Membrane control method and system based on multi-parameter intelligent perception

Abstract

The invention relates to the technical field of membrane control, and discloses a membrane control method and system based on multi-parameter intelligent perception, and the method comprises the steps: obtaining multi-parameter operation data of a membrane separation system, and calculating a transmembrane pressure difference change value, an inlet and outlet water quality change rate and membrane operation load duration according to the operation data; based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate and the membrane operation load duration, a pollution trend early warning model based on an improved long and short term memory network is constructed and analyzed, and a first prediction result is obtained; and performing dynamic pollution risk score calculation based on the first prediction result to obtain a pollution risk comprehensive assessment value. Linkage analysis of the pollution development trend and the system risk level is achieved, the potential pollution risk level can be recognized in advance, cleaning intervention suggestions or alarm strategies are output in a graded mode, and the initiative and timeliness of cleaning decision making are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of membrane control technology, and in particular to a membrane control method and system based on multi-parameter intelligent perception. Background Art

[0002] In recent years, in the field of water treatment and resource reuse, membrane separation technology has been widely used in scenarios such as deep sewage treatment, industrial circulating water purification, and drinking water preparation due to its efficient interception characteristics. The operating efficiency of the membrane system is closely related to its pollution status. Therefore, how to accurately judge the membrane pollution status and intervene in time has become an important issue to improve system stability and reduce operating costs. With the development of intelligent perception and data-driven methods, the use of operating parameters to achieve intelligent monitoring and control of membrane status has become a research hotspot, especially in the fields of multi-parameter fusion and trend prediction, which has broad application prospects.

[0003] Although existing membrane control methods have certain responsiveness, they generally lack dynamic modeling and prediction mechanisms for pollution trends, and are unable to effectively identify the continuity of pollution development and potential change directions. Especially in the early stages of pollution, the system is often unable to perceive the gradual nature of pollution signals in a timely manner, resulting in inaccurate judgment of the timing of cleaning; in the stage of intensified pollution, it may also be delayed due to insufficient prediction capabilities, affecting the long-term operating stability and cleaning efficiency of the membrane system. Therefore, it is necessary to design a membrane control method and system based on multi-parameter intelligent perception to solve the above problems. Summary of the Invention

[0004] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0005] In view of the above-mentioned problems existing in the existing membrane control method and system based on multi-parameter intelligent perception, the present invention is proposed.

[0006] Therefore, the purpose of the present invention is to provide a membrane control method and system based on multi-parameter intelligent perception, which is suitable for solving the problems of lack of dynamic modeling and prediction mechanism of pollution trends, inability to effectively identify the continuity and potential change direction of pollution development, resulting in inaccurate judgment of cleaning timing, affecting the long-term operation stability and cleaning efficiency of the membrane system.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides a membrane control method based on multi-parameter intelligent perception, comprising:

[0009] Acquiring multi-parameter operating data of the membrane separation system, and calculating the transmembrane pressure difference change value, the rate of change of inlet and outlet water quality, and the membrane operation load duration based on the operating data;

[0010] Based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate, and the membrane operation load duration, a pollution trend early warning model based on an improved long short-term memory network is constructed and analyzed to obtain a first prediction result;

[0011] Performing a dynamic pollution risk score calculation based on the first prediction result to obtain a comprehensive pollution risk assessment value;

[0012] A comprehensive judgment of the membrane fouling trend is made based on the first prediction result and the comprehensive assessment value of the fouling risk, and a cleaning intervention is performed.

[0013] As a preferred solution of the membrane control method based on multi-parameter intelligent perception described in the present invention, wherein: the multi-parameter operation data refers to the operation process parameter data obtained by sensors arranged in the membrane separation system;

[0014] The transmembrane pressure difference change value refers to the change amplitude of the difference between the membrane water inlet side pressure and the water outlet side pressure per unit time;

[0015] The rate of change of water quality of the inlet and outlet water refers to the rate of change per unit time of the difference between the key water quality indicators (including turbidity and conductivity) of the membrane inlet and outlet water;

[0016] The membrane operation load duration refers to the length of time the membrane system continuously operates without performing a cleaning intervention operation.

[0017] As a preferred embodiment of the membrane control method based on multi-parameter intelligent perception described in the present invention, a pollution trend warning model is constructed based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate, and the membrane operation load duration, including the following steps:

[0018] Construct the pollution characteristic vector X(t) based on multi-parameter operation data, including the transmembrane pressure difference change rate ΔP(t), inlet and outlet water flow rate Q in (t) and Q out (t), water quality index C COD (t) and C TSS (t) and membrane operation duration τ(t), which are specifically expressed as follows:

[0019] X(t)=[ΔP(t),Q in (t), Q out (t), C COD (t), C TSS (t), τ(t)] T ;

[0020] Introducing the 6th-order pollution weight matrix And define the pollution feature matrix scoring item as:

[0021]

[0022] Among them, M(t) is the pollution feature matrix scoring item, is a sigmoid mapping function, which is used to map the pollution intensity to the standardized interval [0,1]. The transmembrane pressure difference disturbance limit average function term is introduced, and the disturbance accumulation expression based on the Gaussian error function is constructed. The specific formula is:

