Sending instructions: The control unit sends a control signal to the execution unit according to the optimization result in step 3; Adjusting equipment: After receiving the instructions, the execution unit dynamically adjusts the injection pressure and flow rate by adjusting the variable frequency pump and the regulating valve to achieve the optimal state after optimization;

[0022] Step 5: Environmental monitoring and protection:

[0023] Monitoring of groundwater quality: The environmental monitoring module monitors the pH value and heavy metal content of groundwater in real time; Pollution warning: When abnormal groundwater quality is detected, the system automatically adjusts the injection parameters to reduce pollution risks and notifies management personnel through the alarm system;

[0024] Step 6: Remote monitoring and management:

[0025] Real-time monitoring platform: provides a user interface to display the system's operating status, historical data and forecast results; remote operation: users can remotely view the system status through the network and make necessary parameter adjustments and emergency processing.

[0026] Specifically, in step 2, data processing and analysis uses LASSO regression to perform preliminary feature screening to screen out the key parameters that have the most significant impact on the injection effect; uses principal component analysis to reduce the dimension and merges two or more related variables into principal components; then uses random forest regression to analyze the nonlinear relationship between the principal components to obtain the contribution of each principal component to the injection effect.

[0027] Specifically, the data processing and analysis includes:

[0028] Step 2.2.1, use LASSO regression for feature selection. The objective function of LASSO regression is:

[0029]

[0030] Among them, X ij represents the jth characteristic value of the i-th sample, including salinity, temperature, and pressure; y i is the injection effect; β j is the regression coefficient of each parameter; λ is the regularization parameter;

[0031] In the regression process, by adjusting λ, the parameters with little influence on the injection effect are automatically screened out and their coefficients are thinned to zero, so that the key parameters with the most significant influence on the injection effect are obtained among the various parameters of mine water;

[0032] Step 2.2.2, principal component analysis for dimensionality reduction: Use principal component analysis to identify the correlation between the key parameters obtained in step 2.2.1, and generate multiple new principal components to represent the main change trends of these key parameter variables;

[0033] Step 2.2.3, random forest regression analysis: Use the random forest model to perform regression analysis on the principal components after PCA dimensionality reduction to obtain the contribution of each principal component to the injection effect.

[0034] Specifically, the step 2.2.3 includes:

[0035] Input variables: The principal components generated by principal component analysis are used as input variables; Model training: The principal components generated by PCA obtained at each moment in the past are trained with the injection effect, and random forest constructs multiple decision trees to capture the complex nonlinear relationship between the principal components and the injection effect; Prediction output: After the model is trained, new injection operations are predicted. By inputting new principal component values ​​in real time, the random forest regression model can accurately evaluate whether the current injection conditions are optimized and whether the injection parameters need to be adjusted; Variable importance analysis: Random forest can also output the importance score of each principal component to obtain the contribution of each variable to the injection effect.

[0036] Specifically, the step three includes:

[0037] Step 3.1, optimization objective function setting:

[0038] Based on the contribution of each variable to the injection effect obtained in step 2, and to ensure a balance between injection efficiency, environmental protection and safety, the objective function is defined as follows:

[0039] f(P,Q)=α(t)·E(P,Q)-β(t)·G(S,T)-γ(t)·S(P,Q)

[0040] Among them, injection efficiency E(P,Q)=Q: means that the injection efficiency is proportional to the flow rate Q; environmental impact G(S,T)=k1S+k2T: means the impact of salinity S and temperature T on the environment, k1 means the weight of the impact of salinity on the environment, k2 means the weight of the impact of temperature on the environment; system safety S(P,Q)=(P / Pmax) 2 +(Q / Qmax) 2 : Ensure that the injection pressure P and flow rate Q are within the safe range, Pmax and Qmax are the safety thresholds of pressure and flow rate respectively; α(t) is the weight of injection efficiency, β(t) is the weight of environmental impact, and γ(t) is the weight of system safety;

[0041] Step 3.2, dynamic adjustment and feedback mechanism:

[0042] The Bayesian optimization algorithm is used in combination with real-time data feedback to dynamically adjust the injection parameters to continuously optimize the injection process; this is achieved through the following steps: Real-time feedback and iterative adjustment: real-time monitoring of pressure, flow, salinity and temperature through sensors, and comparison of the actual injection effect with the objective function; Dynamic adjustment of weight coefficients: through real-time feedback, the optimization algorithm adjusts the weight coefficients α(t), β(t), γ(t) in the objective function.

