System for determining filter state

Abstract

An exemplary system for determining filter status is provided. The system includes a database configured to electronically store data containing historical performance data of one or more water filtration elements. The system includes a processing unit that communicates with the database. The processing unit receives historical performance data as input and is configured to identify one or more data points in the historical performance data that indicate cleaning events of the water filtration elements. The processing unit divides the historical performance data into individual data segments, each of which represents data in the historical performance data between each identified cleaning event, and is configured to fit a mathematical function to each of the individual data segments and to identify the type of fouling or scaling, or no-fouling or no-scaling conditions, occurring between each of the individual data segments based on the mathematical function fitted to each individual data segment.

Description

[Technical Field]

[0001] Cross-reference of related applications This application claims priority to Spanish Utility Model Application No. 202330840, filed on 12 May 2023, which is incorporated herein by reference in its entirety.

[0002] This invention relates to a method and system for the automatic monitoring of filter membranes in water treatment facilities, and for the detection and identification of the type of fouling, or contamination, of the membranes in water treatment facilities. The method and system can be used to efficiently determine when cleaning of the filter membrane is necessary and to provide accurate guidance on the filter membrane based on the identified type of fouling. [Background technology]

[0003] Water treatment systems are available in various configurations, including systems that perform pressure-driven membrane separation processes for water filtration. Such pressure-driven membrane separation processes allow for the removal of a wide range of neutral and ionic species from a fluid. In order of decreasing pore size, membranes are generally classified into several categories: microfiltration (MF), ultrafiltration (UF), nanofiltration (NF), and reverse osmosis (RO). Microfiltration is used to remove suspended particles with a particle size greater than 0.1 microns. Ultrafiltration typically removes dissolved molecules with a molecular weight greater than 5,000 Daltons. Nanofiltration membranes allow at least some salts to pass through, but typically have a high retention rate for organic compounds with a molecular weight greater than about 200 Daltons. Reverse osmosis membranes have a high retention rate for almost all species.

[0004] UF systems or elements are most commonly used in water treatment facilities, wastewater treatment, and as pretreatment in the food industry. A crucial factor for UF applications is that the membrane achieves high rejection rates for suspended solids, i.e., suspended matter and large molecules, while maintaining high flux.

[0005] NF and RO membranes are most commonly used in applications such as desalination of seawater or brackish water, production of ultrapure water, color removal, wastewater treatment, and concentration of liquids for food use. A crucial factor in almost all NF and RO applications is that the membrane achieves a high rejection rate for small solute molecules while maintaining a high flux.

[0006] A spiral element is the most common configuration for RO and NF membranes. A traditional spiral element design is shown in Figure 1. The element generally includes a membrane envelope 2 and a feed solution spacer sheet 4 wound around a central permeate collection tube 6. The envelope 2 includes two membrane sheets 8 surrounding a permeate carrier sheet 10, and the structure is held together by adhesive 12 along its edges 14, 16, and 18. The fourth edge 20 of the envelope 2 abuts against the permeate collection tube 6 so that the permeate carrier sheet 10 is in fluid contact with an opening 22 in the permeate collection tube 6. Each envelope 2 is separated by a feed solution spacer sheet 4, which is similarly wound around the collection tube 6. The feed solution spacer 4 is in fluid contact with both ends 24, 26 of the element and functions as a conduit for the feed solution across the front surface 28 of the membrane 8. The direction of the feed fluid flow 30 is from the inlet end 24 to the concentric end 26, and this direction is parallel to the axis 32 of the central permeate collection tube 6. As such, the “feed fluid” flows axially through the feed fluid spacer sheet 4 and exits from the opposite end as the “concentrated fluid”. The “permeate” passes through the membrane envelope 2 under pressure and is directed to the permeate collection tube 6 by the permeate carrier sheet 10.