[0023]

[0024] in:

[0025]

[0026] Among them, D(t) is the historical disturbance function term, which is used to represent the impact of historical pressure difference fluctuations on the membrane fouling trend. Ψ(x) is the Gaussian error function, which is used to simulate the cumulative effect of random disturbances. Λ(t) represents the average value of the historical disturbance limit of the transmembrane pressure difference at the tth moment, which is used to reflect the cumulative trend of small disturbances in the evolution of membrane fouling. ΔP(tk) represents the rate of change of the transmembrane pressure difference at the tth moment. n is the selected historical length window. u represents the integral variable, which is used to calculate the probability accumulation function of the disturbance intensity. Considering that the membrane fouling process continues to accumulate with the operating time, the pollution load accumulation term is introduced into the pollution trend warning model. The specific formula is as follows:

[0027]

[0028] Among them, R(t) is the cumulative pollution load term, which represents the total organic pollution intensity experienced by the membrane before time t, and C COD (s) is the influent COD concentration at the sth moment, C TSS (s) is the influent TSS concentration at the sth moment, e -s It is an exponential decay function, which indicates the decreasing trend of pollutant impact over time. Based on the obtained pollution characteristic matrix score item, historical disturbance function item and pollution load accumulation item, a preliminary warning of membrane pollution trend is made, and a pollution trend warning model is constructed.

[0029] As a preferred solution of the membrane control method based on multi-parameter intelligent perception described in the present invention, the specific formula of the pollution trend warning model is:

[0030]

[0031] Among them, S poll(t) is the first prediction result, and α and β are sensitivity adjustment factors.

[0032] As a preferred embodiment of the membrane control method based on multi-parameter intelligent perception described in the present invention, a dynamic pollution risk score calculation is performed based on the first prediction result to obtain a comprehensive pollution risk assessment value, including the following steps:

[0033] Based on the first prediction result, a pollution risk dynamic response function is constructed to characterize the immediate intensity and abnormal fluctuation characteristics of pollution trend changes. The specific formula is as follows:

[0034] F1(t)=α1·S poll (t) 2 +α2·Ω(t);

[0035] Among them, F1(t) represents the dynamic response value of the current pollution trend to the system risk, which is used to express the intensity and mutation of the pollution trend. Ω(t) is the predicted disturbance response term, which represents the fluctuation amplitude of the prediction result and comprehensively reflects the pollution change rate and its acceleration trend. α1 and α2 represent the amplification adjustment parameters, which are used to control the assessment weights of the pollution degree and the disturbance response term, respectively. By introducing the parameter coupling function K(t) and the pollution load accumulation term R(t) and based on the dynamic response value F1(t) of the current pollution trend to the system risk, the comprehensive assessment value of the pollution risk is obtained. The specific calculation formula is as follows:

[0036]

[0037] Among them, R risk (t) represents the comprehensive assessment value of the pollution risk of the membrane system at time t. K(t) reflects the nonlinear correlation degree of multi-source variables such as flow rate and water quality in the membrane system in the form of a Gaussian kernel function, which is used to quantify the pollution enhancement effect caused by the synergy of multiple factors. β1 and β2 are coefficients that adjust the influence of the multi-parameter coupling function and the pollution accumulation term on the comprehensive assessment value of the pollution risk.

[0038] As a preferred solution of the membrane control method based on multi-parameter intelligent perception described in the present invention, the membrane pollution trend includes a normal state, a light pollution state, a moderate pollution risk state and a severe pollution state.

[0039] As a preferred embodiment of the membrane control method based on multi-parameter intelligent perception described in the present invention, a basic pollution determination is performed based on a first prediction result and a comprehensive pollution risk assessment value. If the first prediction result is less than or equal to a pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is less than or equal to a pollution risk threshold S1, the current state is determined to be normal, cleaning intervention is not performed, and the membrane separation system is continuously monitored.

[0040] If the first prediction result is less than or equal to the pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is greater than the pollution risk threshold S1, the current state is preliminarily determined to be a lightly polluted state, and the next step is executed;

[0041] When the preliminary judgment result is a light pollution state, the pollution upgrade judgment is made based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate and the membrane operation load duration. If the transmembrane pressure difference change value is less than or equal to the pressure difference threshold Y1, the inlet and outlet water quality change rate is less than or equal to the water quality change rate threshold Z1, and the membrane operation load duration is less than or equal to the load duration threshold T1, the current state is judged to be light pollution, and a level 1 warning message is sent to the operator, suggesting that the membrane system be inspected;

[0042] If the transmembrane pressure difference change value is greater than the pressure difference threshold Y1, the inlet and outlet water quality change rate is greater than the water quality change rate threshold Z1, or the membrane operation load duration is greater than the load duration threshold T1, the current state is determined to be a moderate pollution risk state, a second-level warning message is sent to the operator, and the membrane system parameters are immediately recorded and analyzed;

[0043] If the first prediction result is greater than the pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is less than or equal to the pollution risk threshold S1, the current state is preliminarily determined to be a moderate pollution risk state, and the next step is executed;