[0043] Specifically, in step 5, when the environmental monitoring module detects that the groundwater quality index exceeds the standard, feedback and adjustment are performed through the following steps:

[0044] Step 5.1, monitoring and alarm: The environmental monitoring module monitors the groundwater quality in real time; if a certain indicator is detected to exceed the set threshold, the system triggers an alarm;

[0045] Step 5.2, data feedback: the environmental monitoring module transmits the exceeding data and alarm information to the control unit; after receiving the feedback, the control unit re-evaluates the current injection strategy;

[0046] Step 5.3, adjust the objective function: adjust the weight coefficient in the objective function and introduce the environmental penalty term F env , and update the objective function as:

[0047] f′(P,Q)=α(t)·E(P,Q)-β(t)·G(S,T)-γ(t)·S(P,Q)-δ(t)·F env

[0048] Among them, F env represents the environmental penalty, δ is the weight coefficient of environmental excess;

[0049] Step 5.4, re-optimize injection parameters: the control unit uses the updated objective function f′(P, Q) to recalculate the optimal injection pressure P and injection flow rate Q; through the optimization algorithm, find new P and Q to maximize the objective function f′(P, Q) while reducing the impact of environmental exceeding the standard;

[0050] Step 5.5, implement adjustments: apply the new injection pressure and flow parameters to the injection system and adjust the operation of the injection equipment in real time; continue to monitor groundwater quality to ensure that the adjustment measures are effective until environmental indicators return to normal range.

[0051] Specifically, when the environmental monitoring module detects that the pH value exceeds the standard, the environmental penalty item F env Defined as:

[0052] F env =k pH ·(pH current -pH target ) 2

[0053] Among them, k pH is the penalty coefficient for exceeding the pH value limit, pH current is the currently monitored pH value, pH target is the target pH.

[0054] Specifically, when the environmental monitoring module detects that heavy metals exceed the standard, the environmental penalty item F env Defined as:

[0055] F env =k HM ·(HM current -HM target ) 2

[0056] Among them, k HM is the penalty coefficient for excessive heavy metal content, HM current is the currently monitored heavy metal content, HM target is the target heavy metal content.

[0057] Compared with the prior art, the present invention has the following technical effects:

[0058] The present invention provides an intelligent control system for deep well injection of mine water, which combines real-time data analysis, dynamic optimization and remote monitoring functions, significantly improves the efficiency and safety of mine water treatment, and reduces the negative impact on the environment.

[0059] The present invention uses LASSO regression, PCA dimensionality reduction and random forest regression to perform multi-dimensional analysis on real-time monitoring data, screen out key variables that affect the injection effect, and dynamically adjust the injection parameters through the Bayesian optimization algorithm. This optimization process ensures that the system can respond flexibly under different working conditions, always maintains maximum injection efficiency, and controls environmental impacts and ensures system safety. Compared with the traditional fixed injection parameter method, the intelligent optimization control of the present invention greatly improves the accuracy and adaptability of the injection process.

[0060] The present invention effectively controls environmental factors during the injection process by setting penalty coefficients. In particular, for water bodies with high salinity, excessive heavy metals, and substandard pH values, the system can automatically adjust parameters to reduce the risk of contamination to groundwater and the surrounding environment. At the same time, the system makes real-time adjustments based on the monitored pressure and flow to ensure that the equipment operates within a safe range and effectively prevent system overload or failure.

[0061] The present invention integrates a real-time monitoring platform and remote operation functions, and users can view the system's operating status, historical data, and prediction results through the network platform. The monitoring platform provides an intuitive user interface, which is convenient for real-time grasp of the system's operating conditions. In addition, the system supports remote parameter adjustment, and users can perform necessary optimization operations and emergency treatments based on the monitoring results to ensure stable operation of the system under different working conditions. This remote control method significantly improves management efficiency, reduces the need for manual intervention, and maintains efficient and reliable operation during the process.

[0062] In addition to being applicable to the treatment of high-salinity mine water, the system architecture and optimization algorithm of the present invention have strong adaptability and can be extended to other similar industrial wastewater treatment scenarios. By adjusting system parameters, it can effectively cope with complex conditions of different water qualities and different treatment requirements, providing an effective solution for other wastewater treatment fields and having broad application prospects.

[0063] The present invention breaks through the limitations of traditional mine water injection systems by combining intelligent control with optimization algorithms, realizes high efficiency, environmental protection, safety and intelligence in the injection process, and has significant technical advantages and practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0065] The present invention provides a control system for optimizing injection parameters of deep well water in high-salinity mines, comprising:

[0066] A data acquisition unit, including multiple sensors and detection equipment, is used to monitor various parameters of mine water in real time; specifically, the multiple sensors and detection equipment include salinity sensors, temperature sensors, pressure sensors and flow meters;

[0067] The control unit, whose core is a high-performance processor, is equipped with advanced data processing algorithms and control logic. It can receive real-time data from the data acquisition unit, analyze and process the data, and calculate the best injection parameters according to the preset optimization algorithm;

[0068] The execution unit includes a variety of actuators and control devices, which can automatically adjust the operating state of the injection equipment according to the instructions of the control unit to ensure that the mine water is injected into the ground in the optimal state; specifically, the various actuators and control devices include variable frequency pumps, regulating valves and injection equipment;

[0069] The environmental monitoring module, including groundwater quality monitoring instruments and data transmission equipment, can monitor the groundwater quality in the injection area and promptly detect and prevent possible pollution problems.

