Ultrafiltration system control optimization method

Abstract

The invention provides an ultrafiltration system control optimization method, and relates to the technical field of water treatment.The method comprises the steps that a chemical cleaning interval serves as an operation cycle, and after each operation cycle is finished, data such as cycle water yield Q and ton water power consumption E are collected; calculating a comprehensive operation score S according to a pre-defined score function model S = w1F (Q) + w2G (E); the S is compared with a preset score threshold value, and differential adjustment (keeping, fine adjustment or great adjustment) is carried out on the control parameters of the next period according to the comparison result. The scoring function model can be expanded to a multi-objective form S = w1F (Q) + w2G (E) + w3H (delta) containing a membrane state scoring item H (delta). The weight coefficients w1 and w2 can be dynamically configured according to the water inlet pollution index SDI value through a mapping relation. According to the method, self-adaptive optimization of the cleaning strategy is achieved through long-period overall evaluation and multi-target scoring, the water production capacity of the ultrafiltration system can be improved at the same time, energy consumption is reduced, and membrane pollution is delayed.

Description

Technical Field

[0001] This invention relates to the field of water treatment technology, and in particular to a method for controlling and optimizing an ultrafiltration system. Background Technology

[0002] In the field of ultrafiltration water treatment, the operating efficiency of an ultrafiltration system is closely related to the cleaning strategy. Currently, the mainstream control methods for ultrafiltration systems mainly rely on two types: one is to operate with fixed parameters such as backwash cycle and duration; the other is for operators to manually adjust the parameters based on experience.

[0003] However, both of these approaches have significant limitations: fixed-parameter operation cannot adapt to changes in influent water quality, water temperature, and membrane fouling status, often leading to insufficient cleaning (resulting in increased membrane fouling and decreased permeate production) or excessive cleaning (resulting in shortened permeate production time and increased energy and water consumption); manual adjustment relies heavily on personal experience, has a delayed response, and makes it difficult to quantify and weigh multiple interdependent objectives such as "permeate production," "energy consumption," and "long-term membrane health," resulting in unstable operating strategies and high total lifecycle costs.

[0004] The root cause is that existing technologies lack a model capable of quantitatively evaluating the overall operational performance of a complete operating cycle (i.e., the time period between two adjacent chemical cleaning processes), and even less capable of automatically and intelligently guiding the optimization and adjustment of operating parameters for the next operating cycle based on this evaluation result. This causes ultrafiltration systems to operate in a suboptimal state for extended periods, hindering further improvements in their energy efficiency and economic benefits. Summary of the Invention

[0005] To address the problems existing in the prior art, the present invention provides a method for controlling and optimizing an ultrafiltration system, comprising the following steps:

[0006] S1. Execute the control parameters of the ultrafiltration system in the current operating cycle and collect the operating result data of the current operating cycle. The operating cycle is the time period between two adjacent chemical cleanings. The control parameters include parameters for adjusting the physical cleaning operation characteristics within the operating cycle. The operating result data includes the total water production Q and the power consumption per ton of water E in the current operating cycle.

[0007] S2. Based on the operation result data of the current operation cycle, calculate the comprehensive operation score S of the current cycle according to a predefined scoring function model, wherein the scoring function model is:

[0008] S = w1•F(Q) + w2•G(E);

[0009] Where F(Q) is the water production score item, G(E) is the energy consumption score item, w1 is the water production weight coefficient, w2 is the energy consumption weight coefficient, and w1+w2=1;

[0010] S3. Compare the comprehensive operation score S with a preset score threshold. The score threshold includes a first score threshold S_high and a second score threshold S_low, and S_high > S_low;

[0011] S4. Adjust the control parameters for the next operation cycle according to the comparison result:

[0012] If S ≥ S_high, keep the control parameters unchanged in the next operation cycle;

[0013] If S_low ≤ S < S_high, slightly adjust the control parameters according to the first preset adjustment rule in the next operation cycle;

[0014] If S < S_low, greatly adjust the control parameters according to the second preset adjustment rule in the next operation cycle.