[0044] When the preliminary judgment result is a moderate contamination risk state, the pollution upgrade judgment is made based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate and the membrane operation load duration. If the transmembrane pressure difference change value is less than or equal to the pressure difference threshold Y1, the inlet and outlet water quality change rate is less than or equal to the water quality change rate threshold Z1, and the membrane operation load duration is less than or equal to the load duration threshold T1, the current state is judged to be a moderate contamination risk state, a secondary warning message is sent to the operator, and the membrane system parameters are immediately recorded and analyzed;

[0045] If the transmembrane pressure difference change value is greater than the pressure difference threshold Y1, the inlet and outlet water quality change rate is greater than the water quality change rate threshold Z1, or the membrane operation load duration is greater than the load duration threshold T1, the current state is determined to be a serious pollution state, and a third-level warning message is sent to notify the operator to immediately perform membrane system cleaning intervention;

[0046] If the first prediction result is greater than the pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is greater than the pollution risk threshold S1, the current state is determined to be a serious pollution state, and a third-level warning message is sent to notify the operator to immediately perform membrane system cleaning intervention.

[0047] Secondly, in order to further solve the problem of inaccurate cleaning timing judgment in the prior art, which affects the long-term operation stability and cleaning efficiency of the membrane system, the present embodiment provides a membrane control system based on multi-parameter intelligent perception, including:

[0048] Data acquisition module: acquires multi-parameter operating data of the membrane separation system and calculates the transmembrane pressure difference change value, the rate of change of inlet and outlet water quality, and the membrane operation load duration based on the operating data;

[0049] Model construction module: based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate and the membrane operation load duration, construct and analyze the pollution trend warning model based on the improved long short-term memory network to obtain a first prediction result;

[0050] Scoring calculation module: performs dynamic pollution risk scoring calculation based on the first prediction result to obtain a comprehensive pollution risk assessment value;

[0051] Comprehensive judgment module: performs comprehensive judgment on membrane pollution trend according to the first prediction result and the comprehensive assessment value of pollution risk, and performs cleaning intervention.

[0052] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the membrane control method based on multi-parameter intelligent perception as described in the first aspect of the present invention is implemented.

[0053] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the membrane control method based on multi-parameter intelligent perception as described in the first aspect of the present invention is implemented.

[0054] Beneficial effects of the present invention:

[0055] By integrating multi-parameter operating data, the pollution feature vector can be constructed in real time, and the membrane pollution trend can be dynamically modeled and predicted, thereby obtaining the first prediction result reflecting the pollution evolution. Combined with the comprehensive assessment value of pollution risk, the pollution status of the membrane system can be accurately determined. Compared with traditional methods that only rely on fixed thresholds or single-parameter triggering control strategies, the present invention introduces nonlinear matrix scoring, Gaussian disturbance response and pollution cumulative integral expression in the modeling structure, which has stronger trend recognition and time series perception capabilities.

[0056] By further constructing a dynamic response function for pollution risk and a parameter coupling function based on the model output, the present invention realizes the linkage analysis of pollution development trends and system risk levels, can identify potential pollution risk levels in advance, and output cleaning intervention suggestions or alarm strategies in a graded manner, effectively improving the initiative and timeliness of cleaning decisions. The invention also divides the membrane operation process into normal, mild, moderate and severe pollution stages through a state division mechanism, ensuring that the system can adaptively adjust the response strategy under different working conditions, thereby significantly improving the operating stability, prediction accuracy and cleaning management efficiency of the membrane system. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0058] Figure 1 This is a schematic diagram of the overall process of a membrane control method based on multi-parameter intelligent perception proposed by the present invention;

[0059] Figure 2 This is a schematic diagram of a framework for constructing a pollution trend warning model for a membrane control method based on multi-parameter intelligent perception proposed in the present invention;

[0060] Figure 3 This is a logical diagram of comprehensive determination of membrane fouling trend of a membrane control method based on multi-parameter intelligent perception proposed in the present invention. DETAILED DESCRIPTION

[0061] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0062] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0063] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0064] Furthermore, the present invention is described in detail with reference to schematic diagrams. For ease of illustration, when describing the embodiments of the present invention, cross-sectional views illustrating device structures may be partially enlarged and not to scale. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of protection of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0065] Example 1

[0066] Reference Figure 1-Figure 3 , as an embodiment of the present invention, provides a membrane control method based on multi-parameter intelligent perception.

[0067] Existing membrane control methods have the following defects: although they have a certain response capability, they generally lack dynamic modeling and prediction mechanisms for pollution trends, and are unable to effectively identify the continuity of pollution development and potential change directions. Especially in the early stages of pollution, the system is often unable to perceive the gradual nature of pollution signals in a timely manner, resulting in inaccurate judgment of the timing of cleaning; in the stage of intensified pollution, the response may also be delayed due to insufficient prediction capabilities, affecting the long-term operating stability and cleaning efficiency of the membrane system.

[0068] The present application provides a method that can effectively solve the above-mentioned problems. Next, multiple embodiments will be combined to explain in detail how to implement the membrane control method based on multi-parameter intelligent perception.