[0070] The present invention also provides a method for optimizing the injection parameters of deep well water in high-salinity mines, and the method is implemented based on a control system for optimizing the injection parameters of deep well water in high-salinity mines; Figure 1 As shown, the following steps are included:

[0071] Step 1: Data collection:

[0072] The salinity, temperature, pressure and flow data of the mine water are collected in real time through the salinity sensor, temperature sensor, pressure sensor and flow meter, and the collected data are transmitted to the control unit;

[0073] Step 2: Data processing and analysis:

[0074] 2.1 Data preprocessing: After receiving the data, the control unit performs data cleaning and normalization to ensure data accuracy;

[0075] 2.2 Real-time data analysis: Use big data analysis technology to conduct multi-dimensional analysis of data and identify key factors that affect the injection effect;

[0076] Specifically, the effect of mine water injection is affected by many factors, such as salinity, temperature, pressure, flow rate, etc. However, different factors have different degrees of influence on the injection effect, and some factors may have little effect or be redundant in some cases. To this end, LASSO regression is used for preliminary feature screening, principal component analysis (PCA) is used for dimensionality reduction, and then random forest regression is used to further analyze the nonlinear relationship between the influencing factors, and finally the contribution of each principal component to the injection effect is obtained; more specifically: LASSO regression screens out the most important variables for the injection effect, such as pressure P, salinity S and temperature T, and the optimization algorithm will adjust the flow rate Q and pressure P based on these variables; after PCA dimensionality reduction, some related variables (such as salinity and temperature) may be merged into principal components, and the optimization algorithm will be adjusted according to these principal components to simplify the calculation complexity of the optimization; then, random forest regression provides nonlinear relationships between variables, which will be used for precise adjustment of parameters during the optimization process.

[0077] Through multiple sensors installed in the mine water treatment system, multi-dimensional data are collected in real time, including: salinity (S) (ppm), temperature (T) (℃), pressure (P) (bar), flow rate (Q) (L / min), injection effect (Y) (score, 0-1, indicating the efficiency of injection); data is collected once every minute to generate a multi-dimensional time series data set, as shown in Table 1;

[0078] Table 1 Example data

[0079] time Salinity (ppm) Temperature(℃) Pressure(bar) Flow rate (L / min) Injection effect Y 12:00 10000 25 50 120 0.85 12:01 9800 26 48 118 0.82 12:02 9600 24 49 119 0.83 12:03 10200 27 51 125 0.87 … … … … … …

[0080] 2.2.1, feature selection using LASSO regression:

[0081] LASSO regression is a regularized regression method for processing high-dimensional data, and is particularly good at selecting the most important features from a large number of variables. The basic idea is to add an L1 regularization term to the regression coefficients, so that some regression coefficients become 0, thereby achieving feature selection. The goal is to find out which parameters of mine water (such as salinity, temperature, pressure, etc.) have the most significant impact on the injection effect (Y). Assume that the data is X1, X2, …X P Indicates that each variable represents salinity, temperature, etc., and the target variable is Y (injection effect).

[0082] Objective function of LASSO regression:

[0083]

[0084] in:

[0085] X ij represents the jth characteristic value of the i-th sample, including salinity, temperature, pressure, etc.;

[0086] y i is the injection effect (target variable);

[0087] β j is the regression coefficient of each parameter;

[0088] λ is a regularization parameter that controls the sparsity of the coefficients.

[0089] In the regression process, by adjusting λ, the parameters with little impact on the injection effect are automatically screened out and their coefficients are thinned to zero, so as to obtain the key parameters that have the most significant impact on the injection effect (Y) among the various parameters of mine water (such as salinity, temperature, pressure, flow rate, etc.). "Small impact" here means that the regression coefficient of a certain parameter is reduced to a small enough size during the calculation process of the LASSO model so that it can be ignored in prediction and optimization. When the regularization parameter λ increases, LASSO will impose a greater penalty on unimportant parameters and thin out their coefficients to 0. At this time, these parameters are deemed to have little impact on the injection effect. For example, the results show that the flow rate (Q) has little impact on the injection effect, and the model will reduce the coefficient β of the flow rate. Q , set to 0, and finally only the three key parameters of salinity, temperature and pressure are retained.

[0090] 2.2.2, Dimensionality reduction by principal component analysis (PCA): After using LASSO regression to screen out the most important variables, the factors that have a greater impact on the injection effect (such as salinity, temperature, and pressure) were retained. Next, principal component analysis (PCA) was used to process these variables.

[0091] Use principal component analysis (PCA) to identify the correlation between the key parameters obtained in step 2.2.1, and generate multiple new principal components to represent the main trend of these key parameter variables; although the number of variables has been reduced at this time, PCA generates several new "principal components" to represent the main trend of these variables by identifying the correlation between these variables. For example, salinity and temperature may have similar trends, and PCA will combine them into a new variable (principal component), so that fewer variables (principal components) can be used to represent most of the information of the original data, thereby simplifying the structure of the model and reducing the computational complexity.

[0092] For example: Assume that there are the following data in Table 2. Through LASSO, the three variables are retained. However, it is found that the trends of salinity and temperature are very similar: when salinity increases, temperature also has a similar trend. PCA will generate a principal component based on this similarity (which may be a weighted combination of "salinity + temperature"), and this principal component can explain the main changes in salinity and temperature. If PCA produces the following principal components: the first principal component: 80% is salinity, 20% is temperature, and the second principal component: mainly pressure; in this way, originally three variables need to be considered, now most of the changes in the data can be explained by two principal components. The first principal component explains the main changes in salinity and temperature, and the second principal component explains the changes in pressure.

