Converter station valve cooling system water abandoning strategy optimization model and method

Abstract

The invention provides a converter station valve cooling system water abandoning strategy optimization model and method. According to the method, the scaling index and the corrosion rate are calculated in real time by collecting the conductivity, the pH value, the temperature and the flow data of the circulating cooling water, so that the system is monitored in real time. And a multi-objective intelligent optimization control method is designed by defining optimization objectives and constraints and establishing a dynamic performance index model. According to the method, a plurality of conflict targets in the wastewater treatment process can be balanced, for example, the water abandoning amount is reduced, the energy consumption is reduced to improve the cooling effect, so that intelligent optimization of the water abandoning strategy of the converter station valve cooling system is realized, the environmental protection performance is improved, the operation state of the converter station is optimized, and fine management and optimization of the water abandoning strategy are realized. The method is simple and easy to operate and has high practical value.

Description

Technical Field

[0001] The present invention belongs to the field of converter station wastewater treatment, and more specifically relates to a converter station valve cooling system water abandonment strategy optimization model and method. Background Art

[0002] Converter stations are crucial components of high-voltage direct current (HVDC) transmission systems. They convert energy between DC and AC. This process generates significant amounts of heat. This raises a key challenge: how to effectively cool the equipment to ensure proper operation while also recovering and utilizing this heat as much as possible.

[0003] To achieve this goal, converter stations are typically equipped with cooling systems. A key component of these systems is the cooling water system. The valve cooling system is one such system, which uses water cooling to lower valve temperatures and ensure the normal and stable operation of the converter station.

[0004] However, converter station valve cooling systems generate a certain amount of waste water during operation, potentially impacting the environment. Furthermore, water resources are inherently valuable, and generating large amounts of waste water during the cooling process can lead to significant waste. Therefore, effective management and optimization of waste water from converter station valve cooling systems is crucial.

[0005] In addition, existing cooling water treatment methods have many shortcomings. For example, traditional water abandonment strategies often find it difficult to comprehensively consider and optimize multiple factors, such as the amount of abandoned water, system energy consumption, and cooling effect. Furthermore, existing optimization methods are often complex and difficult to implement. There is a need for a model and method that can simply and effectively optimize the water abandonment strategy of the converter station valve cooling system. Therefore, the present invention relates to a model and method for optimizing the water abandonment strategy of the converter station valve cooling system to address the shortcomings of the existing technology. Summary of the Invention

[0006] This invention aims to address the inability of existing HVDC converter station valve cooling systems to dynamically optimize multiple objectives, including wastewater volume, energy consumption, and cooling effectiveness, when managing water abandonment strategies. It also addresses the difficulty of implementing traditional optimization methods and their inability to balance wastewater treatment and environmental protection requirements. To this end, this invention proposes an optimization model and method that effectively reduces energy consumption and improves cooling effectiveness while meeting environmental requirements, enabling intelligent optimization control of the valve cooling system's water abandonment strategy.

[0007] In order to achieve the above object, the present invention is implemented by adopting the following technical solutions: the method comprises:

[0008] Collect the conductivity, pH value, temperature and flow rate data of circulating cooling water, and calculate the scaling index and corrosion rate in real time;

[0009] Define optimization objectives and constraints;

[0010] Establish dynamic performance indicator models;

[0011] Design multi-objective intelligent optimization control methods and balance the relationship between multiple conflicting objectives in the dynamic environment of wastewater treatment process.

[0012] In one solution, the conductivity, pH, temperature, and flow rate data of the circulating cooling water in the converter station's valve cooling system are collected to calculate the scaling index and corrosion rate in real time, and to perform outlier detection and threshold warning. This solution includes a high-precision sensor network and real-time computing architecture, including conductivity sensors, pH sensors, turbine flow meters, and platinum resistance temperature sensors.

[0013] Calcium ion concentration is estimated using an online water quality analyzer or soft sensor model, total alkalinity is solved using the pH value and carbonate equilibrium equation, dissolved solid concentration is converted using the empirical formula of conductivity and temperature, corrosion rate is evaluated by combining electrochemical monitoring with the physicochemical parameter correlation model, and a sliding window filtering algorithm is used to eliminate sensor noise.

