Multi-level collaborative optimization control method for ultrapure water production

Abstract

The present invention relates to the technical field of water treatment and intelligent control, and in particular to a multi-stage collaborative optimization control method for ultrapure water production. The present invention combines genetic algorithm and fuzzy logic control technology to dynamically optimize and collaboratively control the operating parameters of multi-stage processes (nanofiltration, reverse osmosis, and electrodeionization), thereby solving the problem of lack of intelligence, real-time performance, and multi-process collaborative optimization in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of water treatment and intelligent control technology, and in particular to a multi-stage collaborative optimization control method for ultrapure water production. The method is applicable to ultrapure water preparation processes in the photovoltaic and semiconductor industries, and meets their strict requirements for high purity, high stability and low energy consumption. Background Art

[0002] Ultrapure water is an indispensable and important resource in the photovoltaic and semiconductor production processes, and is widely used in silicon wafer cleaning, chemical etching, coating and other high-precision processing links. The photovoltaic industry's silicon wafer cleaning and etching processes require extremely high water purity, while the semiconductor industry has stricter control over the stability and particulate matter of ultrapure water. However, in the existing technology, the problems of water quality fluctuations and excessive energy consumption have not been effectively solved for a long time, and it is difficult to simultaneously meet the efficient collaboration needs of photovoltaic and semiconductor production lines. For example, the total organic carbon (TOC) content must be less than 50 ppb, the electrical conductivity (EC) must be less than 0.055 µS / cm, and the particle content (PC) must be less than 10 / mL. These indicators directly affect the performance, yield and service life of photovoltaic modules and semiconductor products.

[0003] The existing ultrapure water preparation process usually adopts a multi-stage treatment process, including raw water pretreatment, nanofiltration (NF), reverse osmosis (RO) and electrodeionization (EDI) and other core process links. However, the operating parameters in traditional processes are mostly fixed values, which makes it difficult to dynamically respond to fluctuations in raw water quality or changes in equipment status, resulting in reduced system operating efficiency and difficulty in ensuring water quality stability. In addition, concentrated water discharge is not fully utilized, and resource waste and energy consumption are particularly prominent.

[0004] In a few technical solutions, some processes use programmable logic controllers (PLCs) to adjust the operating status of a single device, but lack multi-level collaborative optimization and intelligent control of the overall process, and cannot meet the dynamic adaptability and high efficiency requirements of photovoltaic and semiconductor production lines for ultrapure water preparation processes. Especially when water quality fluctuates greatly or load conditions change, traditional technologies are difficult to take into account water quality stability, system energy consumption optimization, and efficient resource utilization at the same time.

[0005] Therefore, how to coordinately optimize the multi-stage process of ultrapure water preparation through intelligent control means, especially to realize intelligent linkage control of multi-stage processes through real-time monitoring and dynamic feedback, has become a technical problem that needs to be urgently solved in the photovoltaic and semiconductor industries. Summary of the invention

[0006] The present invention provides a multi-stage collaborative optimization control method for ultrapure water production, which is applicable to the ultrapure water preparation process in the photovoltaic and semiconductor industries. By combining genetic algorithms and fuzzy logic control technology, the operating parameters of multi-stage processes (nanofiltration, reverse osmosis, and electrodeionization) are dynamically optimized and collaboratively controlled, solving the problem of lack of intelligence, real-time performance, and multi-process collaborative optimization in the prior art.

[0007] The technical solution of the present invention is as follows: a multi-stage collaborative optimization control method for ultrapure water production, comprising the following steps:

[0008] (1) Raw water pretreatment: The raw water is preliminarily treated through sand filtration, activated carbon filtration and softening devices to remove suspended particles, organic matter and residual chlorine. The total organic carbon (TOC), electrical conductivity (EC) and particle content (PC) in the raw water are monitored in real time through a water quality sensor array to provide water quality data for subsequent processes.

[0009] (2) Nanofiltration treatment: Use nanofiltration equipment to perform preliminary desalination on the pretreated water. Combined with water quality data, genetic algorithms are used to optimize the operating pressure and flow rate. Fuzzy logic control is used to fine-tune the operating status and dynamically adjust the concentrate reuse ratio.

[0010] (3) Reverse osmosis deep desalination: The nanofiltration effluent is deeply desalinated by reverse osmosis. The genetic algorithm optimizes the operating pressure and the ratio of concentrated water discharge. The fuzzy logic control adjusts the desalination rate in real time.

