Membrane system early warning regulation and control method based on Kalman filtering and medium

Abstract

The invention discloses a membrane system early warning regulation and control method based on Kalman filtering and a medium. The method comprises the following steps: establishing a membrane pollution physical model based on an ln function; substituting the membrane pollution physical model into the Kalman filtering model to obtain a predicted transmembrane pressure difference; a plurality of early warning limit values are set, and if the transmembrane pressure difference is larger than the early warning limit values, early warning is carried out for the corresponding membrane process stage; the membrane cleaning sequence is reordered, so that the dynamic regulation and control of the cleaning sequence of the membrane bioreactor-the plurality of membrane modules are realized. According to the method, the membrane pollution physical empirical model is coupled with the Kalman filtering, so that the multi-step prediction of the membrane pollution process variable is realized. The problem that membrane pollution is difficult to predict is solved, harm caused by membrane pollution is reduced, and safe, stable and efficient operation of the MBR system is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of membrane system early warning and control, and more specifically, to a membrane system early warning and control method and medium based on Kalman filtering. Background Art

[0002] Membrane bioreactor is a wastewater treatment technology that combines activated sludge method with membrane separation technology. It has the advantages of good effluent quality, high load and small footprint.

[0003] However, membrane fouling can cause an increase in transmembrane pressure difference and a decrease in membrane flux, leading to a decrease in membrane performance, a reduction in membrane service life, and an increase in operating energy consumption and usage costs.

[0004] Therefore, in order to reduce the occurrence of membrane fouling, it is necessary to develop a membrane system early warning control method and medium based on Kalman filtering.

[0005] The information disclosed in the background technology section of the present invention is only intended to deepen the understanding of the general background technology of the present invention, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0006] The present invention proposes a membrane system early warning control method and medium based on Kalman filtering, which realizes multi-step prediction of membrane pollution process variables by coupling the physical empirical model of membrane pollution with Kalman filtering. It solves the problem of membrane pollution being difficult to predict, reduces the harm caused by membrane pollution, and ensures the safe, stable and efficient operation of the MBR system.

[0007] In a first aspect, an embodiment of the present disclosure provides a membrane system early warning and control method based on Kalman filtering, comprising:

[0008] Establish a physical model of membrane fouling based on ln function;

[0009] Substituting the membrane fouling physical model into the Kalman filter model to obtain a predicted transmembrane pressure difference;

[0010] Setting multiple warning limits, if the transmembrane pressure difference is greater than the warning limit, a warning is issued for the corresponding membrane process stage;

[0011] The membrane cleaning sequence is reordered to achieve dynamic regulation of the cleaning sequence of membrane bioreactor-multiple membrane modules.

[0012] Preferably, the membrane fouling physical model based on the ln function is:

[0013] ln(T)=ln(T 0 )+kt

[0014] Where T is the transmembrane pressure difference of the membrane bioreactor during operation, T 0 is the transmembrane pressure difference of the membrane bioreactor during operation at the initial moment, k is the membrane fouling coefficient, and t is the time.

[0015] Preferably, the Kalman filter model is:

[0016] a i =C i a i-1 +A i v i +R i

[0017] s i =C i a i +m i

[0018] Among them, a i represents the predicted value, C i represents the state transfer matrix, A i represents the control matrix, v i represents the control function, v i =Δt; R i represents the process noise, s i is the system observation value; m i represents the observation noise.

[0019] Preferably, substituting the membrane fouling physical model into the Kalman filter model to obtain the predicted transmembrane pressure difference comprises:

[0020] Performing single-step time prediction for the Kalman filter model;

[0021] Update filtering, using the difference between the observed value and the predicted value at the current moment to correct the predicted value at the current moment;

[0022] Perform multi-step time prediction for filtering.

[0023] Preferably, the single-step time prediction is:

[0024]

[0025] B i|i-1 =C i B i-1|i-1 C T i +D i

[0026] in, is the prior value; B i|i-1is the prior covariance of the deviation between the system observations and the model predictions.