[0093] Table 2 Example data

[0094] Salinity (ppm) Temperature(℃) Pressure(bar) 10000 25 50 9800 26 48 9600 24 49 10200 27 51

[0095] 2.2.3 Random forest regression analysis: An ensemble learning method that uses the results of multiple decision trees for integration. It can not only handle linear problems, but also complex nonlinear relationships, and can give the contribution of each feature. Random forest constructs multiple decision trees, each tree is trained on a different random subset, and finally the prediction results of each tree are averaged to obtain the final prediction value. This method is very suitable for processing nonlinear data, so it can further help analyze the nonlinear impact of each feature on the injection effect.

[0096] After LASSO regression and PCA, the key variables that have a greater impact on the "injection effect" have been screened out, and PCA has been used to further simplify the data structure and reduce redundant information. However, even if these variables or principal components are simplified, the relationship between them and the "injection effect" is often still not a simple linear relationship. Therefore, the following random forest regression is used to deal with the nonlinear relationship between these key variables or principal components and the "injection effect" and further improve the prediction accuracy.

[0097] The random forest model is used to perform regression analysis on the principal components after PCA dimensionality reduction to obtain the contribution of each principal component to the injection effect;

[0098] After completing LASSO and PCA, the simplified main variables (or principal components) have been obtained. Random forest regression will be further used to model the relationship between these variables and the injection effect. Its specific steps are as follows:

[0099] 1. Input variables: The principal components generated by PCA (e.g., principal component 1 representing salinity and temperature, and principal component 2 representing pressure) are used as input variables. The nonlinear effects of these principal components on the injection effect will be predicted through the random forest model.

[0100] 2. Model training: The principal components generated by PCA and the injection effects obtained at each moment in the past are trained, and random forest constructs multiple decision trees to capture the complex nonlinear relationship between the principal components and the injection effects.

[0101] 3. Prediction output: After training the model, new injection operations are predicted. By inputting new principal component values ​​in real time, the random forest regression model can accurately evaluate whether the current injection conditions are optimized and whether the injection parameters need to be adjusted.

[0102] 4. Variable importance analysis: Random forest can also output the importance score of each principal component to obtain the contribution of each variable to the injection effect. For example, the random forest model may find that principal component 1 (mainly salinity and temperature) has the greatest impact on the injection effect, accounting for 60%, while principal component 2 (pressure) accounts for 40%. This helps to further optimize the injection strategy, such as focusing on controlling the fluctuation range of salinity and temperature.

[0103] For example:

[0104] Suppose two main principal components are obtained through LASSO and PCA:

[0105] Principal component 1: mainly composed of salinity and temperature, explaining most of the variance in the data;

[0106] Principal component 2: mainly composed of pressure, explaining the remaining variance information.

[0107] These principal components are used as input to train the random forest model. The model builds 100 decision trees, each of which predicts the injection effect (Y) based on the different values ​​of principal component 1 and principal component 2.

[0108] In practical applications, when new real-time data is collected (including salinity, temperature, and pressure), new principal components 1 and 2 are obtained through PCA dimensionality reduction. Based on these inputs, the random forest model obtains the importance scores of principal components 1 and 2, for example:

[0109] The contribution of principal component 1 (salinity and temperature) to the injection effect is 60%;

[0110] The contribution of main component 2 (pressure) to the injection effect is 40%.

[0111] This information can help further optimize the injection process and determine which variables to exercise more control and adjustment on.

[0112] By performing multi-dimensional analysis on known real-time data, the system identifies the variables that most significantly affect the injection effect (step two), and then adjusts the injection parameters based on these analysis results to optimize future injection operations (step three).

[0113] Step 3: Optimization and calculation:

[0114] After completing the data processing and analysis in step 2, the system has identified the key variables (such as pressure P, salinity S, temperature T, etc.) that have the most significant impact on the current injection effect Y through LASSO regression, PCA dimensionality reduction and random forest regression, and through the analysis of these variables, clarified the relationship between them and the injection effect.

[0115] Based on the key variables with the most significant impact on the current injection effect obtained in step 2, the optimization phase is entered to dynamically adjust the injection parameters to ensure that the system can achieve a balance between injection efficiency, environmental protection and safety.

[0116] 3.1. Optimization objective function setting

[0117] The optimization goal is to optimize the injection efficiency, environmental impact and system safety by dynamically adjusting the pressure P and flow Q. Based on the contribution of each variable to the injection effect obtained in step 2, and ensuring a balance between injection efficiency, environmental protection and safety, the objective function is defined as follows:

[0118] f(P,Q)=α(t)·E(P,Q)-β(t)·G(S,T)-γ(t)·S(P,Q)

[0119] Among them, injection efficiency E(P,Q)=Q: It means that the injection efficiency is proportional to the flow rate Q;

[0120] Environmental impact G(S,T)=k1S+k2T: represents the impact of salinity S and temperature T on the environment. k1 and k2 are constant coefficients used to characterize the degree of impact of salinity and temperature on the environment.