[0015] Optionally, the control parameters further include parameters for controlling the maintenance chemical cleaning process.

[0016] Optionally, the water production score item F(Q) = Q / Q_ref, the energy consumption score item G(E) = 1 - E / E_ref, E_ref is the reference benchmark value of power consumption per ton of water, Q_ref is the reference benchmark value of water production, and the value-taking methods of Q_ref and E_ref include any one of the following:

[0017] Adopt a preset fixed constant;

[0018] Or,

[0019] Adopt the arithmetic mean of the corresponding data of the previous M operation cycles, where M ≥ 3;

[0020] Or,

[0021] Adopt the moving average of the corresponding data of the previous M operation cycles.

[0022] Optionally, the operation result data further includes the transmembrane pressure difference growth rate Δ during the current operation cycle;

[0023] Correspondingly, the score function model is:

[0024] S = w1•F(Q) + w2•G(E) + w3•H(Δ);

[0025] Among them, H(Δ) is the membrane state score item, w3 is the membrane state weight coefficient, and w1 + w2 + w3 = 1.

[0026] Optionally, the membrane state rating item H(Δ) = 1 - Δ / Δ_ref, where Δ_ref is a reference value for the rate of increase of the permeate pressure difference, and the value of Δ_ref is obtained through experimental calibration based on the material and designed service life of the ultrafiltration membrane element.

[0027] Optionally, the method for determining the water production weighting coefficient w1 and the energy consumption weighting coefficient w2 includes:

[0028] Obtain the SDI (Spectrum Indices) of the influent water quality of the ultrafiltration system;

[0029] Based on the preset pollution interval to which the pollution index SDI belongs, a set of intermediate weight coefficients w1_temp and w2_temp are determined, wherein the preset pollution interval is divided into at least three consecutive pollution intervals, and each pollution interval corresponds to a set of preset intermediate weight coefficients w1_temp and w2_temp.

[0030] The intermediate weight coefficients are normalized to obtain the final weight coefficients:

[0031] w1=w1_temp / (w1_temp+w2_temp);

[0032] w2=w2_temp / (w1_temp+w2_temp).

[0033] Optionally, the method for determining the membrane state weighting coefficient w3 includes any of the following:

[0034] Use a preset fixed constant;

[0035] or,

[0036] Determined based on the cumulative operating time of the ultrafiltration membrane element;

[0037] or,

[0038] It is determined based on the long-term trend of the permeabilization pressure differential growth rate Δ during historical operating cycles.

[0039] Optionally, the parameters used to adjust the characteristics of the physical cleaning operation include the backwash cycle T and the backwash duration t.

[0040] Accordingly, the first preset adjustment rule is: the backwashing cycle T is adjusted to T*k1, and the backwashing duration t is adjusted to t*k2; where 0.95≤k1≤1.05, 0.95≤k2≤1.05, and k1 and k2 are not both 1;

[0041] The second preset adjustment rule is: adjust the backwashing cycle T to T*K1, and adjust the backwashing duration t to t*K2; where 0.8≤K1≤0.95, 1.05≤K2≤1.2.

[0042] Optionally, the method further includes: after running N consecutive cycles, calibrating the first scoring threshold S_high and the second scoring threshold S_low according to a preset statistical update rule, including:

[0043] Update the first scoring threshold S_high to α times the highest value of the comprehensive running score S in the past N running cycles, where 0 < α < 1;

[0044] The second scoring threshold S_low is updated to be β times the average of the comprehensive running scores S over the past N running cycles, where 0 < β < 1.