[0027] Preferably, correcting the predicted value at the current moment by using the difference between the observed value at the current moment and the predicted value comprises:

[0028] E i =B i|i-1 C T i (C i B i|i-1 C T i +P i ) -i

[0029]

[0030] B i|i =(IE i C i )B i|i-1

[0031] Among them, E i is the Kalman gain; B i This is the posterior covariance; is the filter estimate, I is the state transfer matrix,

[0032] Preferably, performing multi-step time prediction for filtering includes:

[0033] Search for time periods with the same historical time interval from the current moment, calculate the percentage average relative error with the current operation stage, and find the period that satisfies the historical date time period [t-Δt, t] and the current operation stage with the smallest percentage average relative error as the positioning date of the pseudo observation value;

[0034] The pseudo observation value of the future prediction time period Δt is equal to the end time t of the historical date of positioning, and the data of t+Δt is found as the pseudo measurement value data of the prediction data;

[0035] In the prediction process, the pseudo-measurement s DM The value replaces the system observation value s i Perform filter estimation calculations.

[0036] Preferably, reordering the membrane cleaning sequence includes:

[0037] The growth rate is calculated based on the predicted transmembrane pressure difference and the membrane cleaning sequence is determined.

[0038] Preferably, the growth rate is:

[0039]

[0040] Among them, P rate is the growth rate, T t1 is the transmembrane pressure difference corresponding to time t1, T t2 is the transmembrane pressure difference corresponding to time t2.

[0041] In a second aspect, an embodiment of the present disclosure further provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the membrane system early warning and control method based on Kalman filtering is implemented.

[0042] Its beneficial effects are:

[0043] The present invention couples the membrane pollution model in the Kalman filter formula, combines the pseudo-observation value judgment method, realizes the multi-step prediction process of the Kalman filter, and improves the accuracy and performance of membrane pollution prediction. In addition, it also provides data support for the order of membrane cleaning, realizes the regulation of dynamic membrane cleaning sequence, and solves the problem of serious over-cleaning of membrane pollution caused by fixed-cycle fixed-sequence cleaning and insufficient series cleaning of serious membrane pollution. The present invention greatly reduces the cost of membrane pollution treatment, improves the use efficiency of MBR, and ensures the safe, stable and efficient operation of sewage treatment plants.

[0044] The methods and apparatus of the present invention have other features and advantages that will be apparent from, or will be described in detail in, the accompanying drawings and subsequent detailed descriptions incorporated herein, which together serve to explain the specific principles of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.

[0046] Figure 1 A flow chart showing the steps of a membrane system early warning and control method based on Kalman filtering according to an embodiment of the present invention.

[0047] Figure 2 A schematic diagram of 48-hour membrane fouling prediction according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0048] The preferred embodiments of the present invention will be described in more detail below. Although the preferred embodiments of the present invention are described below, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0049] Figure 1 A flow chart showing the steps of a membrane system early warning and control method based on Kalman filtering according to an embodiment of the present invention.

[0050] like Figure 1 As shown, the present invention provides a membrane system early warning control method based on Kalman filtering, comprising:

[0051] Step 101, establishing a membrane fouling physical model based on the ln function;

[0052] Step 102, substituting the membrane fouling physical model into the Kalman filter model to obtain a predicted transmembrane pressure difference;

[0053] Step 103, setting a plurality of warning limits, if the transmembrane pressure difference is greater than the warning limit, a warning is issued for the corresponding membrane process stage;

[0054] Step 104, re-arrange the membrane cleaning sequence to achieve dynamic regulation of the cleaning sequence of the membrane bioreactor-multiple membrane modules.

[0055] In one example, the physical model of membrane fouling based on the ln function is:

[0056] ln(T)=ln(T 0 )+kt

[0057] Where T is the transmembrane pressure difference of the membrane bioreactor during operation, T 0 is the transmembrane pressure difference of the membrane bioreactor during operation at the initial moment, k is the membrane fouling coefficient, and t is the time.

[0058] In one example, the Kalman filter model is:

[0059] a i =C i a i-1 +A i v i +R i

[0060] s i =C i a i +m i

[0061] Among them, a i represents the predicted value, C i represents the state transfer matrix, A i represents the control matrix, v i represents the control function, v i =Δt; R irepresents the process noise, s i is the system observation value; m i represents the observation noise.