[0007] Ultrafiltration and microfiltration can be carried out using membranes fabricated in one of two forms—flat sheet or tubular. Information on these ultrafiltration and microfiltration processes and associated equipment can be found, for example, in Wenten, I Gede, Ultrafiltration in Water Treatment and Its Evaluation as Pretreatment for Reverse Osmosis System (2008). Ultrafiltration or microfiltration membranes in flat sheet form can be wound into a spiral module in a manner similar to that shown in Figure 1. Such a module operates in a cross-flow mode in which feedwater flows past the membrane surface and is separated into a permeate flow that flows through the membrane and a concentrate flow that continues past the membrane surface. In cross-flow mode, the permeate flow is substantially free of suspended solids and macromolecules. The concentrate flow has a considerable concentration of suspended solids and macromolecules. Alternatively, tubular ultrafiltration or microfiltration membranes often appear in the form of multiple hollow fibers bundled together, where the filtrate flow is extracted from the feed fluid flow through the membrane, driven by the pressure difference across the membrane. Such hollow fiber designs are often used in a "dead-end" or "filtration" mode, where only one feed fluid and one filtrate flow are present, and water passes through while solids or large organic matter accumulate on the membrane surface. The water flow can be from the outside of the fibers to the inside (out-in) or from the inside of the fibers to the outside (in-out).

[0008] Information on ultrafiltration systems can be found, for example, in the DuPont Ultrafiltration Technical Manual, Version 4 (June 2019), and an updated version is available at https: / / inaqua.de / assets / NEUE-Datenblaetter / Neue-DuPont / Manua-ultrafiltration-technical.pdf, last accessed May 2, 2024. Ultrafiltration systems typically operate in filtration mode most of the time. Feed water is pumped through the membrane and converted into filtrate. Filtration cycles typically range from approximately 20 to 90 minutes, depending on the feed source and quality. Ultrafiltration systems are typically designed to operate at a constant flow rate. As solids accumulate on the membrane surface, the transmembrane pressure (TMP) increases, and eventually, foulants must be removed by a backwash sequence. The backwash sequence is generally initiated based on time. Alternatively, it may be initiated based on the volume of the filtrate and / or the TMP setpoint (the latter being more suitable for feedwater quality with greater variability). The backwash sequence essentially involves pumping filtrate (or clean water), or a mixture of air and filtrate, from the filtrate side of the membrane to the feedwater side, in order to push out accumulated foulant from the membrane pore channels and / or the membrane surface. The backwash sequence includes one or more steps: air scrub, gravity drain, backwash through the top module outlet, backwash through the bottom module outlet, and a final forward flush or rinse. Once the air scrub step is complete, the module is drained by gravity, allowing the material removed from the membrane surface by the preceding air scrub step to be washed away from the system. The duration of this step depends on the volume of the system and the piping layout, but is typically set to about 30-60 seconds. If gravity drainage is not possible due to the system configuration, or if gravity drainage takes too long, the process can be replaced by a forced flush through the module's bottom outlet using a backwash pump. However, this consumes more water and energy.

[0009] After the backwash step through the module's upper outlet, the filtrate continues to flow from the inside to the outside of the fibers, but is washed away through the module's bottom outlet to ensure that the entire length of the fibers is cleaned. The backwash pump is not stopped during the transition between the upper backwash (module's upper outlet) and the lower backwash (module's lower outlet). Valves must be sequenced to prevent membrane damage. Similar to the upper backwash step, the duration of the lower backwash is typically about 30–45 seconds, and chlorine may be added to help remove foulants and / or inhibit microbial activity. The backwash step may be repeated multiple times depending on the degree of fouling. Monitoring TMP and backwash wastewater quality may be useful in optimizing the duration of these steps. Less frequently, chemically enhanced cleaning (CEC) processes such as chemically enhanced backwash (CEB) may be performed. During CEB, chemicals (e.g., chlorine, acid, or base) are added to the backwash flow to enhance the cleaning effect. CEB is performed in a similar manner to backwashing, except that an immersion step is added after the addition of chemicals. The frequency of CEB depends on the supply water quality, but is typically once a day to once a week. The order of the backwashing step can be reversed to ensure that the system remains filled with the chemical solution during the immersion time. Occasionally, a Clean-in-Place (CIP) is performed to restore system performance. CIP is typically performed about every 3 to 12 months and involves immersion and recirculation of the chemical solution via a UF or MF module. CIP chemicals can include acids, alkalis, or special solutions.