[0045] By adopting the above technical solution, the present invention has at least the following beneficial effects:

[0046] 1. This invention achieves long-term overall optimization of ultrafiltration systems: breaking through the limitations of traditional real-time feedback control, it uses the chemical cleaning cycle as the optimization unit and anchors the optimization target on overall indicators such as total water production per cycle and average energy consumption per cycle, thus avoiding the drawbacks of traditional real-time optimization that may lead to accelerated membrane fouling and damage to the long-term stability of the system in pursuit of instantaneous indicators.

[0047] 2. This invention overcomes the technical challenge of multi-objective collaborative optimization: by constructing a comprehensive scoring model that integrates water production (F(Q)), energy consumption (G(E)) and membrane state (H(Δ)), and innovatively introducing a dynamic weight configuration mechanism based on water quality (SDI), the ultrafiltration system can intelligently and dynamically adjust the optimization focus (whether to prioritize water production or energy saving and membrane preservation) when water quality changes, thus achieving the best balance between multiple conflicting objectives.

[0048] 3. This invention possesses intelligent adaptive capabilities based on explicit rules: through a pre-set comprehensive scoring model, mapping table, and adjustment rules, the ultrafiltration system acquires adaptive optimization capabilities similar to expert experience. Its core innovation lies in the transparency, controllability, and predictability of this intelligence. For example, through statistical update rules, the scoring threshold can be automatically recalibrated to follow system performance drift, always maintaining the rationality of the evaluation criteria. This rule-driven intelligence ensures the stability and reliability of the ultrafiltration system. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart illustrating a method for controlling and optimizing an ultrafiltration system, as provided in an embodiment of this disclosure. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Example 1

[0053] This embodiment fully demonstrates the entire process from data acquisition to parameter adjustment through a specific example, including:

[0054] Step S1: Execution and Data Collection

[0055] The system executes the control parameters of the ultrafiltration system in the current operating cycle and collects the operating result data for that cycle. The operating cycle is the time period between two adjacent chemical cleaning operations.

[0056] The control parameters include parameters used to adjust the physical cleaning operation characteristics within the operating cycle, including a backwash cycle of T=30 minutes and a backwash duration of t=60 seconds.

[0057] After the operation cycle ended, the collected operational data included: total water production Q = 2550 m³, power consumption per ton of water E = 0.38 kWh / m³, and membrane pressure differential growth rate Δ = 3.0 kPa / h. The current water pollution index SDI = 4.2.

[0058] Step S2: Calculate the comprehensive performance score S based on the predefined scoring function model.

[0059] Based on the data collected in step S1, the comprehensive operating score S for the current period is calculated according to a predefined scoring function model with fixed parameters.

[0060] Calculate each scoring item:

[0061] Set the water production reference value Q_ref=3000m³, and calculate the water production score item: F(Q)=Q / Q_ref=2550 / 3000=0.85.

[0062] Set the reference benchmark value for electricity consumption per ton of water to E_ref=0.45kWh / m³, and calculate the energy consumption score item: G(E)=1-E / E_ref=1-0.38 / 0.45≈0.156.

[0063] Set the reference value of the transmembrane pressure difference growth rate Δ_ref = 5.0 kPa / h (calibrated through experiments according to the membrane material), and calculate the membrane state scoring item: H(Δ)=1 - Δ / Δ_ref = 1 - 3.0 / 5.0 = 0.4.

[0064] Determine the weight coefficients:

[0065] The ultrafiltration system pre-stores the SDI-weight mapping relationship for determining a set of intermediate weight coefficients according to the influent SDI value. For example, a feasible mapping relationship is: when the SDI value is higher, the value of w2_temp is relatively larger (i.e., more emphasis on energy consumption and membrane protection). For example, the medium interval is preset as (0.7, 0.3); for a higher and poorer interval, a larger w2_temp value will be preset.

[0066] According to the current SDI = 4.2, query the SDI-weight mapping relationship and know that it belongs to the medium interval, and the corresponding intermediate weight coefficients are w1_temp = 0.7 and w2_temp = 0.3.

[0067] Set the membrane state weight w3 = 0.1.