[0069] Figure 1 The overall flow chart of the membrane control method based on multi-parameter intelligent perception is shown, including:

[0070] S1: Obtain multi-parameter operating data of the membrane separation system, and calculate the transmembrane pressure difference change value, the inlet and outlet water quality change rate, and the membrane operation load duration based on the operating data.

[0071] It should be noted that, in the present invention, the membrane separation system refers to an automated water treatment device with liquid inlet and outlet passages, membrane components, an operation control module and an online monitoring unit, which is mainly used to achieve filtration, interception and water production operations of the target liquid. It is often used in scenarios such as deep sewage treatment, industrial circulating water purification and drinking water preparation. By using it in conjunction with the method of the present invention, it can significantly improve the pollution prediction ability and operation stability of the membrane separation system. It belongs to a very mature existing technology.

[0072] Multi-parameter operation data refers to the operation process parameter data obtained by sensors deployed in the membrane separation system;

[0073] Specifically, multi-parameter operation data includes the following data:

[0074] By laying pressure sensors in the membrane inlet pipe and outlet pipe respectively, the pressure data on both sides of the membrane can be obtained in real time. The instantaneous flow data on the membrane inlet and outlet sides are monitored by flow sensors to assist in judging the membrane flux and water production efficiency, and to analyze the changes in hydraulic load. The turbidity of the inlet and outlet water quality is measured by an online turbidity meter, the conductivity of the inlet and outlet water quality is measured by an online conductivity sensor, the concentration of pollutants in the inlet water quality (including COD, chemical oxygen demand concentration, and TSS, suspended solids concentration) is analyzed through online or laboratory sampling, and the continuous operation time of the membrane under uncleaned conditions is recorded by the control system.

[0075] The transmembrane pressure difference change value refers to the change in the difference between the pressure on the water inlet side and the pressure on the water outlet side of the membrane per unit time;

[0076] The rate of change of water quality of inlet and outlet water refers to the rate of change per unit time of the difference between the key water quality indicators (including turbidity and conductivity) of the membrane inlet and outlet water;

[0077] Membrane operating load duration refers to the length of time the membrane system operates continuously without performing cleaning intervention operations.

[0078] S2: Based on the change in transmembrane pressure difference, the rate of change of inlet and outlet water quality, and the duration of membrane operation load, a pollution trend warning model based on the improved long-short-term memory network is constructed and analyzed to obtain the first prediction result.

[0079] like Figure 2 As shown in Figure 2, the construction of the pollution trend early warning model includes the following steps:

[0080] Construct the pollution characteristic vector X(t) based on multi-parameter operation data, including the transmembrane pressure difference change rate ΔP(t), inlet and outlet water flow rate Q in (t) and Q out (t), water quality index C COD (t) and C TSS (t) and membrane operation duration τ(t), which are specifically expressed as follows:

[0081] X(t)=[ΔP(t),Q in (t), Q out (t), C COD (t), C TSS (t), τ(t)] T ;

[0082] Introducing the 6th-order pollution weight matrix And define the pollution feature matrix scoring item as:

[0083]

[0084] Among them, M(t) is the pollution feature matrix scoring item, is a sigmoid mapping function, which is used to map the pollution intensity to the standardized interval [0,1]. The transmembrane pressure difference disturbance limit average function term is introduced, and the disturbance accumulation expression based on the Gaussian error function is constructed. The specific formula is:

[0085]

[0086] in:

[0087]

[0088] Among them, D(t) is the historical disturbance function term, which is used to represent the impact of historical pressure difference fluctuations on the membrane fouling trend. Ψ(x) is the Gaussian error function, which is used to simulate the cumulative effect of random disturbances. Λ(t) represents the average value of the historical disturbance limit of the transmembrane pressure difference at the tth moment, which is used to reflect the cumulative trend of small disturbances in the evolution of membrane fouling. ΔP(tk) represents the rate of change of the transmembrane pressure difference at the tth moment. n is the selected historical length window. u represents the integral variable, which is used to calculate the probability accumulation function of the disturbance intensity. Considering that the membrane fouling process continues to accumulate with the operating time, the pollution load accumulation term is introduced into the pollution trend warning model. The specific formula is as follows:

[0089]

[0090] Among them, R(t) is the cumulative pollution load term, which represents the total organic pollution intensity experienced by the membrane before time t, and C COD (s) is the influent COD concentration at the sth moment, C TSS (s) is the influent TSS concentration at the sth moment, e -s It is an exponential decay function, which indicates the decreasing trend of pollutant impact over time. Based on the obtained pollution characteristic matrix score item, historical disturbance function item and pollution load accumulation item, a preliminary warning of membrane pollution trend is made, and a pollution trend warning model is constructed.

[0091] The specific formula of the pollution trend early warning model is:

[0092]

[0093] Among them, S poll (t) is the first prediction result, and α and β are sensitivity adjustment factors.