[0014] In one solution, the optimization objectives and constraints are defined as follows: the objectives are: minimizing the amount of discarded water, pumping energy consumption, and water quality deviation; the constraints are: conductivity ≤ 1500 μS / cm, valve opening range, and system water balance equation.

[0015] In one solution, the establishment of a dynamic performance index model includes: an energy consumption model, a water quality deviation model, and a comprehensive performance index model.

[0016] In one scheme, the described design multi-objective intelligent optimization control method proposes a multi-objective intelligent optimization control method - DMIOC; in DMIOC, in order to accurately obtain the dynamic characteristics of the wastewater treatment process, an energy consumption and effluent water quality model based on an adaptive kernel function is established; at the same time, in order to balance the relationship between energy consumption and effluent water quality, a DMOPSO algorithm based on an adaptive flight parameter adjustment mechanism is designed to obtain suitable optimized set values of the controlled variables.

[0017] In one embodiment, the multi-objective intelligent optimization control method includes:

[0018] (1) Establish a dynamic performance indicator model;

[0019] (2) Dynamic optimization of control variable setpoints;

[0020] (3) Tracking of control variable set values.

[0021] In one solution, the circulating cooling water parameters and system operation data collected by the energy consumption model are used to establish a dynamic model of pumping energy consumption by means of an adaptive kernel function method.

[0022] In one solution, the water quality deviation model uses an adaptive kernel function method to establish a dynamic model of water quality deviation, including conductivity and pH value deviation.

[0023] In one solution, the comprehensive performance index model combines the energy consumption model and the water quality deviation model to construct a comprehensive performance index function as the basis for multi-objective optimization; the integration of the comprehensive performance index model performs dimensionless and dynamic weighting on energy consumption and water quality deviation. Beneficial effects of the present invention:

[0024] The waste water strategy optimization model and method for the valve cooling system of the converter station of the present invention can achieve multi-objective dynamic optimization, including reducing the amount of waste water, reducing energy consumption, improving the cooling effect, etc., thereby helping to improve the operation efficiency and reliability of the converter station. Its intelligent optimization control method can achieve fine management of the waste water strategy of the valve cooling system of the converter station, which not only meets the environmental protection requirements but also optimizes the operation status of the converter station, significantly improves the reasonable utilization rate of cooling water and effectively reduces environmental pollution. At the same time, compared with traditional optimization methods, it is simple to operate and has low implementation difficulty, and is widely used in the field of high-voltage direct current transmission, with high practical value. Description of the Drawings

[0025] Figure 1 It is a flow chart of the method of the present invention;

[0026] Figure 2 It is a flow chart of establishing a dynamic performance index model of the present invention;

[0027] Figure 3 It is a flow chart of the multi-objective intelligent optimization control method. Detailed Embodiments

[0028] To facilitate the understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Typical embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the disclosure of the present invention more thorough and comprehensive.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as those understood by those skilled in the art to which the present invention pertains. The terms used in the present specification are for the purpose of describing specific embodiments only and are not intended to limit the present invention. To facilitate understanding of the present invention, a more comprehensive description of the present invention will be provided below with reference to the accompanying drawings. Typical embodiments of the present invention are shown in the drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present invention.

[0030] like Figure 1 As shown, the present invention provides a water abandonment strategy optimization model and method for a converter station valve cooling system, including the following steps:

[0031] Step 1: Collect the conductivity, pH value, temperature, and flow rate data of the circulating cooling water, and calculate the scaling index (LSI) and corrosion rate (CR) in real time.