[0062] In one example, substituting the membrane fouling physical model into the Kalman filter model to obtain the predicted transmembrane pressure difference includes:

[0063] Single-step time prediction for the Kalman filter model;

[0064] Update filtering, using the difference between the observed value and the predicted value at the current moment to correct the predicted value at the current moment;

[0065] Perform multi-step time prediction for filtering.

[0066] In one example, a one-step time forecast is:

[0067]

[0068] B i|i-1 =C i B i-1|i-1 C T i +D i

[0069] in, is the prior value; B i|i-1 is the prior covariance of the deviation between the system observations and the model predictions.

[0070] In one example, using the difference between the observed value and the predicted value at the current moment to correct the predicted value at the current moment includes:

[0071] E i =B i|i-1 C Y i (C i B i|i-1 C T i +P i ) -1

[0072]

[0073] B i|i =(IE i C i )B i|i-1

[0074] Among them, E i is the Kalman gain; B i This is the posterior covariance; is the filter estimate, I is the state transfer matrix,

[0075] In one example, performing multi-step time prediction for filtering includes:

[0076] Search for time periods with the same historical time interval from the current moment, calculate the percentage average relative error with the current operation stage, and find the period that satisfies the historical date time period [t-Δt, t] and the current operation stage with the smallest percentage average relative error as the positioning date of the pseudo observation value;

[0077] The pseudo observation value of the future prediction time period Δt is equal to the end time t of the historical date of positioning, and the data of t+Δt is found as the pseudo measurement value data of the prediction data;

[0078] In the prediction process, the pseudo-measurement s DM The value replaces the system observation value s i Perform filter estimation calculations.

[0079] In one example, reordering the membrane cleaning sequence includes:

[0080] The growth rate is calculated based on the predicted transmembrane pressure difference and the membrane cleaning sequence is determined.

[0081] In one example, the growth rate is:

[0082]

[0083] Among them, P rate is the growth rate, T t1 is the transmembrane pressure difference corresponding to time t1, T t2 is the transmembrane pressure difference corresponding to time t2.

[0084] Specifically, the physical model of membrane fouling is established as follows:

[0085] T=T 0 exp (k×t) (1)

[0086] Substitute the ln function into both ends of the above formula to obtain the membrane fouling physical model based on the ln function:

[0087] ln (T) = ln (T 0 )+kt (2)

[0088] Where T is the transmembrane pressure difference of the membrane bioreactor during operation, in kPa; T 0 is the transmembrane pressure difference of the membrane bioreactor during operation at the initial moment, in kPa; k is the membrane fouling coefficient, in 1 / day, which represents the characteristics of the biological fouling process; t is the time, in days.

[0089] The transformed physical empirical model of membrane fouling can be used in the Kalman filter formulation.

[0090] Substituting the membrane pollution physical model based on the ln function into the Kalman filter model, we can obtain:

[0091] a i =C i a i-1 +A i v i +R i (3)

[0092] Among them, a i Represents the predicted value, and its formula is: C i Represents the state transfer matrix, and its formula is: A i represents the control matrix, and its formula is: v i represents the control function, and its formula is: i =Δt; R i represents process noise, which has a mean of zero and a covariance of D i , set it to a fixed value.

[0093] s i =C i a i +m i (4)

[0094] Among them, s i is the system observation value; m i represents the observation noise, and R i Same, mean is zero, covariance is P i , set it to a fixed value.

[0095] Substituting the membrane fouling physical model into the Kalman filter model, the predicted transmembrane pressure difference is obtained, including:

[0096] Single-step time prediction for the Kalman filter model:

[0097]

[0098] B i|i-1 =C i B i-1|i-1 C T i +D i (6)

[0099] in, is the prior value; B i|i-1 is the prior covariance of the deviation between the system observations and the model predictions.