[0094] It should be noted that although both α and β are sensitivity adjustment factors, their objects of action and adjustment targets are different. α is used to adjust the weight of the pollution load accumulation item on the prediction result, and tends to suppress the excessive score caused by the continuous accumulation of pollution. β is used to adjust the influence of the historical disturbance function item on the prediction result, and tends to strengthen the early warning effect of sudden pollution signals. α minimizes the mean square error between the prediction result and the actual pollution state by fitting the historical membrane system operation data, and then optimizes the α value that is most suitable for the pollution accumulation characteristics of the current membrane system. β maximizes the model's early warning accuracy for short-term disturbance signals by fitting the correlation between historical pressure difference disturbance data and pollution mutations, and adopts the gradient descent algorithm for iterative optimization, thereby determining the optimal value of β. In practical applications, the initial parameter value can also be determined in combination with expert experience or empirical regression analysis, and then the parameter configuration can be optimized through online fine-tuning. This embodiment does not make specific limitations on this.

[0095] It should be noted that in order to improve the accuracy and foresight of membrane fouling trend identification, the present invention adopts an improved long short-term memory network to construct a fouling trend warning model, which is reflected in the following two aspects:

[0096] First, a pollution feature vector construction mechanism is introduced before the input layer of the original LSTM. Specifically, the multi-parameter pollution operation data is normalized and constructed into a 6-dimensional pollution feature vector. The vector is multiplied by the learnable 6th-order pollution weight matrix to form a pollution score item, which serves as part of the input of the pollution trend warning model. At the same time, the historical disturbance function item and the pollution load accumulation item are introduced to form the input triplet of the long and short-term memory network. Second, a pollution trend scoring function is added to the traditional output layer to nonlinearly fuse multiple pollution factors and output trend prediction values, thereby improving the sensitivity of the pollution trend warning model to pollution changes and its ability to identify abnormal mutations. The pollution trend warning model constructed based on the improved long and short-term memory network receives the pollution feature combination of the current moment and the historical moment, and learns its evolution trend through a time recursion mechanism. It finally outputs the first prediction result, which is directly used as the first judgment basis of the pollution warning model for subsequent comprehensive judgment of pollution trends.

[0097] S3: Calculating a dynamic pollution risk score based on the first prediction result to obtain a comprehensive pollution risk assessment value, including the following steps:

[0098] Based on the first prediction result, a pollution risk dynamic response function is constructed to characterize the immediate intensity and abnormal fluctuation characteristics of pollution trend changes. The specific formula is as follows:

[0099] F1(t)=α1·S poll (t) 2 +α2·Ω(t);

[0100] Among them, F1(t) represents the dynamic response value of the current pollution trend to the system risk, which is used to express the intensity and mutation of the pollution trend. Ω(t) is the predicted disturbance response term, which represents the fluctuation amplitude of the prediction result and comprehensively reflects the pollution change rate and its acceleration trend. α1 and α2 represent the amplification adjustment parameters, which are used to control the assessment weights of the pollution degree and the disturbance response term, respectively. By introducing the parameter coupling function K(t) and the pollution load accumulation term R(t) and based on the dynamic response value F1(t) of the current pollution trend to the system risk, the comprehensive assessment value of the pollution risk is obtained. The specific calculation formula is as follows:

[0101]

[0102] Among them, R risk (t) represents the comprehensive assessment value of the pollution risk of the membrane system at time t. K(t) reflects the nonlinear correlation degree of multi-source variables such as flow rate and water quality in the membrane system in the form of a Gaussian kernel function, which is used to quantify the pollution enhancement effect caused by the synergy of multiple factors. β1 and β2 are coefficients that adjust the influence of the multi-parameter coupling function and the pollution accumulation term on the comprehensive assessment value of the pollution risk.

[0103] In an embodiment of the present application, the amplified adjustment parameters in the pollution risk dynamic response function α1 and α2 can be obtained by constructing a training set through historical operation data and performing fitting optimization by minimizing the loss function (such as mean square error). β1 and β2 can be supervised and trained through historical pollution level label data, and an optimization algorithm can be used to search for the optimal value to improve the evaluation accuracy; they can also be set based on empirical rules or on-site parameter adjustment.

[0104] S4: Perform a comprehensive judgment on the membrane fouling trend based on the first prediction result and the comprehensive assessment value of the fouling risk, and perform cleaning intervention.

[0105] Preferably, the membrane fouling trend includes a normal state, a lightly fouled state, a moderately fouled risk state, and a severely fouled state.

[0106] Specifically, such as Figure 3 Figure 2 shows a logic diagram of comprehensive membrane fouling trend determination for a membrane control method based on multi-parameter intelligent perception proposed in the present invention. Basic fouling determination is performed based on the first prediction result and the comprehensive assessment value of the fouling risk. If the first prediction result is less than or equal to the fouling trend warning threshold R1, and the comprehensive assessment value of the fouling risk is less than or equal to the fouling risk threshold S1, the current state is determined to be normal, and no cleaning intervention is performed, but the membrane separation system is continuously monitored.