[0121] System safety S(P,Q)=(P / Pmax) 2 +(Q / Qmax) 2 : Ensure that the injection pressure and flow rate are within the safe range. Pmax and Qmax are the safety thresholds of pressure and flow rate respectively.

[0122] α(t): The weight of the injected efficiency, which indicates the priority of the system for efficiency under the current conditions.

[0123] β(t): The weight of environmental impact, indicating the priority of the system in protecting the environment.

[0124] γ(t): The weight of system security, ensuring that the system can operate within a safe range at all times.

[0125] k1 represents the weight of the impact of salinity on the environment. Higher salinity may lead to groundwater contamination, reduced formation permeability or other negative environmental impacts. Therefore, the value of k1 is determined based on the sensitivity of salinity to the environment during mine water injection, through experimental data and industry standards.

[0126] k2 represents the weight of the impact of temperature on the environment. Changes in temperature may cause changes in the chemical reaction rate between mine water and groundwater, affecting the biological environment of the water body. The value of k2 is also determined by experimental data and industry standards.

[0127] 3.2. Dynamic adjustment and feedback mechanism

[0128] The Bayesian optimization algorithm is used, combined with real-time data feedback, to dynamically adjust the injection parameters to continuously optimize the injection process. The optimization process is achieved through the following steps:

[0129] ① Real-time feedback and iterative adjustment: The pressure, flow, salinity and temperature are monitored in real time through sensors, and the actual injection effect is compared with the objective function. Based on the feedback results, the system will dynamically adjust the flow Q and pressure P.

[0130] ② Dynamic adjustment of weight coefficients: Through real-time feedback, the optimization algorithm adjusts the weight coefficients α(t), β(t), and γ(t) in the objective function. For example, when the salinity is too high, the system automatically increases the weight of β(t) to give priority to reducing environmental impact; if the pressure is close to the safety threshold, the system will increase the weight of γ(t) to give priority to ensuring system safety.

[0131] Specific optimization examples:

[0132] Assume that through step 2: data processing and analysis, the system identifies that salinity S, pressure P and temperature T are the main factors affecting the injection effect. The current flow rate Q = 100L / min, pressure P = 45bar, salinity S = 10000ppm, temperature T = 25℃.

[0133] The initial objective function of the optimization algorithm is:

[0134]

[0135] The system initially sets α(t) = 1, β(t) = 0.5, and γ(t) = 0.8, and dynamically adjusts based on real-time feedback. If the salinity is found to be too high, the system may increase β(t) to reduce the flow and control the impact of salinity on the environment.

[0136] Through this feedback and adjustment mechanism, the system is able to achieve a balance and optimization between injection efficiency, environmental protection and system safety.

[0137] Step 4: Execution and Adjustment:

[0138] 4.1. Sending instructions: The control unit sends a control signal to the execution unit according to the optimization result in step 3, including adjustment instructions for injection pressure P and flow rate Q.

[0139] 4.2. Adjusting equipment: After receiving the instruction, the execution unit dynamically adjusts the injection pressure and flow rate by adjusting the variable frequency pump and regulating valve to make the system reach the optimal state after optimization.

[0140] Step 5: Environmental monitoring and protection:

[0141] 1. Monitoring groundwater quality: The environmental monitoring module monitors groundwater pH, heavy metal content and other indicators in real time. 2. Pollution warning: When abnormal groundwater quality is detected, the system automatically adjusts injection parameters to reduce pollution risks and notifies management personnel through the alarm system.

[0142] The objective function f(P,Q) = α(t)·E(P,Q)-β(t)·G(S,T)-γ(t)·S(P,Q) mainly considers injection efficiency, salinity, temperature and system safety, and does not directly consider environmental indicators such as pH value and heavy metals. However, when an over-standard alarm occurs in the environmental monitoring link, the system can still indirectly affect these environmental indicators by adjusting the injection pressure and flow rate.

[0143] Below is a detailed explanation and process on how to reduce negative environmental impacts by adjusting injection parameters when an over-limit alarm occurs during environmental monitoring.

[0144] Feedback mechanism and adjustment logic: When the environmental monitoring module detects that groundwater quality indicators (such as pH value, heavy metals) exceed the standard, the system will provide feedback and make adjustments through the following steps:

[0145] 5.1 Monitoring and alarm:

[0146] The environmental monitoring module monitors groundwater quality in real time; if an indicator (such as pH value or heavy metal content) is detected to exceed the set threshold, the system triggers an alarm.

[0147] 5.2 Data Feedback:

[0148] The environmental monitoring module transmits the exceeding data and alarm information to the control unit; after receiving the feedback, the control unit re-evaluates the current injection strategy.

[0149] 5.3 Adjust the objective function:

[0150] The control unit temporarily adjusts the weight coefficient in the objective function to give priority to solving the problem of exceeding environmental standards.