[0100] Update the filter and use the difference between the observed value and the predicted value at the current moment to correct the predicted value at the current moment:

[0101] E i =B i|i-1 C T i ( C i B i|i-1 C T i +P i ) -1 (7)

[0102]

[0103] B i|i =(IE i C i )B i|i-1 (9)

[0104] Among them, E i That is the Kalman gain; B i This is the posterior covariance; is the filter estimate, I is the state transfer matrix,

[0105] Perform multi-step time prediction for filtering. The observation values ​​of the future time period prediction process are pseudo-observations, and the calculation method is as follows: search for time periods with the same historical time interval from the current moment, calculate the percentage average relative error with the current operation stage, and find the time period that satisfies the historical date time period [t-Δt, t] and the current operation stage The minimum percentage average relative error is used as the positioning date of the pseudo-observation value. The pseudo-observation value of the future prediction time period (Δt) is equal to the end moment t of the historical date of positioning, and the data of t+Δt is found as the pseudo-measurement value data of the prediction data. Pseudo-measurement s in the prediction process DM The value replaces the system observation value s i Substitute it into formula (8) as the system observation value for filtering estimation.

[0106] Membrane pollution warning: Set the transmembrane pressure difference limit as the warning limit, T Max1 , T Max2 ……T Maxn , the trend of future transmembrane pressure difference is calculated by Kalman filter prediction formula. If the predicted transmembrane pressure difference T p >(T Max1 , T Max2 ...or T Maxn If one of them is used, the corresponding membrane process stage will be warned, which can provide early warning for the use of MBR, which is conducive to the process personnel to take effective measures in advance.

[0107] The purpose of setting multiple warning limits is to take different targeted measures according to different warning levels, take effective measures in advance before irreversible pollution occurs, avoid serious membrane blockage problems, effectively improve cleaning recovery rate and water production, reduce operating costs, reduce the effect of reagents, and increase the life of membrane products. The larger the transmembrane pressure difference, the more serious the membrane pollution, which is a direct indicative indicator. The warning level can be set to a first-level warning. If the first transmembrane pressure difference limit is exceeded, it means that the membrane unit has encountered pollution and the operating parameters need to be adjusted according to the operating status; a second-level warning, if the second transmembrane pressure difference limit is exceeded, it means that the membrane unit has encountered serious pollution and chemical cleaning is required in the near future; a third-level warning, if the third transmembrane pressure difference limit is exceeded, it means that the membrane unit has encountered extreme pollution and chemical cleaning is required as soon as possible.

[0108] Membrane bioreactor - Dynamic regulation of the cleaning sequence of multiple membrane modules to determine the arrangement of membrane cleaning sequence. By calculating the above model, the transmembrane pressure difference growth rate of each series of membrane systems in the future period is calculated as the basis for the maintenance cleaning and restorative cleaning arrangement sequence of large water plants and MBR systems. The formula is as follows:

[0109]

[0110] P=Q / (A*t) (11)

[0111] Among them, Q is the water production flow rate, A is the area of ​​the membrane, and t is the membrane water production time.

[0112] The method can effectively improve the utilization efficiency of MBR and solve the problems of severe membrane pollution caused by fixed-cycle fixed-sequence cleaning and insufficient series cleaning of membrane pollution.

[0113] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the Kalman filter-based membrane system early warning and control method is implemented.

[0114] To facilitate understanding of the solutions and effects of the embodiments of the present invention, two specific application examples are given below. Those skilled in the art should understand that the examples are only for facilitating understanding of the present invention, and any specific details thereof are not intended to limit the present invention in any way.

[0115] Example 1

[0116] Figure 2 A schematic diagram of 48-hour membrane fouling prediction according to an embodiment of the present invention is shown.

[0117] This method is used to predict the operation process of the MBR membrane. The model prediction indicators are shown in Table 1. The 48-hour prediction results are shown in Table 1. Figure 2 As shown in the figure, the average relative percentage error (MAPE) is 2.22%, the equality coefficient (EC) is 98.59%, the mean absolute error (MAE) is 0.14, and the mean square error (MSE) is 0.04.

[0118] Table 1 Model prediction indicators

[0119]

[0120] Example 2

[0121] An embodiment of the present disclosure provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the membrane system early warning and control method based on Kalman filtering is implemented.

[0122] According to the computer-readable storage medium of the embodiment of the present disclosure, non-transitory computer-readable instructions are stored thereon. When the non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the above-mentioned methods of each embodiment of the present disclosure are executed.