[0107] If the first prediction result is less than or equal to the pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is greater than the pollution risk threshold S1, the current state is preliminarily determined to be a lightly polluted state, and the next step is executed;

[0108] When the preliminary judgment result is a light pollution state, the pollution upgrade judgment is made based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate and the membrane operation load duration. If the transmembrane pressure difference change value is less than or equal to the pressure difference threshold Y1, the inlet and outlet water quality change rate is less than or equal to the water quality change rate threshold Z1, and the membrane operation load duration is less than or equal to the load duration threshold T1, the current state is judged to be light pollution, and a level 1 warning message is sent to the operator, suggesting that the membrane system be inspected;

[0109] If the transmembrane pressure difference change value is greater than the pressure difference threshold Y1, the inlet and outlet water quality change rate is greater than the water quality change rate threshold Z1, or the membrane operation load duration is greater than the load duration threshold T1, the current state is determined to be a moderate pollution risk state, a second-level warning message is sent to the operator, and the membrane system parameters are immediately recorded and analyzed;

[0110] If the first prediction result is greater than the pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is less than or equal to the pollution risk threshold S1, the current state is preliminarily determined to be a moderate pollution risk state, and the next step is executed;

[0111] When the preliminary judgment result is a moderate contamination risk state, the pollution upgrade judgment is made based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate and the membrane operation load duration. If the transmembrane pressure difference change value is less than or equal to the pressure difference threshold Y1, the inlet and outlet water quality change rate is less than or equal to the water quality change rate threshold Z1, and the membrane operation load duration is less than or equal to the load duration threshold T1, the current state is judged to be a moderate contamination risk state, a secondary warning message is sent to the operator, and the membrane system parameters are immediately recorded and analyzed;

[0112] If the transmembrane pressure difference change value is greater than the pressure difference threshold Y1, the inlet and outlet water quality change rate is greater than the water quality change rate threshold Z1, or the membrane operation load duration is greater than the load duration threshold T1, the current state is determined to be a serious pollution state, and a third-level warning message is sent to notify the operator to immediately perform membrane system cleaning intervention;

[0113] If the first prediction result is greater than the pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is greater than the pollution risk threshold S1, the current state is determined to be a serious pollution state, and a third-level warning message is sent to notify the operator to immediately perform membrane system cleaning intervention.

[0114] For example, assume that at a municipal sewage treatment plant, a membrane separation system is in operation. After more than 120 hours of continuous and stable operation, the data acquisition module in the membrane separation system continuously detects a slight increase in the transmembrane pressure difference and a gradually increasing difference in the turbidity of the inlet and outlet water. Although the first prediction result output by the pollution trend warning model is still in the low-risk range at this time, the comprehensive pollution risk assessment value is close to the set pollution risk threshold. Based on the first prediction result and the pollution risk assessment value, the system determines that it is currently in a mild pollution warning state and initiates a further judgment mechanism. After analysis, the current transmembrane pressure difference change value has exceeded the pressure difference threshold Y1, and the membrane operating load duration exceeds the set threshold T1. The current state is comprehensively judged to be a moderate pollution risk state, and a secondary warning message is immediately sent to the operation and maintenance personnel. After receiving the alarm, the operation and maintenance personnel record the parameters of the membrane system and inspect the water samples on site, confirming the presence of an initial organic pollution trend on the membrane surface. They complete a low-intensity chemical cleaning of the membrane components before the pollution intensifies, effectively preventing the pollution from further developing into a serious state and ensuring that the treated water quality remains stable and meets the standards.

[0115] In the embodiments of the present application, for the various thresholds in the comprehensive judgment of membrane pollution trend, the pollution trend warning threshold is established by analyzing the distribution of prediction results output by the improved LSTM model under normal operation and mild pollution state of the membrane system, combining a large amount of historical label data for statistical modeling, and selecting the optimal balance point of sensitivity and specificity based on the ROC curve analysis method; the pollution risk threshold is established by modeling and analyzing the distribution law of the pollution risk score results under various typical working conditions, combining the cleaning response requirements corresponding to different risk levels, and adopting the quantile method or error minimization strategy to determine whether the risk reaches the intervention condition that triggers the cleaning recommendation; the pressure difference threshold is established based on the transmembrane pressure difference fluctuation characteristics of the membrane component under different flux loads, and is determined by long-term monitoring of the change range of the transmembrane pressure difference under different pollution states. The range is modified in combination with the operating condition limit and anti-pollution characteristics provided by the membrane manufacturer; the water quality change rate threshold is set based on the response speed change trend of the difference between the inlet and outlet turbidity and conductivity during the development of membrane pollution, combined with the sensor response characteristics, detection accuracy and actual water quality disturbance amplitude statistics; the load duration threshold is obtained by analyzing the pollution accumulation characteristics corresponding to different operating time in a large number of membrane operation cycle records, combining the cleaning cycle experience and the pollution turning point characteristics, and evaluating the failure mode and cleaning delay risk; all of the above thresholds can also be initially set in combination with the small-scale trial operation data of the membrane system in a specific application scenario, and optimized and adjusted through dynamic feedback during online operation; an adaptive threshold learning strategy based on machine learning can also be introduced to achieve model-threshold co-evolution during continuous training. This embodiment does not make specific restrictions on this.