[0068] Normalize the above weights to obtain the final weight coefficients:

[0069] w1 = 0.7 / (0.7 + 0.3 + 0.1) ≈ 0.636, w2 = 0.3 / 1.1 ≈ 0.273, w3 = 0.1 / 1.1 ≈ 0.091.

[0070] Calculate the comprehensive operation score S:

[0071] According to the scoring function model: S = w1•F(Q)+w2•G(E)+w3•H(Δ), substitute and calculate: S = 0.636*0.85 + 0.273*0.156 + 0.091*0.4 = 0.541 + 0.043 + 0.036 = 0.620.

[0072] Step S3: Compare the comprehensive operation score S with the preset score threshold

[0073] The preset score thresholds include: the first score threshold S_high = 0.85 and the second score threshold S_low = 0.70, and S_high > S_low.

[0074] Compare the calculated S = 0.620 with the preset score threshold: Since S < S_low, enter the corresponding adjustment branch.

[0075] Step S4: Adjust the control parameters according to the comparison result

[0076] Since S < S_low, it is determined that the current control parameters need to be optimized and should be adjusted significantly according to the second preset adjustment rule.

[0077] The second preset adjustment rule is: adjust the backwashing period T to T * K1, and adjust the backwashing duration t to t * K2. Set K1 = 0.9 and K2 = 1.1.

[0078] Therefore, the control parameters for the next operation cycle are automatically adjusted to: the new backwashing period T' = 30 min * 0.9 = 27 min, and the new backwashing duration t' = 60 s * 1.1 = 66 s.

[0079] After the ultrafiltration system runs continuously for N = 5 cycles according to this rule, its each operation cycle is successively [0.75, 0.68, 0.71, 0.52, 0.64]. The system will recalibrate the first scoring threshold S_high and the second scoring threshold S_low according to the preset statistical update rule, update S_high to α = 0.9 times of the historical highest score 0.75, that is, 0.675; update S_low to β = 0.8 times of the historical average score 0.66, that is, 0.528. The updated S_high and S_low will be used for score comparison in subsequent cycles.

[0080] Embodiment 2

[0081] To realize the intelligent configuration of the weight coefficients w1 and w2, the system pre-stores the SDI-weight mapping relationship, as shown in Table 1.

[0082] Table 1 SDI-weight mapping relationship

[0083]

[0084] After the above intermediate weights are normalized as w1 = w1_temp / (w1_temp + w2_temp) and w2 = w2_temp / (w1_temp + w2_temp), the final weight coefficients are obtained. Table 1 is only an example, and the specific values can be adjusted according to the actual situation of the project.

[0085] When the measured SDI = 4.5 (in the medium range), take w1_temp = 0.65 and w2_temp = 0.35. If the system sets the membrane state weight w3 = 0.1, then perform normalization: w1 = 0.65 / (0.65 + 0.35 + 0.1) ≈ 0.591, w2 = 0.35 / 1.1 ≈ 0.318, w3 = 0.1 / 1.1 ≈ 0.091. This mapping relationship enables the system to automatically adjust the optimization strategy according to the water quality.

[0086] Embodiment 3

[0087] Taking the operation of an ultrafiltration system in a wastewater reuse project under poor influent water quality as an example, this paper demonstrates the calculation and decision-making process of a complete cycle of the method of the present invention. The system is designed to produce water at a flow rate of 220 m³ / h. 3 / h, the duration of this operating cycle is 17 hours. The total water production during the cycle is Q=2676m³. 3 Electricity consumption per ton of water E = 0.38 kWh / m 3 Due to poor influent water quality, the transmembrane pressure differential (TMP) increased from 65 kPa to 183 kPa during the cycle. The calculated TMP growth rate was Δ = (183 - 65) / 17 ≈ 6.94 kPa / h. The current influent pollution index (SDI) was 5.5. The system's preset reference baseline value was: permeate flow rate baseline Q_ref = 3000 m³ / h. 3 The benchmark for electricity consumption per ton of water is E_ref = 0.45 kWh / m³. 3 The baseline permeasurand pressure differential growth rate is Δ_ref = 5.0 kPa / h. The initial system control parameters are: backwash cycle T = 30 min, backwash duration t = 60 s, first scoring threshold S_high = 0.85, and second scoring threshold S_low = 0.70.