[0011] (4) Electrodeionization and fine desalination: Electrodeionization (EDI) devices are used to perform final fine desalination on the reverse osmosis effluent. Combined with real-time monitoring data, the operating current and water flow rate are dynamically adjusted to ensure that the ultrapure water quality meets the standards.

[0012] (5) Real-time monitoring and control: Water quality data is obtained through a water quality sensor array, and the operating parameters of the nanofiltration, reverse osmosis and EDI devices are optimized using genetic algorithms. Fuzzy logic control is combined to fine-tune the operating status to form an intelligent dynamic optimization closed loop.

[0013] Preferably, the real-time monitoring includes using a TOC sensor, a conductivity sensor and a particle sensor, and the monitoring data is transmitted to the control unit through a PLC system for use by the optimization algorithm and the control module.

[0014] Preferably, the genetic algorithm optimizes the process parameter x={ , ,..., }, the fitness function is defined as:

[0015] ,

[0016] in: : Water quality target achievement rate, defined as:

[0017] ,

[0018] E(x): Energy consumption per unit of water production, defined as:

[0019] ,

[0020] in, is the device power, is the running time, is the water production;

[0021] S(x): The fluctuation range of water quality parameters, defined as:

[0022] ,

[0023] , , is the weight coefficient, satisfying + + =1, TOC is total organic carbon, EC is electrical conductivity, PC is particle content, and the process parameters are unified as follows: is the nanofiltration pressure, is the ratio of concentrated water reuse, is the EDI current.

[0024] Preferably, the population initialization of the genetic algorithm contains 10 to 50 individuals, the crossover probability and the mutation probability are set to 0.8 and 0.2 respectively, and the optimization target parameters include the operating pressure of the nanofiltration device. , Reverse osmosis device concentrated water reuse ratio And the operating current of the EDI device .

[0025] Preferably, the input of the fuzzy logic control includes the deviation of the water quality parameter, specifically:

[0026] ,

[0027] ,

[0028] The output of fuzzy logic control is the process parameter adjustment value Δx, which is calculated as follows:

[0029] ,

[0030] in, and is the output of the fuzzy membership function, and is the fuzzy control gain.

[0031] Preferably, it also includes concentrated water recycling, and the concentrated water recycling is combined with the concentrated water recycling ratio calculated by genetic algorithm optimization, and the concentrated water part of the reverse osmosis device is recycled to the nanofiltration inlet water to achieve efficient utilization of water resources. The concentrated water recycling ratio calculation formula is:

[0032] ,

[0033] in: : solute concentration in concentrated water; : solute concentration in raw water; : Solute concentration in the permeate water.

[0034] Preferably, the dynamic adjustment range of the concentrated water reuse ratio is 10%-30%, and is determined by the optimization result of the genetic algorithm.

[0035] The present invention provides a multi-stage collaborative optimization control method for ultrapure water production. Through real-time monitoring and dynamic optimization, combined with genetic algorithms and fuzzy logic control technology, it achieves technical breakthroughs in many aspects such as improved water quality stability, optimized energy consumption and efficient utilization of resources, especially in the collaborative control and intelligentization of multi-stage processes, showing remarkable effects.

[0036] 1. Technical effects brought about by method improvement:

[0037] Significantly improved water quality stability: Total organic carbon (TOC), electrical conductivity (EC) and particle content (PC) are monitored in real time through a water quality sensor array to capture water quality fluctuations. The operating parameters of the multi-stage process (nanofiltration, reverse osmosis, EDI) are dynamically optimized based on real-time monitoring to avoid substandard water quality caused by raw water quality fluctuations. Quantified effect: TOC fluctuation range is reduced by 20%-25%, EC fluctuation range is reduced by 15%-20%, particle content is controlled below 10 / mL, and water quality compliance rate is increased to more than 98%.

[0038] Improved resource utilization efficiency: Dynamic adjustment of the ratio of concentrated water reuse (10%-30%) can significantly reduce wastewater discharge and improve resource utilization.

[0039] Quantitative effect: The discharge of brine is reduced by 15%-20%, and the proportion of brine reuse is increased by 20%-25%.

[0040] 2. Technical effects of the algorithm alone

[0041] Technical effect of genetic algorithm: Through the global optimization of the operating parameters of nanofiltration, reverse osmosis and EDI (such as pressure, flow rate and concentrated water reuse ratio) by genetic algorithm, the system can adaptively adjust the operating status under various working conditions. Quantitative effect: The system operating pressure is reduced by 10%-12%, the unit energy consumption is reduced by 10%-15%, and the water quality compliance rate is maintained at more than 98%.