[0032] In the actual operation of the converter station valve cooling system, step 1 must be implemented through a high-precision sensor network and real-time computing architecture. Key parameters of the circulating cooling water (conductivity, pH, temperature, and flow) are collected in real time using industrial-grade sensors. The conductivity sensor uses a four-electrode design to eliminate polarization effects and is installed in the cooling tower outlet main pipe. The pH sensor uses a glass electrode with an automatic temperature compensation module and is deployed at the circulating water pump inlet. Turbine flowmeters and platinum resistance temperature sensors are distributed at key system nodes. All sensor signals are transmitted to the central controller via 4-20mA or Modbus protocol, and timestamp synchronization technology is used to ensure data consistency. The scaling index (LSI) is calculated based on the Langelier saturation index model. It requires dynamic acquisition of water parameters such as calcium hardness, total alkalinity, and ionic strength. Calcium ion concentration is indirectly estimated using an online water quality analyzer or soft sensor model. Total alkalinity is calculated by solving the pH and carbonate equilibrium equations simultaneously. Dissolved solids concentration is converted using an empirical formula based on conductivity and temperature. The evaluation of corrosion rate (CR) combines electrochemical monitoring with the correlation model of physicochemical parameters. For example, the linear polarization resistance method is used to measure the instantaneous corrosion current density in real time. At the same time, the empirical corrosion rate formula of carbon steel materials is constructed based on the chloride ion concentration, dissolved oxygen content and flow rate of the water body. The data processing link uses a sliding window filtering algorithm to eliminate sensor noise, and the outlier detection module (such as outlier analysis based on Mahalanobis distance) is used to remove interference data. Finally, the calculation results of LSI and CR are output at a refresh frequency of 1 to 5 seconds and written to the real-time database for subsequent optimization module calls. At the same time, the threshold warning mechanism is triggered (such as starting the anti-scaling dosing program when LSI>0.5, and issuing a material degradation alarm when CR>0.1mm / a), forming a closed-loop link from data perception to decision support.

[0033] Step 2: Define the optimization objectives and constraints. Objectives: Minimize the amount of wasted water, pumping energy consumption, and water quality deviation (conductivity / pH deviation). Constraints: Conductivity ≤ 1500 μS / cm, valve opening range, and system water balance equation.

[0034] After completing the real-time data collection and dynamic assessment of scaling and corrosion in Step 1, Step 2 needs to transform the multi-objective optimization problem into a mathematical formal expression. The core lies in constructing an objective function that takes into account economy, energy consumption, and water quality stability, and ensuring the feasibility of system operation through physical constraints. The mathematical description of the optimization objectives needs to fuse the amount of wasted water Q 弃 , pumping energy consumption E 泵 and water quality deviation (Δσ, ΔpH) through multi-objective fusion, and define the comprehensive objective function using the dynamic weighted summation method:

[0035]

[0036] where α(t), β(t), and γ(t) are time-varying weight coefficients, and their priorities are dynamically adjusted according to the real-time states of the scaling index LSI and corrosion rate CR through an online learning algorithm (such as fuzzy logic or reinforcement learning). For example, when LSI > 0, increase γ(t) to strengthen water quality control, and when CR exceeds the standard, increase α(t) to reduce the enrichment of corrosive ions. The pumping energy consumption model needs to construct a non-linear relationship based on the flow rate Q, pump efficiency η, and head H in Step 1:

[0037]

[0038] where ρ is the density of cooling water, g is the acceleration due to gravity, n is the pump speed, Δt is the control period, and the efficiency η is corrected in real time through the adaptive kernel function model in Step 3.

[0039] The mathematical expression of the constraint conditions needs to strictly follow the physical limitations and chemical stability requirements of the system. The hard constraint of conductivity is expressed in the form of an inequality σ ≤ 1500 μS / cm, which is directly related to the cumulative effect of ion concentration. When the online prediction model (such as ARIMA or LSTM) predicts that it may exceed the limit in the future period, an early water waste action is triggered. The valve opening constraint reflects the physical limit of the mechanical actuator and is defined as k min ≤ k ≤ k max , where k is the valve opening percentage, and its boundary values are calibrated according to the flow characteristic curves of valve types (such as butterfly valves and ball valves). The system water balance equation needs to satisfy the law of conservation of mass:

[0040]