[0088] First, calculate the scores for each component: permeate production score F(Q) = 2676 / 3000 = 0.892, energy consumption score G(E) = 1 - 0.38 / 0.45 ≈ 0.156, and membrane state score H(Δ) = 1 - 6.94 / 5.0 ≈ -0.388. Since SDI = 5.5, referring to Table 1, it falls within the poor range, corresponding to intermediate weights w1_temp = 0.6 and w2_temp = 0.4. Setting the membrane state weight w3 = 0.1, after normalization, we get w1 = 0.6 / (0.6 + 0.4 + 0.1) ≈ 0.545, w2 = 0.4 / 1.1 ≈ 0.364, and w3 = 0.1 / 1.1 ≈ 0.091. The overall operating score was then calculated as S = 0.545 * 0.892 + 0.364 * 0.156 + 0.091 * (-0.388) ≈ 0.486 + 0.057 - 0.035 ≈ 0.508. Since the overall operating score S ≈ 0.508 is less than the second scoring threshold S_low (0.70), the current control parameters are determined to need optimization. According to the second preset adjustment rule, adjustment coefficients K1 = 0.9 and K2 = 1.1 are taken. The system automatically adjusts the control parameters for the next cycle to: backwash cycle T' = 30 * 0.9 = 27 min, backwash duration t' = 60 * 1.1 = 66 s.

[0089] From an experiential perspective, when water quality deteriorates and pressure differential increases rapidly, it is feasible and necessary to shorten the backwash cycle and extend the backwash time in a timely manner; otherwise, irreversible losses may easily occur.

[0090] Example 4

[0091] In this embodiment, the ultrafiltration system operated according to this rule for 5 cycles (N=5), and its historical scores were 0.75, 0.68, 0.71, 0.52, and 0.64, respectively. The system will then automatically recalibrate the first scoring threshold S_high and the second scoring threshold S_low according to statistical update rules (e.g., α=0.9, β=0.8): S_high will be updated to 0.9 times the historical highest score of 0.75, i.e., 0.675; S_low will be updated to 0.8 times the historical average score of 0.66, i.e., 0.528. The updated thresholds will be used for score comparison in the next stage.

[0092] Example 5

[0093] The ultrafiltration module used in this embodiment has 76 membrane elements per module, with each membrane element having an area of ​​50m². 2 The ultrafiltration membrane is an imported membrane with a pore size of 0.1 microns, made of PVDF material, cylindrical membrane fibers, and external pressure full-volume filtration.

[0094] Designed single-unit inlet water volume is 220m³. 3 / h, factory standard membrane flux 160L / m 2 / h, inlet pressure 120-240kPa; the raw water is Class A compliant effluent from a wastewater treatment plant, which enters the ultrafiltration after coagulation and sedimentation in the high-efficiency tank; initially, the manual control and the initial parameters of the method of this invention are the same.

[0095] Four ultrafiltration systems were compared. Two systems were compared during the summer, using both manual control and the method of this invention. 90 days of operational data were analyzed. Before comparison, both ultrafiltration systems were manually controlled with identical operating parameters: w1_temp=0.8, w2_temp=0.2, w3=0.05. The other two systems were compared during the winter, again using both manual control and the method of this invention. 90 days of operational data were analyzed. Before comparison, both ultrafiltration systems were manually controlled with identical operating parameters: w1_temp=0.7, w2_temp=0.3, w3=0.05. Manual control involved experienced engineers manually adjusting the operating parameters according to standard procedures.