[0116] It should be noted that existing membrane control methods mostly rely on static thresholds or single parameters to judge the pollution status, which is difficult to fully cope with the pollution evolution process under complex operating conditions. This embodiment designs a joint judgment mechanism based on the first prediction result and the comprehensive pollution risk assessment value to establish a multi-level intelligent pollution judgment system including normal state, mild pollution warning state, moderate pollution risk state and severe pollution state. It can combine the current model output with the parameter evolution characteristics during the pollution development process for comprehensive analysis. This multi-level trend judgment mechanism breaks through the limitations of traditional binary judgment, realizes the progressive and dynamic identification of the membrane pollution process, and automatically matches the corresponding warning level and intervention response strategy according to the severity and evolution speed of the pollution trend, significantly improving the accuracy and adaptability of membrane pollution control, effectively reducing the risk of unplanned shutdown during system operation, and avoiding the waste of resources and loss of membrane life caused by excessive cleaning, thereby achieving dual optimization of the operating efficiency and maintenance cost of the membrane separation system.

[0117] In summary, the present invention provides a membrane control method based on multi-parameter intelligent perception. By fusing multi-parameter operation data, a pollution feature vector is constructed in real time, and the membrane pollution trend is dynamically modeled and predicted, thereby obtaining a first prediction result reflecting the evolution of pollution, and combined with the comprehensive assessment value of pollution risk, an accurate judgment of the pollution state of the membrane system is achieved. Compared with the traditional method that only relies on fixed thresholds or single parameter triggering control strategies, the present invention introduces nonlinear matrix scoring, Gaussian perturbation response and pollution cumulative integral expression in the modeling structure, and has stronger trend recognition ability and time series perception ability; by further constructing a pollution risk dynamic response function and a parameter coupling function based on the model output, the present invention realizes the linkage analysis of pollution development trends and system risk levels, can identify potential pollution risk levels in advance, and output cleaning intervention suggestions or alarm strategies in a graded manner, effectively improving the initiative and timeliness of cleaning decisions. The invention also divides the membrane operation process into normal, mild, moderate and severe pollution stages through a state division mechanism, ensuring that the system can adaptively adjust the response strategy under different working conditions, thereby significantly improving the operating stability, prediction accuracy and cleaning management efficiency of the membrane system.

[0118] Example 2, an embodiment of the present invention, provides a membrane control system based on multi-parameter intelligent perception, comprising:

[0119] Data acquisition module: acquires multi-parameter operating data of the membrane separation system and calculates the transmembrane pressure difference change value, the rate of change of inlet and outlet water quality, and the duration of membrane operation load based on the operating data;

[0120] Model building module: Based on the change in transmembrane pressure difference, the rate of change of inlet and outlet water quality, and the duration of membrane operation load, a pollution trend warning model based on an improved long-short-term memory network is constructed and analyzed to obtain the first prediction result;

[0121] Scoring calculation module: performs dynamic pollution risk scoring calculation based on the first prediction result to obtain a comprehensive assessment value of pollution risk;

[0122] Comprehensive judgment module: Performs comprehensive judgment on membrane fouling trend based on the first prediction result and the comprehensive assessment value of fouling risk, and performs cleaning intervention.

[0123] Example 3 is an embodiment of the present invention, which is different from the previous embodiment in that:

[0124] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0125] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0126] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0127] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0128] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A membrane control method based on multi-parameter intelligent perception, characterized in that: include: Acquiring multi-parameter operating data of the membrane separation system, and calculating the transmembrane pressure difference change value, the rate of change of inlet and outlet water quality, and the membrane operation load duration based on the operating data; Based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate, and the membrane operation load duration, a pollution trend early warning model based on an improved long short-term memory network is constructed and analyzed to obtain a first prediction result; Performing a dynamic pollution risk score calculation based on the first prediction result to obtain a comprehensive pollution risk assessment value; A comprehensive judgment of the membrane fouling trend is made based on the first prediction result and the comprehensive assessment value of the fouling risk, and a cleaning intervention is performed.

2. The membrane control method based on multi-parameter intelligent perception according to claim 1, characterized in that: The multi-parameter operation data refers to the operation process parameter data obtained by sensors arranged in the membrane separation system; The transmembrane pressure difference change value refers to the change amplitude of the difference between the membrane water inlet side pressure and the water outlet side pressure per unit time; The rate of change of water quality of the inlet and outlet water refers to the rate of change per unit time of the difference between the key water quality indicators of the membrane inlet and outlet water; The membrane operation load duration refers to the length of time the membrane system continuously operates without performing a cleaning intervention operation.

3. The membrane control method based on multi-parameter intelligent perception according to claim 2, characterized in that: Based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate and the membrane operation load duration, a pollution trend early warning model based on an improved long short-term memory network is constructed, comprising the following steps: Based on multi-parameter operation data, a pollution feature vector is constructed, a 6th-order pollution weight matrix is ​​introduced, and pollution feature matrix scoring items are defined; The perturbation limit average function term of transmembrane pressure difference is introduced, and the perturbation accumulation expression based on Gaussian error function is constructed; The pollution load accumulation term is introduced to simulate the continuous accumulation of membrane fouling over the operating time; Based on the obtained pollution characteristic matrix score item, historical disturbance function item and pollution load accumulation item, a preliminary warning of membrane pollution trend is given, and a pollution trend warning model is constructed.