[0042] Technical effect of fuzzy logic control: In response to water quality fluctuations and operation deviations, fuzzy logic control is used to make real-time fine-tuning of operating parameters (such as pressure and flow rate), quickly respond to dynamic changes in raw water quality, and make up for the shortcomings of genetic algorithms in terms of real-time performance.

[0043] Quantification of effect: When the water quality fluctuates by 20%-25%, the operating system can still remain stable, and there is no obvious decline in desalination efficiency and water quality indicators.

[0044] 3. Synergistic effect of combining genetic algorithm and fuzzy logic

[0045] Dynamic coordination of global optimization and local fine-tuning: Genetic algorithms provide global optimization operating parameters (such as pressure, flow rate, current, etc.), and fuzzy logic control performs real-time fine-tuning of these optimization results, solving the problem that a single optimization method is difficult to take into account both globality and real-time performance.

[0046] Quantification of the effect: When the water quality fluctuates by up to 25%, the system energy consumption growth is controlled within 3%, and the water quality compliance rate is stabilized at more than 95%.

[0047] Collaborative optimization control of multi-stage processes: Nanofiltration, reverse osmosis and EDI devices form a dynamic closed-loop control system under the synergy of real-time monitoring and algorithm optimization. Through data sharing and feedback adjustment, each process can operate efficiently in the optimal state. Quantitative effect: The system operation efficiency is improved by 15%-20%, and the parameter adjustment time between each device is shortened by 30%-40%, which significantly reduces the water quality fluctuation caused by process switching or load changes.

[0048] Combination of brine reuse and energy consumption optimization: Genetic algorithm dynamically calculates the brine reuse ratio, and fuzzy logic adjusts the flow rate and pressure in real time to ensure maximum resource utilization without causing pollution risks to the previous process. Quantified effect: The brine reuse ratio is increased to more than 25%, while keeping the system energy consumption below 3.5 kWh / m³.

[0049] The present invention effectively solves the problems of insufficient dynamic control, high energy consumption and waste of resources in the ultrapure water preparation process through method improvement, genetic algorithm optimization, fuzzy logic control rapid response, and the synergistic optimization of the combination of the two. The synergistic optimization and closed-loop control of multi-stage processes significantly improve the adaptability and intelligence level of photovoltaic and semiconductor production lines to ultrapure water preparation processes, and meet the needs of high purity, high stability and low energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a structural schematic diagram of the present invention. DETAILED DESCRIPTION

[0051] Embodiment 1:

[0052] A multi-stage collaborative optimization control method for ultrapure water production comprises the following steps:

[0053] (1) Raw water pretreatment: The raw water is preliminarily treated through sand filtration, activated carbon filtration and softening devices to remove suspended particles, organic matter and residual chlorine. The total organic carbon (TOC), electrical conductivity (EC) and particle content (PC) in the raw water are monitored in real time through a water quality sensor array to provide water quality data for subsequent processes.

[0054] (2) Nanofiltration treatment: Use nanofiltration equipment to perform preliminary desalination on the pretreated water. Combined with water quality data, genetic algorithms are used to optimize the operating pressure and flow rate. Fuzzy logic control is used to fine-tune the operating status and dynamically adjust the concentrate reuse ratio.

[0055] (3) Reverse osmosis deep desalination: The nanofiltration effluent is deeply desalinated by reverse osmosis. The genetic algorithm optimizes the operating pressure and the ratio of concentrated water discharge. The fuzzy logic control adjusts the desalination rate in real time.

[0056] (4) Electrodeionization and fine desalination: Electrodeionization (EDI) devices are used to perform final fine desalination on the reverse osmosis effluent. Combined with real-time monitoring data, the operating current and water flow rate are dynamically adjusted to ensure that the ultrapure water quality meets the standards.

[0057] (5) Real-time monitoring and control: Water quality data is obtained through a water quality sensor array, and the operating parameters of the nanofiltration, reverse osmosis and EDI devices are optimized using genetic algorithms. Fuzzy logic control is combined to fine-tune the operating status to form an intelligent dynamic optimization closed loop.

[0058] Embodiment 2:

[0059] The difference between this embodiment and embodiment 1 is that this embodiment also includes the following technical features:

[0060] The real-time monitoring includes using a TOC sensor, a conductivity sensor and a particle sensor, and the monitoring data is transmitted to the control unit through a PLC system for use by the optimization algorithm and the control module.