[0096] Analyze 90-day operating data for both winter (December-February) and summer (July-September). The time for water production is shorter than the analysis period; the difference represents the time consumed for cleaning and backwashing. The amount of water produced is shorter than the net water production; the difference represents the amount of water used for cleaning. It can be seen that in summer, this invention increases net water production by 10.53% and reduces pressure differential by 2.06% (energy saving); in winter, this invention increases net water production by 9.03% and reduces pressure differential by 1.53% (energy saving), as shown in Table 2.

[0097] Table 2. Operational data for 90 days each in winter (December-February) and summer (July-September)

[0098]

[0099] To further analyze the reasons for the performance differences, the average values ​​of the main control parameters during the operation period were retrieved and statistically analyzed within the system, as shown in Table 3.

[0100] Table 3. Average values ​​of key control parameters during winter (December-February) and summer (July-September) operation.

[0101]

[0102] Data shows that, under the control of the method of this invention, the system adopts longer chemical cleaning intervals and backwashing intervals, as well as shorter backwashing times. This indicates that, through the optimization model of this invention, the system reduces non-productive time while ensuring cleaning effectiveness, thereby increasing effective water production time and total water production.

[0103] The short-term comparative test in this embodiment shows that, under different seasonal water quality conditions, the method of the present invention can adjust the cleaning strategy compared with traditional manual experience control, and obtain a higher system water production time and net water production while maintaining or reducing the operating pressure difference.

[0104] Example 6

[0105] The ultrafiltration module used in this embodiment has 76 membrane elements per module, with each membrane element having an area of ​​50.12 m². 2 The ultrafiltration membrane is a domestically produced membrane with a pore size of 0.1 microns, made of PVDF material, cylindrical membrane fibers, and external pressure full-volume filtration.

[0106] The design capacity for a single inlet is 215m³. 3 / h, factory standard membrane flux 150L / m 2 / h, inlet pressure 120-250kPa; the raw water is the quasi-IV standard effluent from a sewage treatment plant, which enters the ultrafiltration after coagulation and sedimentation in the high-efficiency tank; at the beginning, the manual control and the initial parameters of the method of this invention are the same.

[0107] To ensure a rigorous comparison, four parallel and identical ultrafiltration units were set up and operated synchronously under the same influent conditions:

[0108] Control group: The conventional manual control strategy based on human experience was adopted.

[0109] Experimental group 1: Using the method of this invention, the intermediate weight values ​​w1_temp=0.8, w2_temp=0.2, and membrane state weight w3=0 were set.

[0110] Experimental group 2: Using the method of this invention, the same values ​​were set: w1_temp=0.8, w2_temp=0.2, and membrane state weight w3=0.1.

[0111] Experimental group 3: Using the method of this invention, the same values ​​were set: w1_temp=0.8, w2_temp=0.2, and membrane state weight w3=0.15.

[0112] All experimental groups had the same initial operating parameters as the control group to eliminate differences in initial conditions.

[0113] A comparative analysis of operational data over 370 days was conducted, including 8640 hours of production time and the remainder being maintenance time. Key data comparisons after cumulative operation are shown in Table 4 below.

[0114] Table 4 Comparison of key data after cumulative operation

[0115]

[0116] Even in the simplest mode without considering membrane state (w3=0), this invention, through long-cycle scoring and rule adjustment, has significantly improved water production time and water production (+13.63%), and reduced operating pressure drop (energy saving 6.03%). This demonstrates the effectiveness of the core optimization framework constructed by claims 1-3, 8.

[0117] To directly quantify the performance degradation of membrane materials, the industry-recognized standardized membrane flux (permeate flux under conditions of 25℃ and 100kPa) was used as the core evaluation index. The membrane elements of each system were tested after the operation was completed, and the results are shown in Table 5.