4. The membrane control method based on multi-parameter intelligent perception according to claim 3, characterized in that: The specific formula of the pollution trend early warning model is as follows: Among them, S poll (t) is the first prediction result, α and β are sensitivity adjustment factors, M(t) is the pollution characteristic matrix score item, D(t) is the historical disturbance function item, and R(t) is the pollution load accumulation item.

5. The membrane control method based on multi-parameter intelligent perception according to claim 4, characterized in that: Performing a dynamic pollution risk score calculation based on the first prediction result to obtain a comprehensive pollution risk assessment value includes the following steps: Based on the first prediction results, a dynamic response function for pollution risk is constructed to characterize the immediate intensity and abnormal fluctuation characteristics of pollution trend changes; By introducing parameter coupling function and pollution load accumulation term and based on the dynamic response value of current pollution trend to system risk, a comprehensive assessment value of pollution risk is obtained.

6. The membrane control method based on multi-parameter intelligent perception according to claim 1, characterized in that: The membrane fouling trends include a normal state, a light fouling state, a moderate fouling risk state, and a severe fouling state.

7. The membrane control method based on multi-parameter intelligent perception according to claim 6, characterized in that: A basic pollution determination is performed based on the first prediction result and the comprehensive pollution risk assessment value. If the first prediction result is less than or equal to the pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is less than or equal to the pollution risk threshold S1, the current state is determined to be normal, and no cleaning intervention is performed, and the membrane separation system is continuously monitored. If the first prediction result is less than or equal to the pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is greater than the pollution risk threshold S1, the current state is preliminarily determined to be a lightly polluted state, and the next step is executed; When the preliminary judgment result is a light pollution state, the pollution upgrade judgment is made based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate and the membrane operation load duration. If the transmembrane pressure difference change value is less than or equal to the pressure difference threshold Y1, the inlet and outlet water quality change rate is less than or equal to the water quality change rate threshold Z1, and the membrane operation load duration is less than or equal to the load duration threshold T1, the current state is judged to be light pollution, and a level 1 warning message is sent to the operator, suggesting that the membrane system be inspected; If the transmembrane pressure difference change value is greater than the pressure difference threshold Y1, the inlet and outlet water quality change rate is greater than the water quality change rate threshold Z1, or the membrane operation load duration is greater than the load duration threshold T1, the current state is determined to be a moderate pollution risk state, a second-level warning message is sent to the operator, and the membrane system parameters are immediately recorded and analyzed; If the first prediction result is greater than the pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is less than or equal to the pollution risk threshold S1, the current state is preliminarily determined to be a moderate pollution risk state, and the next step is executed; When the preliminary judgment result is a moderate contamination risk state, the pollution upgrade judgment is made based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate and the membrane operation load duration. If the transmembrane pressure difference change value is less than or equal to the pressure difference threshold Y1, the inlet and outlet water quality change rate is less than or equal to the water quality change rate threshold Z1, and the membrane operation load duration is less than or equal to the load duration threshold T1, the current state is judged to be a moderate contamination risk state, a secondary warning message is sent to the operator, and the membrane system parameters are immediately recorded and analyzed; If the transmembrane pressure difference change value is greater than the pressure difference threshold Y1, the inlet and outlet water quality change rate is greater than the water quality change rate threshold Z1, or the membrane operation load duration is greater than the load duration threshold T1, the current state is determined to be a serious pollution state, and a third-level warning message is sent to notify the operator to immediately perform membrane system cleaning intervention; If the first prediction result is greater than the pollution trend warning threshold R1, and the comprehensive pollution risk assessment value is greater than the pollution risk threshold S1, the current state is determined to be a serious pollution state, and a third-level warning message is sent to notify the operator to immediately perform membrane system cleaning intervention.

8. A membrane control system based on multi-parameter intelligent perception, based on the membrane control method based on multi-parameter intelligent perception according to any one of claims 1 to 7, characterized in that: include: Data acquisition module: acquires multi-parameter operating data of the membrane separation system and calculates the transmembrane pressure difference change value, the rate of change of inlet and outlet water quality, and the membrane operation load duration based on the operating data; Model construction module: based on the transmembrane pressure difference change value, the inlet and outlet water quality change rate and the membrane operation load duration, construct and analyze the pollution trend warning model based on the improved long short-term memory network to obtain a first prediction result; Scoring calculation module: performs dynamic pollution risk scoring calculation based on the first prediction result to obtain a comprehensive pollution risk assessment value; Comprehensive judgment module: performs comprehensive judgment on membrane pollution trend according to the first prediction result and the comprehensive assessment value of pollution risk, and performs cleaning intervention.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the membrane control method based on multi-parameter intelligent perception according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the membrane control method based on multi-parameter intelligent perception according to any one of claims 1 to 7 are implemented.

Images