[0061] Preferably, the genetic algorithm optimizes the process parameter x={ , ,..., }, the fitness function is defined as:

[0062] ,

[0063] in: : Water quality target achievement rate, defined as:

[0064] ,

[0065] E(x): Energy consumption per unit of water production, defined as:

[0066] ,

[0067] in, is the device power, is the running time, is the water production;

[0068] S(x): The fluctuation range of water quality parameters, defined as:

[0069] ,

[0070] , , is the weight coefficient, satisfying + + =1, TOC is total organic carbon, EC is electrical conductivity, PC is particle content, and the process parameters are unified as follows: is the nanofiltration pressure, is the ratio of concentrated water reuse, is the EDI current.

[0071] Preferably, the population initialization of the genetic algorithm contains 10 to 50 individuals, the crossover probability and the mutation probability are set to 0.8 and 0.2 respectively, and the optimization target parameters include the operating pressure of the nanofiltration device. , Reverse osmosis device concentrated water reuse ratio And the operating current of the EDI device .

[0072] Preferably, the input of the fuzzy logic control includes the deviation of the water quality parameter, specifically:

[0073] ,

[0074] ,

[0075] The output of fuzzy logic control is the process parameter adjustment value Δx, which is calculated as follows:

[0076] ,

[0077] in, and is the output of the fuzzy membership function, and is the fuzzy control gain.

[0078] Preferably, it also includes brine reuse, which is combined with the brine reuse ratio calculated by genetic algorithm optimization, and the brine of the reverse osmosis device is partially reused as nanofiltration inlet water to achieve efficient utilization of water resources. The dynamic adjustment range of the brine reuse ratio is 10%-30%, and is determined by the optimization results of the genetic algorithm.

[0079] The calculation formula for the concentrated water reuse ratio is:

[0080] ,

[0081] in: : solute concentration in concentrated water; : solute concentration in raw water; : Solute concentration in the permeate water.

[0082] Embodiment 3:

[0083] The following is an example of the technical solution of the present invention with data and calculation process, demonstrating the effect of optimizing the ultrapure water preparation process through genetic algorithm and fuzzy logic control.

[0084] Example: Optimization of ultrapure water preparation in photovoltaic production lines

[0085] Background conditions

[0086] Raw water parameters:

[0087] Total Organic Carbon (TOC) = 2.5 ppm

[0088] Electrical conductivity (EC) = 500 µS / cm

[0089] Particle content (PC) = 3000 particles / mL

[0090] Target ultrapure water parameters:

[0091] TOC ≤ 50 ppb

[0092] EC ≤ 0.055 µS / cm

[0093] PC ≤ 10 cells / mL

[0094] System settings:

[0095] Nanofiltration (NF): preliminary desalination, initial operating pressure 4 bar, initial flow rate 2.5 m / s.

[0096] Reverse Osmosis (RO): Deep desalination, initial pressure 6 bar, brine discharge ratio 30%.

[0097] Electrodeionization (EDI): fine desalination, initial current 5 A, initial flow rate 2.0 m / s.

[0098] Optimization process

[0099] 1. Data monitoring and input

[0100] The water quality sensor array monitors:

[0101] Nanofiltration outlet: TOC = 1.8 ppm, EC = 200 µS / cm, PC = 1200 pieces / mL;

[0102] RO outlet: TOC = 200 ppb, EC = 2 µS / cm, PC = 150 pieces / mL;

[0103] EDI outlet: TOC = 80 ppb, EC = 0.08 µS / cm, PC = 20 pieces / mL (not up to standard).

[0104] 2. Genetic Algorithm Optimization

[0105] Optimization objectives: maximize water quality target achievement rate Q(x), minimize energy consumption E(x), and control water quality fluctuation S(x).

[0106] Fitness function:

[0107] ,

[0108] Fitness parameters:

[0109] ,

[0110] ,

[0111] ,

[0112] After 10 generations of genetic algorithm iterations:

[0113] Optimization results:

[0114] The nanofiltration pressure was optimized to 3.5 bar, and the flow rate was adjusted to 2.8 m / s;

[0115] The reverse osmosis pressure is optimized to 5.5 bar, and the concentrate ratio is reduced to 25%;

[0116] The EDI current was reduced to 4.2 A and the flow rate was increased to 2.2 m / s.

[0117] 3. Fuzzy logic control fine-tuning

[0118] Fuzzy logic control real-time response:

[0119] TOC deviation: ΔTOC=80−50=30 ppb

[0120] EC deviation: ΔEC = 0.08 − 0.055 = 0.025 μS / cm

[0121] Adjustment results:

[0122] The nanofiltration flow rate was fine-tuned to 2.9 m / s;

[0123] The proportion of reverse osmosis concentrate was further optimized to 22%.