[0151] Specifically, a new environmental penalty term F is introduced env , and update the objective function:

[0152] f′(P,Q)=α(t)·E(P,Q)-β(t)·G(S,T)-γ(t)·S(P,Q)-δ(t)·F env

[0153] Among them, F env represents the environmental penalty, δ is the weight coefficient of environmental penalty; environmental penalty term, F env Definition: When the environmental monitoring module detects that the pH value exceeds the standard, F env Defined as:

[0154] F env =k pH ·(pH current -pH target ) 2

[0155] Among them, k pH It is the penalty coefficient for excessive pH value. When the pH value is too high or too low, it may cause changes in the acidity and alkalinity of groundwater, affecting the chemical balance and ecological environment of the water body. Therefore, the value of kpH is determined by experimental data and industry standards based on the degree of impact of excessive pH value on the environment during mine water injection. current is the currently monitored pH value, pH target is the target pH.

[0156] Similarly, when heavy metals exceed the standard, F env Defined as:

[0157] F env =k HM ·(HM current -HM target ) 2

[0158] Among them, k HM is the penalty coefficient for excessive heavy metals, which may lead to water pollution, soil quality degradation, and toxic effects on organisms. HM The value is determined based on the impact of excessive heavy metal content on the environment, through experimental data and industry standards. current is the currently monitored heavy metal content, HM target is the target heavy metal content.

[0159] 5.4 Re-optimize injection parameters:

[0160] The control unit uses the updated objective function f′(P, Q) to recalculate the optimal injection pressure P and injection flow rate Q. Through the optimization algorithm, new P and Q are found to maximize the objective function f′(P, Q) while reducing the impact of environmental over-limit.

[0161] 5.5 Implementation adjustments:

[0162] Apply new injection pressure and flow parameters to the injection system and adjust the operation of the injection equipment in real time; continue to monitor groundwater quality to ensure that the adjustment measures are effective until environmental indicators return to normal range.

[0163] Step 6: Remote monitoring and management:

[0164] 6.1. Real-time monitoring platform: provides a friendly user interface to display the system's operating status, historical data and prediction results;

[0165] 6.2. Remote operation: Users can remotely view the system status through the network and make necessary parameter adjustments and emergency processing.

[0166] Through the above technical solution, the present invention provides an efficient and reliable mine water deep well injection control system, which can significantly improve the treatment efficiency, reduce the impact on the environment, and has a wide range of application prospects. The adaptive control system of the present invention can not only be used for the treatment of high-salinity mine water, but also can be applied to other similar industrial wastewater treatment fields, providing strong support for environmental protection and resource utilization.

[0167] Specific embodiments of the present invention are given below. It should be noted that the present invention is not limited to the following specific embodiments, and all equivalent changes made on the basis of the technical solution of this application fall within the protection scope of the present invention.

[0168] Example:

[0169] A mining company produced a large amount of high-salinity mine water during the mining process, which posed a serious risk of pollution to the environment. To this end, the company decided to implement a deep well injection system to optimize the mine water treatment process and reduce the impact on the environment.

[0170] Step 1: Data collection:

[0171] 1.1. Installation of sensors: Salinity sensors, temperature sensors, pressure sensors and flow meters were installed at key locations of the mine water treatment system (such as the injection wellhead), a total of 10 sensors.

[0172] 1.2. Real-time monitoring: The sensor collects the salinity, temperature, pressure and flow data of the mine water in real time every minute. For example, the result of a data collection is as follows:

[0173] Time: 12:00, Salinity: 10000ppm, Temperature: 25℃, Pressure: 50bar, Flow rate: 120L / min.

[0174] 1.3. Data transmission: The collected data is transmitted to the control unit through the wireless module for subsequent data processing.

[0175] Step 2: Data processing and analysis:

[0176] 2.1. Data preprocessing: The control unit cleans and normalizes the received data, removes outliers, and ensures the accuracy of the data.

[0177] 2.2. Real-time data analysis: Using big data analysis technology, we can identify the key factors that affect the injection effect. Through LASSO regression, principal component analysis (PCA) and random forest regression, we can determine that salinity, temperature and pressure are the main factors affecting the injection efficiency.

[0178] Step 3: Optimization and calculation:

[0179] 3.1. Construction of adaptive optimization model: Based on real-time data, an adaptive optimization model was established to dynamically adjust the injection parameters (pressure and flow rate).

[0180] 3.2. Definition of optimization objective function: The system defines the objective function as:

[0181] f(P,Q)=α(t)·E(P,Q)-β(t)·G(S,T)-γ(t)S(P,Q)

[0182] 3.3. Dynamically adjust weight coefficients: According to real-time data and environmental monitoring feedback, dynamically adjust weight coefficients α(t), β(t), and γ(t) to ensure that injection efficiency is reduced first when environmental indicators exceed standards.

[0183] 3.4. Evaluation and adjustment of the optimization process: When the system detects that the salinity is high, the optimization algorithm is automatically triggered. The injection parameters after adjustment are:

[0184] Injection pressure: 45bar;

[0185] Injection flow rate: 110L / min.

[0186] Step 4: Execution and Adjustment:

[0187] 4.1. Execute injection operation: According to the optimized parameters, the control unit instructs the execution unit to adjust the variable frequency pump and the regulating valve to start injection.

[0188] 4.2. Dynamic adjustment process: During the injection process, the parameters are monitored in real time and feedback adjustments are made. When the pressure is detected to be close to the set upper limit, the flow rate is automatically reduced to ensure the safety of the injection process.