[0041] where Q 补 is the make-up water volume, Q 循环 is the circulating flow rate, Q蒸发 Calculated from the cooling tower heat load and ambient temperature and humidity, Q 泄漏 Estimated by monitoring the pipeline pressure drop through a pressure sensor is the change rate of the system water volume (usually approaching zero during steady-state operation). This equation introduces an optimization model through the Lagrange multiplier method to ensure that the solution space satisfies the mass balance. In addition, the dynamic coupling effect of scaling and corrosion needs to be implicitly constrained. For example, when LSI > 2.0, the weight of the discarded water volume is forced to increase to prevent pipeline blockage. Such logic is embedded in the optimization framework through mixed integer programming or a fuzzy rule base to form a complete constraint system that takes into account both explicit constraints and implicit experience.

[0042] Step 3: Establish a dynamic performance index model;

[0043] As Figure 2 shown, S301: Establish an energy consumption model. According to the circulating cooling water parameters and system operation data collected in Step 1, use the adaptive kernel function method to establish a dynamic model of pumping energy (PE). This model can accurately describe the energy consumption changes of the system under different operating conditions.

[0044] In the dynamic performance index modeling of Step 3, first, a non-linear regression model based on the adaptive kernel function needs to be constructed for the pumping energy (PE). This model takes the flow rate, head, pump speed, cooling water density ρ, and valve opening k collected in real-time in Step 1 as input variables, and constructs a feature space mapping through the radial basis kernel function κ(x i ,x j )=exp(-γ||x i -x j || 2 ), where γ is the dynamically adjusted kernel bandwidth parameter, updated by the backpropagation of the mean square error of historical data within a sliding time window. The mathematical expression of the energy consumption model is:

[0045]

[0046] where α i (t) is the time-varying weight coefficient, b(t) is the bias term, and the parameters are updated online using the recursive least squares method (RLS). The update rule is:

[0047]

[0048] where φ(t) is the mapping of the current input vector in the kernel space, λ is the forgetting factor (usually taken as 0.95 - 0.99), and the covariance matrix P(t) is calculated recursively through the Sherman-Morrison formula to ensure that the model can capture dynamic characteristics such as cooling tower load fluctuations and pipeline resistance changes.

[0049] S302. Similarly using the adaptive kernel function method, a water quality deviation model is established, including the conductivity and pH deviation models. This model can reflect the changes of the system water quality parameters with the operating conditions and time.

[0050] The establishment of the water quality deviation model needs to synchronously handle the multivariable coupling effect of conductivity σ and pH value. Define the conductivity deviation Δσ = |σ - σ set | / σ set and the pH deviation ΔpH = |pH - pH set |, and adopt a mixed kernel function (the convex combination of linear kernel and Gaussian kernel) to construct a multi-output regression model:

[0051]

[0052] The input vector u(t) includes the makeup water volume Q 补 、the circulation flow rate Q 循环 、the chemical dosage C chem and the environmental temperature T amb , and the weight coefficient β j (t) is jointly optimized through the multi-task learning framework, and the loss function is defined as:

[0053]

[0054] where μ is the sparsity penalty factor, and the coordinate descent method is used to update the model parameters online to cope with the slow time-varying disturbances such as the lag of chemical reaction kinetics and the growth of microorganisms.

[0055] S303. The comprehensive performance index model combines the energy consumption model and the water quality deviation model to construct a comprehensive performance index function as the basis for multi-objective optimization.

[0056] The integration of the comprehensive performance index model needs to dimensionless and dynamically weight the energy consumption and the water quality deviation. Define the normalized energy consumption index and the comprehensive water quality deviation Fuse the two through the time-varying coupling coefficient ξ(t):

[0057]

[0058] where ξ(t) is dynamically derived from the optimization objective weights α(t), β(t), γ(t) in step 2, satisfying ξ(t) = α(t) / (α(t) + γ(t)), and the sigmoid function is introduced for smooth transition to prevent the oscillation caused by the objective conflict. This comprehensive index is used as the rolling optimization benchmark of the model predictive control (MPC), embedded in the multi-objective optimizer in step 2, and finally realizes the dynamic trade-off between the minimum energy consumption and the stable water quality.