[0118] Table 5. Performance Comparison of Membrane Materials After Cumulative Operation

[0119]

[0120] Comparing experimental group 1 (w3=0) and experimental group 2 (w3=0.1), it can be seen that although the permeate production was slightly sacrificed (from +13.63% to +11.69%), the standardized membrane flux achieved a significant leap from 1.16% to 7.12%. This indicates that the introduction of the membrane state weight w3 and the role of the H(Δ) scoring term (see claims 4 and 5) are not simply supplements to the permeate / energy consumption model. The data reveals a fundamental shift, not obvious but of great engineering value—from pursuing short-term output maximization to pursuing optimal life-cycle cost. The long-term benefits of membrane health resulting from this shift far outweigh the short-term permeate loss.

[0121] The basic optimization framework provided by this invention can significantly improve system capacity and energy efficiency. More importantly, the "membrane state scoring model" (H(Δ)) and the independently configurable "membrane state weight" (w3) of this invention enable the system to make a scientific and quantifiable trade-off between "short-term water production benefits" and "long-term asset health". From the perspective of total life cycle cost, the strategy of w3=0.1 (experimental group 2) achieves an improvement of more than 7% in membrane performance retention and nearly 10% energy saving by sacrificing only about 2% increase in water production, demonstrating excellent comprehensive benefits. By adjusting w3, a flexible switch from "high-production mode" to "membrane protection mode" can be achieved. This function solves the dilemma of "ensuring production" and "ensuring equipment" that has long been faced by the industry. The preferred range of w3=0.1 to 0.15 is the range that achieves the best techno-economic balance found in long-term practice. The data in this embodiment show that by introducing a membrane state scoring model and configuring an appropriate w3 value, the long-term performance retention rate of the membrane can be significantly improved and the operating pressure drop can be reduced with a slight impact on short-term water production. This synergistic optimization effect is difficult to achieve by conventional methods such as single parameter adjustment or traditional experience control.

[0122] In summary, this invention provides an optimization framework based on explicit rules and prior knowledge. Its "intelligence" lies in encoding expert experience (such as prioritizing energy consumption when SDI is high) into preset mapping tables and adjustment rules. This method has outstanding advantages such as logical transparency, computational efficiency, rapid deployment, and stable and interpretable results, making it particularly suitable for control scenarios in industrial settings with extremely high reliability requirements.

[0123] It is important to note that the core protection of this invention lies in the aforementioned "predefined and parameter-fixed" rule-based implementation. This is fundamentally different from the technical path of using data-driven models such as machine learning (e.g., neural networks) and reinforcement learning to iterate during operation and generate black-box decision-making solutions. While the latter may possess stronger adaptive potential, it typically faces challenges such as large training data requirements, uncertain model generalization, and uninterpretable decision-making processes. This invention complements, rather than encompasses, these potential data-driven solutions, each applicable to different technical needs and scenarios.

[0124] The present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A method for controlling and optimizing an ultrafiltration system, characterized in that, It includes the following steps: S1. Execute the control parameters of the ultrafiltration system in the current operation cycle and collect the operation result data of the current operation cycle. Here, the operation cycle is the time period between two adjacent chemical cleanings. The control parameters include the parameters for adjusting the physical cleaning operation characteristics within the operation cycle. The operation result data includes the total water production Q in the current operation cycle and the power consumption per ton of water E in the current operation cycle; S2. Based on the operation result data of the current operation cycle, calculate the comprehensive operation score S of the current cycle according to a pre-defined scoring function model. The scoring function model is: S = w1•F(Q) + w2•G(E); where, F(Q) is the water production scoring item, G(E) is the energy consumption scoring item, w1 is the water production weight coefficient, w2 is the energy consumption weight coefficient, and w1 + w2 = 1; S3. Compare the comprehensive operation score S with a preset scoring threshold. Here, the scoring threshold includes the first scoring threshold S_high and the second scoring threshold S_low, and S_high > S_low; S4. Adjust the control parameters of the next operation cycle according to the comparison result: If S ≥ S_high, keep the control parameters unchanged in the next operation cycle; If S_low ≤ S < S_high, slightly adjust the control parameters according to the first preset adjustment rule in the next operation cycle; If S < S_low, greatly adjust the control parameters according to the second preset adjustment rule in the next operation cycle.