[0189] Step 5: Environmental monitoring and protection:

[0190] 5.1. Monitoring of groundwater quality: The environmental monitoring module monitors the key indicators of groundwater in real time, including pH value, heavy metal content and other indicators. The monitoring data is sent to the control unit through the data transmission device. pH value: 9.0 (exceeding the standard, the normal range is 6.5-8.5), heavy metal content: 0.02mg / L (normal range);

[0191] 5.2. Pollution warning: If the pH value or heavy metal content is detected to be excessive, the system will immediately trigger an alarm and adjust the injection parameters to reduce environmental impact. If the pH value exceeds the limit, the system will trigger an alarm mechanism;

[0192] 5.3. Introducing environmental penalty items: Since the pH value exceeds the standard, the system will introduce environmental penalty items F env ;

[0193] 5.4. Adjustment of objective function: f′(P,Q)=α(t)·E(P,Q)-β(t)·G(S,T)-γ(t)·S(P,Q)-δ(t)·F env

[0194] 5.5. Re-optimize injection parameters: The control unit uses the updated objective function f'(P,Q) to recalculate the optimal injection pressure P and flow rate Q. The new parameters obtained during the optimization process are: new injection pressure: 30 bar; new injection flow rate: 85 L / min.

[0195] 5.6. Implement adjustments: Apply the new injection pressure and flow parameters to the injection system and adjust the operation of the injection equipment in real time. The system will continue to monitor groundwater quality to ensure that the adjustment measures are effective until the pH value returns to the normal range (between 6.5 and 8.5).

Claims

1. A control system for optimizing injection parameters of deep well water in high-salinity mines, characterized in that: include: Data acquisition unit, including multiple sensors and detection equipment, used to monitor various parameters of mine water in real time; The control unit can receive real-time data from the data acquisition unit, analyze and process the data, and calculate the best injection parameters according to the preset optimization algorithm; The execution unit, including a variety of actuators and control devices, can automatically adjust the operating status of the injection equipment according to the instructions of the control unit to ensure that the mine water is injected into the ground in the optimal state; The environmental monitoring module, including groundwater quality monitoring instruments and data transmission equipment, can monitor the groundwater quality in the injection area and promptly detect and prevent possible pollution problems.

2. The control system for optimizing injection parameters of deep well water in high-salinity mines according to claim 1, characterized in that: The plurality of sensors and detection devices include a salinity sensor, a temperature sensor, a pressure sensor and a flow meter; The various actuators and control devices include variable frequency pumps, regulating valves and injection equipment.

3. A method for optimizing injection parameters of deep well water in high-salinity mines, characterized in that: The method is implemented based on the control system for optimizing the injection parameters of deep well water in high-salinity mines as described in claim 2; and comprises the following steps: Step 1: Data collection: The salinity, temperature, pressure and flow data of the mine water are collected in real time through the salinity sensor, temperature sensor, pressure sensor and flow meter, and the collected data are transmitted to the control unit; Step 2: Data processing and analysis: Data preprocessing: After receiving the data, the control unit performs data cleaning and normalization to ensure data accuracy; Real-time data analysis: Perform multi-dimensional analysis of the data to identify key factors that affect the injection effect; Step 3: Optimization and calculation: Based on the key variables affecting the current injection effect obtained in step 2, enter the optimization stage and dynamically adjust the injection parameters to ensure that the system can achieve a balance between injection efficiency, environmental protection and safety; Step 4: Execution and Adjustment: Sending instructions: The control unit sends a control signal to the execution unit according to the optimization result in step 3; Adjusting equipment: After receiving the instructions, the execution unit dynamically adjusts the injection pressure and flow rate by adjusting the variable frequency pump and the regulating valve to achieve the optimal state after optimization; Step 5: Environmental monitoring and protection: Monitoring of groundwater quality: The environmental monitoring module monitors the pH value and heavy metal content of groundwater in real time; Pollution warning: When abnormal groundwater quality is detected, the system automatically adjusts the injection parameters to reduce pollution risks and notifies management personnel through the alarm system; Step 6: Remote monitoring and management: Real-time monitoring platform: provides a user interface to display the system's operating status, historical data and forecast results; remote operation: users can remotely view the system status through the network and make necessary parameter adjustments and emergency processing.

4. The method for optimizing injection parameters of deep wells in high-salinity mines according to claim 3, characterized in that: In the step 2, data processing and analysis are to use LASSO regression for preliminary feature screening to screen out the key parameters that have the most significant impact on the injection effect; use principal component analysis to reduce the dimension and merge two or more related variables into principal components; then use random forest regression to analyze the nonlinear relationship between the principal components to obtain the contribution of each principal component to the injection effect.