[0124] Optimized results

[0125] Through the above optimization, the system outlet water quality reaches the target:

[0126] TOC: reduced from 80 ppb to 45 ppb;

[0127] EC: from 0.08 µS / cm to 0.050 µS / cm;

[0128] PC: decreased from 20 cells / mL to 8 cells / mL.

[0129] System performance improvement

[0130] Water quality stability:

[0131] The TOC fluctuation range is reduced by 25%, and the EC fluctuation range is reduced by 20%.

[0132] Energy consumption optimization:

[0133] The system's unit water production energy consumption was reduced from 3.5 kWh / m³ to 3.0 kWh / m³, with an energy saving rate of approximately 14%.

[0134] Resource Utilization:

[0135] The proportion of concentrated water reuse increased from 30% to 35%, reducing wastewater discharge by approximately 15%.

[0136] This example shows that by combining genetic algorithm with fuzzy logic control, dynamic closed-loop control can be achieved to achieve water quality compliance, optimize energy consumption and efficiently utilize resources under complex water quality fluctuation conditions.

Claims

1. A multi-stage collaborative optimization control method for ultrapure water production, characterized in that: The following steps are involved: (1) Raw water pretreatment: The raw water is preliminarily treated through sand filtration, activated carbon filtration and softening devices to remove suspended particles, organic matter and residual chlorine. The total organic carbon (TOC), electrical conductivity (EC) and particle content (PC) in the raw water are monitored in real time through a water quality sensor array to provide water quality data for subsequent processes. (2) Nanofiltration treatment: Use nanofiltration equipment to perform preliminary desalination on the pretreated water. Combined with water quality data, genetic algorithms are used to optimize the operating pressure and flow rate. Fuzzy logic control is used to fine-tune the operating status and dynamically adjust the concentrate reuse ratio. (3) Reverse osmosis deep desalination: The nanofiltration effluent is deeply desalinated by reverse osmosis. The genetic algorithm optimizes the operating pressure and the ratio of concentrated water discharge. The fuzzy logic control adjusts the desalination rate in real time. (4) Electrodeionization (EDI) fine desalination: The EDI device is used to perform final fine desalination on the reverse osmosis effluent, and the operating current and water inlet flow rate are dynamically adjusted in combination with real-time monitoring data; (5) Real-time monitoring and control: Water quality data is obtained through a water quality sensor array, and the operating parameters of the nanofiltration, reverse osmosis and EDI devices are optimized using genetic algorithms. Fuzzy logic control is combined to fine-tune the operating status to form an intelligent dynamic optimization closed loop.

2. The multi-stage collaborative optimization control method for ultrapure water production according to claim 1, characterized in that: The real-time monitoring includes using a TOC sensor, a conductivity sensor and a particle sensor, and the monitoring data is transmitted to the control unit through a PLC system for use by the optimization algorithm and the control module.

3. The multi-stage collaborative optimization control method for ultrapure water production according to claim 1, characterized in that: The population initialization of the genetic algorithm includes 10 to 50 individuals, the crossover probability and the mutation probability are set to 0.8 and 0.2 respectively, and the optimization target parameters include nanofiltration pressure, concentrate reuse ratio and EDI current.

4. The multi-stage collaborative optimization control method for ultrapure water production according to claim 1, characterized in that: The input of the fuzzy logic control includes the deviation of water quality parameters, specifically: ， ， The output of fuzzy logic control is the process parameter adjustment value △x, which is calculated as follows: ， in, and is the output of the fuzzy membership function, and is the fuzzy control gain; △TOC is the deviation value of organic carbon concentration, TOC actual To actually measure the total organic carbon concentration, TOC target is the target total organic carbon concentration; △EC is the deviation value of conductivity; EC actual To actually measure conductivity; EC target is the target conductivity.

5. The multi-stage collaborative optimization control method for ultrapure water production according to claim 1, characterized in that: It also includes concentrated water reuse, which combines concentrated water reuse ratio calculated by genetic algorithm optimization to reuse concentrated water part of the reverse osmosis device as nanofiltration inlet water.

6. The multi-stage collaborative optimization control method for ultrapure water production according to claim 5, characterized in that: The calculation formula for the concentrated water reuse ratio is: ， Where: C conc. : solute concentration in concentrated water; C feed : solute concentration in raw water; C perm. : Solute concentration in the permeate water.

7. The multi-stage collaborative optimization control method for ultrapure water production according to claim 5, characterized in that: The dynamic adjustment range of the concentrated water reuse ratio is 10%-30%, and is determined by the optimization results of the genetic algorithm.

Images