[0059] Step 4. Design a multi-objective intelligent optimization control method. In order to effectively balance the relationship between multiple conflicting objectives in the dynamic environment of the wastewater treatment process, a multi-objective intelligent optimization control method, DMIOC, is proposed. In DMIOC, in order to accurately obtain the dynamic characteristics of the wastewater treatment process, an energy consumption and effluent water quality model based on an adaptive kernel function is established; at the same time, in order to balance the relationship between energy consumption and effluent water quality, a DMOPSO algorithm based on an adaptive flight parameter adjustment mechanism is designed to obtain suitable optimized set values of the controlled variables.

[0060] like Figure 3 As shown, S401, establishment of dynamic performance index model

[0061] The main goal of the wastewater treatment operation process is to reduce the operating energy consumption while ensuring that the effluent water quality meets the standards. Accurately describing the dynamic performance indicators PE, AE and EQ of the wastewater treatment process is the key to improving the operating performance. The dynamic adjustment cycle of the wastewater treatment process performance indicators PE, AE and EQ is 2 hours. Therefore, the dynamic characteristics of the performance indicators should be obtained in real time during each optimization cycle. This paper obtains the process variables related to PE, AE and EQ by analyzing the dynamic characteristics and operation data of the wastewater treatment process, which are the inlet flow rate Q in ,S O ,S NO , ammonia nitrogen (Ammonian itrogen,) and suspended solids concentration (Suspended solids, SS). According to the analyzed relevant process variables, the adaptive kernel function method is used to establish the dynamic relationship between PE and relevant process variables f1(x(t)), the dynamic relationship between AE and relevant process variables f2(x(t)), and the dynamic relationship between EQ and relevant process variables f3(x(t)). The expressions are:

[0062]

[0063] Among them, f1(x(t)) is the PE model at time t, f2(x(t)) is the AE model at time t, and f3(x(t)) is the EQ model at time t. x(t) is the input variable at time t, x(t) = [Q in (t),S O (t),S NO (t),S NH (t),SS(t)].c 1r (t),c 2r (t), c 3r (t) are the centers of the rth kernel function in the PE, AE and EQ models at time t, b 1r (t), b 2r (t), b 3r(t) are the widths of the r-th kernel function in the PE, AE, and EQ models at time t, respectively, W 1r (t), W 2r (t), W 3r (t) are the connection weights of the r-th kernel function in the PE, AE, and EQ models at time t, respectively, W 10 (t), W 20 (t), W 30 (t) are the biases of the PE, AE, and EQ models at time t, respectively, r = 1, 2, …, R, and R is the number of kernel functions.

[0064] The PE, AE, and EQ models of the wastewater treatment process based on adaptive kernel functions can not only adaptively adjust the performance index model parameters according to the error between the actual output and the desired output to ensure the accuracy of the performance index model; at the same time, the proposed performance index model establishes the relationship between PE, AE, and EQ and the key controlled variables, which is conducive to realizing the dynamic multi-objective intelligent optimization control of the wastewater treatment process.

[0065] S402. Dynamic optimization of the control variable setpoint

[0066] In order to optimize the dynamic performance indices PE, AE, and EQ simultaneously and obtain the optimized setpoints of the controlled variables, the established PE model f1(x(t)), AE model f2(x(t)), and EQ model f3(x(t)) at time t are used as the optimization objective functions, and the DMO PSO algorithm based on the adaptive flight parameter adjustment mechanism is used to optimize them. The optimization objective function F(t) is

[0067] F(t) = min{f1(x(t)), f2(x(t)), f3(x(t))}

[0068] During the optimization iteration process, the particle velocity and position update formulas are

[0069]