5. The method for optimizing injection parameters of deep wells in high-salinity mines according to claim 4, characterized in that: The data processing and analysis include: Step 2.2.1, use LASSO regression for feature selection. The objective function of LASSO regression is: Among them, X ij represents the jth characteristic value of the i-th sample, including salinity, temperature, and pressure; y i is the injection effect; j is the regression coefficient of each parameter; λ is the regularization parameter; In the regression process, by adjusting λ, the parameters with little influence on the injection effect are automatically screened out and their coefficients are thinned to zero, so that the key parameters with the most significant influence on the injection effect are obtained among the various parameters of mine water; Step 2.2.2, principal component analysis for dimensionality reduction: Use principal component analysis to identify the correlation between the key parameters obtained in step 2.2.1, and generate multiple new principal components to represent the main change trends of these key parameter variables; Step 2.2.3, random forest regression analysis: Use the random forest model to perform regression analysis on the principal components after PCA dimensionality reduction to obtain the contribution of each principal component to the injection effect.

6. The method for optimizing injection parameters of deep wells in high-salinity mines according to claim 5, characterized in that: The step 2.2.3 includes: Input variables: The principal components generated by principal component analysis are used as input variables; Model training: The principal components generated by PCA obtained at each moment in the past are trained with the injection effect, and random forest constructs multiple decision trees to capture the complex nonlinear relationship between the principal components and the injection effect; Prediction output: After the model is trained, new injection operations are predicted. By inputting new principal component values ​​in real time, the random forest regression model can accurately evaluate whether the current injection conditions are optimized and whether the injection parameters need to be adjusted; Variable importance analysis: Random forest can also output the importance score of each principal component to obtain the contribution of each variable to the injection effect.

7. The method for optimizing injection parameters of deep wells in high-salinity mines according to claim 1, characterized in that: The step three comprises: Step 3.1, optimization objective function setting: Based on the contribution of each variable to the injection effect obtained in step 2, and to ensure a balance between injection efficiency, environmental protection and safety, the objective function is defined as follows: f(P,Q)=α(t)·E(P,Q)-β(t)·G(S,T)-γ(t)·S(P,Q) Among them, injection efficiency E(P,Q)=Q: means that the injection efficiency is proportional to the flow rate Q; environmental impact G(S,T)=k1S+k2T: means the impact of salinity S and temperature T on the environment, k1 means the weight of the impact of salinity on the environment, k2 means the weight of the impact of temperature on the environment; system safety S(P,Q)=(P / Pmax) 2 +(Q / Qmax) 2 : Ensure that the injection pressure P and flow rate Q are within the safe range, Pmax and Qmax are the safety thresholds of pressure and flow rate respectively; α(t) is the weight of injection efficiency, β(t) is the weight of environmental impact, and γ(t) is the weight of system safety; Step 3.2, dynamic adjustment and feedback mechanism: The Bayesian optimization algorithm is used in combination with real-time data feedback to dynamically adjust the injection parameters to continuously optimize the injection process; this is achieved through the following steps: Real-time feedback and iterative adjustment: real-time monitoring of pressure, flow, salinity and temperature through sensors, and comparison of the actual injection effect with the objective function; Dynamic adjustment of weight coefficients: through real-time feedback, the optimization algorithm adjusts the weight coefficients α(t), β(t), γ(t) in the objective function.

8. The method for optimizing injection parameters of deep wells in high-salinity mines according to claim 1, characterized in that: In step 5, when the environmental monitoring module detects that the groundwater quality index exceeds the standard, feedback and adjustment are performed through the following steps: Step 5.1, monitoring and alarm: The environmental monitoring module monitors the groundwater quality in real time; if a certain indicator is detected to exceed the set threshold, the system triggers an alarm; Step 5.2, data feedback: the environmental monitoring module transmits the exceeding data and alarm information to the control unit; after receiving the feedback, the control unit re-evaluates the current injection strategy; Step 5.3, adjust the objective function: adjust the weight coefficient in the objective function and introduce the environmental penalty term F env , and update the objective function as: f′(P,Q)=α(t)·E(P,Q)-β(t)·G(S,T)-γ(t)·S(P,Q)-δ(t)·F env Among them, F env represents the environmental penalty, δ is the weight coefficient of environmental excess; Step 5.4, re-optimize injection parameters: the control unit uses the updated objective function f′(P, Q) to recalculate the optimal injection pressure P and injection flow rate Q; through the optimization algorithm, find new P and Q to maximize the objective function f′(P, Q) while reducing the impact of environmental exceeding the standard; Step 5.5, implement adjustments: apply the new injection pressure and flow parameters to the injection system and adjust the operation of the injection equipment in real time; continue to monitor groundwater quality to ensure that the adjustment measures are effective until environmental indicators return to normal range.

9. The method for optimizing injection parameters of deep wells in high-salinity mines according to claim 8, characterized in that: When the environmental monitoring module detects that the pH value exceeds the standard, the environmental penalty item F env Defined as: F env =k pH ·(pH current -pH target ) 2 Among them, k pH is the penalty coefficient for exceeding the pH value limit, pH current is the currently monitored pH value, pH target is the target pH.

10. The method for optimizing injection parameters of deep wells in high-salinity mines according to claim 8, characterized in that: When the environmental monitoring module detects that heavy metals exceed the standard, the environmental penalty item F env Defined as: F env =k HM ·(HM current -HM target ) 2 Among them, k HM is the penalty coefficient for excessive heavy metal content, HM current is the currently monitored heavy metal content, HM target is the target heavy metal content.

Images