2. The ultrafiltration system control optimization method according to claim 1, characterized in that, The control parameters also include the parameters for controlling the maintenance chemical cleaning process.

3. The ultrafiltration system control optimization method according to claim 1, characterized in that, The water production scoring item F(Q) = Q / Q_ref, the energy consumption scoring item G(E) = 1 - E / E_ref. E_ref is the reference benchmark value of the power consumption per ton of water, Q_ref is the reference benchmark value of the water production. The value-taking methods of Q_ref and E_ref include any one of the following: Adopt a preset fixed constant; Or, Adopt the arithmetic mean of the corresponding data of the previous M operation cycles, where M ≥ 3; Or, Adopt the moving average of the corresponding data of the previous M operation cycles.

4. The ultrafiltration system control optimization method according to claim 1 or 3, characterized in that, The operation result data also includes the transmembrane pressure difference growth rate Δ in the current operation cycle; Correspondingly, the scoring function model is: S = w1•F(Q) + w2•G(E) + w3•H(Δ); where, H(Δ) is the membrane state scoring item, w3 is the membrane state weight coefficient, and w1 + w2 + w3 = 1.

5. The ultrafiltration system control optimization method according to claim 4, characterized in that, The membrane state scoring item H(Δ) = 1 - Δ / Δ_ref, Δ_ref is the reference benchmark value of the transmembrane pressure difference growth rate, and the value of Δ_ref is obtained through experimental calibration according to the material and designed service life of the ultrafiltration membrane element.

6. The ultrafiltration system control optimization method according to claim 1, characterized in that, The determination methods of the water production weight coefficient w1 and the energy consumption weight coefficient w2 include: Obtain the pollution index SDI of the ultrafiltration system inlet water quality; Based on the preset pollution interval to which the pollution index SDI belongs, a set of intermediate weight coefficients w1_temp and w2_temp are determined, wherein the preset pollution interval is divided into at least three consecutive pollution intervals, and each pollution interval corresponds to a set of preset intermediate weight coefficients w1_temp and w2_temp. The intermediate weight coefficients are normalized to obtain the final weight coefficients: w1=w1_temp / (w1_temp+w2_temp); w2=w2_temp / (w1_temp+w2_temp).

7. The ultrafiltration system control optimization method according to claim 4 or 5, characterized in that, The method for determining the membrane state weighting coefficient w3 includes any of the following: Use a preset fixed constant; or, Determined based on the cumulative operating time of the ultrafiltration membrane element; or, It is determined based on the long-term trend of the permeabilization pressure difference growth rate Δ during historical operating cycles.

8. The ultrafiltration system control optimization method according to claim 1 or 2, characterized in that, The parameters used to adjust the characteristics of physical cleaning operations include the backwash cycle T and the backwash duration t. Accordingly, the first preset adjustment rule is: the backwashing cycle T is adjusted to T*k1, and the backwashing duration t is adjusted to t*k2; where 0.95≤k1≤1.05, 0.95≤k2≤1.05, and k1 and k2 are not both 1; The second preset adjustment rule is: adjust the backwashing cycle T to T*K1, and adjust the backwashing duration t to t*K2; where 0.8≤K1≤0.95, 1.05≤K2≤1.

2.

9. The ultrafiltration system control optimization method according to claim 1, characterized in that, The method further includes: after running for N consecutive cycles, calibrating the first scoring threshold S_high and the second scoring threshold S_low according to a preset statistical update rule, including: The first scoring threshold S_high is updated to α times the highest value of the comprehensive running score S in the past N running cycles, where 0 < α < 1; The second scoring threshold S_low is updated to be β times the average of the comprehensive running scores S over the past N running cycles, where 0 < β < 1.

Images