[0070] where k is the iteration number in the evolution process, d is the dimension in the search space, d = 1, 2, …, D; ω is the inertia weight, which is used to control the influence of the previous velocity on the current velocity; c1 and c2 are the acceleration constants, r1 and r2 are random values uniformly distributed in [0, 1], p i,(k) ) is the historical best position of the i-th particle at the k-th iteration, expressed as Meanwhile, g (k) is the global best position obtained through population optimization, expressed as The flight parameters ω, c1, and c2 in the DMOPSO algorithm are the main factors affecting the global exploration ability and local development ability. During the optimization process, the larger the inertia weight ω, the larger the learning parameter c1, and the smaller the learning parameter c2, the stronger the global search ability of the particle swarm. When the inertia weight ω is smaller, the learning parameter c1 is smaller, and the learning parameter c2 is larger, the local development ability of the entire particle swarm is stronger. To effectively balance the diversity and convergence of particles, the present invention proposes a parameter adjustment method based on population spacing information. The expression of the population spacing information is

[0071]

[0072] where l t,i (k + 1) is the minimum Manhattan distance between the i-th particle and other particles at the (k + 1)-th iteration at time t, is the average Manhattan distance of all particles at the (k + 1)-th iteration at time t, and S is the number of particles.

[0073] Since the flight process of particles has complex non-linear characteristics, this paper proposes a non-linear function to express the flight process of particles

[0074]

[0075] where θ t (k + 1) is the non-linear function of the particle flight process at time t. The iterative update formula for the flight parameter ω t =[ω t1 (k), ω t2 (k), …, ω t5 (k)] is

[0076]

[0077] c t1 =[c t1,1 (k), c t1,2 (k), …, c t1,5 (k)], c t2 =[c t2,1 (k), c t2,2 (k), …, c t2,5 (k)], and the iterative update formula for

[0078] is

[0079]

[0080] where ω ti (k + 1) is the inertia weight of the i-th dimension at the (k + 1)-th iteration at time t, c t1i (k + 1) and c t2,i(k + 1) are the local and global learning factors of the i-th dimension at the (k + 1)-th iteration at time t, respectively.

[0081] The change in the population spacing information can reflect the distribution of the particle population. When the PS value is large, it indicates that the distribution of the particle swarm is uneven, and the particle swarm at this time needs to improve its global search ability. Therefore, in order to increase the diversity of the particle swarm, the parameters ω and c1 should be increased, and the parameter c2 should be decreased. On the other hand, a small PS value indicates good diversity of the particle swarm, and the particle swarm should improve its local development ability. The parameters ω and c1 should be decreased, and the parameter c2 should be increased to improve the convergence performance of the particle swarm. After optimizing the objective function in Equation (5) of the DMOPSO based on the designed parameter adjustment mechanism based on population spacing information, non-dominated solutions can be obtained and saved in the archive. In the process of selecting the global optimal solution, according to the criterion of minimum energy consumption when the organic matter S NH and SS in the effluent meet the standards (S NH < 4 mg / l, SS < 18 mg / l), a set of optimal solutions g is selected. It can be used to balance the relationship between PE, AE, and EQ, and save the optimized setpoint values of the controlled variables required. and S403, tracking of the setpoint values of the control variables

[0082] Use the multi-loop PID control strategy to complete the control of the optimized setpoint values of the controlled variables and . The multi-loop PID control strategy adjusts the operating variable error in the proportional, integral, and derivative links to achieve the adjustment of the controlled variable. The adjustment time of the controlled variable is 15 min. The output expression of the multi-loop PID control strategy is

[0083]

[0084] where, Δu(t) is the operating variable, Δu(t) = [ΔK La5 (t), ΔQ a (t)] T , ΔK La5 (t) is the change in the oxygen transfer coefficient in the fifth zone, ΔQ a (t) is the change in the internal reflux flow rate, K p is the proportional coefficient matrix, H i is the integral coefficient matrix, H d is the derivative coefficient matrix, e(t) is the error between the actual output and the optimized setpoint value of S O and S NO , expressed as

[0085] e(t) = z(t) - y(t) (16)

[0086] where \(e(t)=[e_1(t),e_2(t)]\) T , \(y(t)=[y_1(t),y_2(t)]\) T =[S O (t),S NO (t)] T , \(y_1(t)\) is the actual S O concentration at time \(t\), and \(y_2(t)\) is the actual S NO concentration at time \(t\).