[0123] The above-mentioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or mobile hard disk), media with built-in rewritable non-volatile memory (e.g., memory card) and media with built-in ROM (e.g., ROM box).

[0124] Those skilled in the art should understand that the purpose of the above description of the embodiments of the present invention is only to exemplarily illustrate the beneficial effects of the embodiments of the present invention, and is not intended to limit the embodiments of the present invention to any given examples.

[0125] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A membrane system early warning and control method based on Kalman filtering, characterized in that: include: Establish a physical model of membrane fouling based on ln function; Substituting the membrane fouling physical model into the Kalman filter model to obtain a predicted transmembrane pressure difference; Setting multiple warning limits, if the transmembrane pressure difference is greater than the warning limit, a warning is issued for the corresponding membrane process stage; The membrane cleaning sequence is reordered to achieve dynamic regulation of the cleaning sequence of membrane bioreactor-multiple membrane modules.

2. The membrane system early warning and control method based on Kalman filtering according to claim 1, wherein: The physical model of membrane fouling based on the ln function is: ln(T)=ln(T0)+kt Wherein, T is the transmembrane pressure difference of the membrane bioreactor during operation, T0 is the transmembrane pressure difference of the membrane bioreactor during operation at the initial moment, k is the membrane fouling coefficient, and t is time.

3. The membrane system early warning and control method based on Kalman filtering according to claim 2, wherein: The Kalman filter model is: a i =C i a i-1 +A i v i +R i s i =C i a i +m i Among them, a i represents the predicted value, C i represents the state transfer matrix, A i represents the control matrix, v i represents the control function, v i =Δt; R i represents the process noise, s i is the system observation value; m i represents the observation noise.

4. The membrane system early warning and control method based on Kalman filtering according to claim 1, wherein: Substituting the membrane fouling physical model into the Kalman filter model, the predicted transmembrane pressure difference is obtained, including: Performing single-step time prediction for the Kalman filter model; Update filtering, using the difference between the observed value and the predicted value at the current moment to correct the predicted value at the current moment; Perform multi-step time prediction for filtering.

5. The membrane system early warning and control method based on Kalman filtering according to claim 4, wherein: The single-step time prediction is: B i|i-1 =C i B i-1|i-1 C T i +D i in, is the prior value; B i|i-1 is the prior covariance of the deviation between the system observations and the model predictions.

6. The membrane system early warning and control method based on Kalman filtering according to claim 4, wherein: The difference between the observed value and the predicted value at the current moment is used to correct the predicted value at the current moment, including: E i =B i|i-1 C T i (C i B i|i-1 C T i +P i ) -i B i|i =(I-E i C i )B i|i-1 Among them, E i is the Kalman gain; B i This is the posterior covariance; is the filter estimate, I is the state transfer matrix, 7. The membrane system early warning and control method based on Kalman filtering according to claim 6, wherein: Multi-step time prediction for filtering includes: Search for time periods with the same historical time interval from the current moment, calculate the percentage average relative error with the current operation stage, and find the period that satisfies the historical date time period [t-Δt, t] and the current operation stage with the smallest percentage average relative error as the positioning date of the pseudo observation value; The pseudo observation value of the future prediction time period Δt is equal to the end time t of the historical date of positioning, and the data of t+Δt is found as the pseudo measurement value data of the prediction data; In the prediction process, the pseudo-measurement s DM The value replaces the system observation value s i Perform filter estimation calculations.

8. The membrane system early warning and control method based on Kalman filtering according to claim 1, wherein: Re-ordering of membrane cleaning sequence includes: The growth rate is calculated based on the predicted transmembrane pressure difference and the membrane cleaning sequence is determined.

9. The membrane system early warning and control method based on Kalman filtering according to claim 8, wherein: The growth rate is: Among them, P rate is the growth rate, T t1 is the transmembrane pressure difference corresponding to time t1, T t2 is the transmembrane pressure difference corresponding to time t2.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the membrane system early warning and control method based on Kalman filtering described in any one of claims 1 to 9 is implemented.

Images