[0087] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The described program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above various methods. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0088] It should be understood that the detailed description of the technical solutions of the present invention with the aid of the preferred embodiments is illustrative rather than restrictive. Those of ordinary skill in the art can modify the technical solutions recorded in each embodiment on the basis of reading the specification of the present invention, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of each embodiment of the present invention.

Claims

1. An optimization model and method for the waste water strategy of the valve cooling system in a converter station, characterized in that: The method includes: Collect the conductivity, pH value, temperature and flow rate data of circulating cooling water, and calculate the scaling index and corrosion rate in real time; Define optimization objectives and constraints; Establish dynamic performance indicator models; Design multi-objective intelligent optimization control methods and balance the relationship between multiple conflicting objectives in the dynamic environment of wastewater treatment process.

2. The optimization model and method for the waste water strategy of the valve cooling system of a converter station according to claim 1, characterized in that: By collecting the conductivity, pH value, temperature and flow data of the circulating cooling water in the converter station valve cooling system, the scaling index and corrosion rate are calculated in real time, and abnormal value detection and threshold warning are carried out. The installation method includes: a high-precision sensor network and real-time computing architecture, covering conductivity sensors, pH sensors, turbine flow meters and platinum resistance temperature sensors; Calcium ion concentration is estimated using an online water quality analyzer or soft sensor model, total alkalinity is solved using the pH value and carbonate equilibrium equation, dissolved solid concentration is converted using the empirical formula of conductivity and temperature, corrosion rate is evaluated by combining electrochemical monitoring with the physicochemical parameter correlation model, and a sliding window filtering algorithm is used to eliminate sensor noise.

3. An optimization model and method for the waste water strategy of a converter station valve cooling system according to claim 1, characterized in that: The defined optimization objectives and constraints include: objectives: minimizing the amount of discarded water, pumping energy consumption, and water quality deviation; constraints: conductivity ≤ 1500 μS / cm, valve opening range, and system water balance equation.

4. The optimized model and method for the waste water strategy of the valve cooling system in a converter station according to claim 1, characterized in that: The dynamic performance index model includes: energy consumption model, water quality deviation model, and comprehensive performance index model.

5. The optimized model and method for the waste water strategy of the valve cooling system of a converter station according to claim 1, characterized in that: The described design multi-objective intelligent optimization control method proposes a multi-objective intelligent optimization control method - DMIOC; in DMIOC, in order to accurately obtain the dynamic characteristics of the wastewater treatment process, an energy consumption and effluent water quality model based on an adaptive kernel function is established; at the same time, in order to balance the relationship between energy consumption and effluent water quality, a DMOPSO algorithm based on an adaptive flight parameter adjustment mechanism is designed to obtain suitable optimized set values of the controlled variables.

6. The optimized model and method for the waste water strategy of the valve cooling system of a converter station according to claim 5, characterized in that: The multi-objective intelligent optimization control method includes: (1) Establish a dynamic performance indicator model; (2) Dynamic optimization of control variable setpoints; (3) Tracking of control variable set values.

7. An optimization model and method for the waste water strategy of the valve cooling system in a converter station according to claim 4, characterized in that: The energy consumption model collects circulating cooling water parameters and system operation data, and uses an adaptive kernel function method to establish a dynamic model of pumping energy consumption.

8. The optimization model and method for the waste water strategy of the valve cooling system of a converter station according to claim 4, characterized in that: The water quality deviation model uses an adaptive kernel function method to establish a dynamic model of water quality deviation, including conductivity and pH value deviation.

9. An optimization model and method for the waste water strategy of the valve cooling system of a converter station according to claim 4, characterized in that: The comprehensive performance index model is combined with the energy consumption model and the water quality deviation model to construct a comprehensive performance index function as the basis for multi-objective optimization; The integration of comprehensive performance indicator model makes energy consumption and water quality deviation dimensionless and dynamically weighted.

Images