[0010] Over time, the performance level of the filter elements in any of the filtration systems discussed herein may decrease due to the accumulation of foreign matter on the feed liquid spacer or membrane (e.g., fouling) and / or the precipitation of salts (e.g., scaling). Such a decrease in performance level may affect the quality of the filtered water and / or place greater pressure on the water filtration system during operation, potentially leading to increased maintenance efforts over time.

[0011] Manufacturers of filtration elements and / or water treatment systems generally establish specific guidelines for cleaning filtration elements. Typically, this involves monitoring the overall operation of the water treatment system and shutting down the system to clean the filters when certain operational characteristics are detected (e.g., pressure drops across the elements exceed recommended thresholds). For proper cleaning, it is generally recommended to shut down the system, clean the affected filtration elements, and send some of the affected elements to a laboratory for a detailed analysis of fouling and / or scaling types. However, typically, water treatment plant operators push the filtration system beyond the recommended guidelines to avoid prolonged system shutdowns. This can lead to further problems, such as pump overheating or poor product quality, which would then force a system shutdown.

[0012] At this point, instead of sending the filter element to the lab, the operator might rely on experience to guess the type of fouling / scaling occurring and clean the system using chemicals they think might be appropriate. In such a case, the type of chemical used may not be appropriate and may not address the actual fouling / scaling occurring. This could result in the cleaned filter element having a lower performance baseline than the performance level it would have had if a proper clean had been performed. After repeated cleaning attempts, the filter element and / or water treatment facility may face further downtime, resulting in increased operating costs and potentially substandard filtered water. [Overview of the Initiative] [Problems that the invention aims to solve]

[0013] It is clear that there is still a need for efficient and reliable automated systems to evaluate historical data, determine filter status, and recommend appropriate measures to correct fouling or scaling conditions. [Means for solving the problem]

[0014] Embodiments of the present disclosure provide an exemplary system for determining filter conditions. The system includes a database configured to electronically store data, the data including historical performance data of one or more water filtration elements. The system includes a processing unit that communicates with the database. The processing unit receives historical performance data as input, identifies one or more data points in the historical performance data indicating cleaning events of one or more water filtration elements, divides the historical performance data into individual data segments, each of the individual data segments representing data in the historical performance data between each identified cleaning event, fits a mathematical function to each of the individual data segments, and is configured to identify the type of fouling occurring between each of the individual data segments based on the mathematical function fitted to each of the individual data segments.

[0015] Embodiments of the present disclosure provide an exemplary method for determining a filter state. The method includes receiving historical performance data of one or more water filtration elements as input to a system for determining the filter state. The system for determining the filter state includes a database configured to electronically store the historical performance data and a processing unit that communicates with the database. The method includes identifying one or more data points in the historical performance data that indicate a cleaning event of one or more water filtration elements. The method includes dividing the historical performance data into individual data segments, each of which represents data in the historical performance data between each identified cleaning event. The method includes fitting a mathematical function to each of the individual data segments. The method includes identifying the type of fouling that occurs between each of the individual data segments based on the mathematical function fitted to each of the individual data segments.

[0016] Embodiments of the present disclosure provide an exemplary non-temporary computer-readable medium for storing instructions for determining a filter state that can be executed by a processing unit. Execution of an instruction by the processing unit causes the processing unit to receive historical performance data of one or more water filtration elements as input to a system for determining the filter state. The system for determining the filter state includes a database configured to electronically store the historical performance data and a processing unit that communicates with the database. Execution of an instruction by the processing unit causes the processing unit to identify one or more data points in the historical performance data that indicate a cleaning event for one or more water filtration elements. Execution of an instruction by the processing unit causes the processing unit to divide the historical performance data into individual data segments, each of which represents data in the historical performance data between each identified cleaning event. Execution of an instruction by the processing unit causes the processing unit to fit a mathematical function to each of the individual data segments. Execution of an instruction by the processing unit causes the processing unit to identify the type of fouling occurring between each of the individual data segments based on the mathematical function fitted to each individual data segment.

[0017] In some embodiments, with respect to NF filtration systems and RO filtration systems, a filtration module failure mode in a data segment can be identified as scaling if at least one of the following conditions is met: (a) the coefficient of determination or Akaike Information Criterion (AIC) of dP for a linear polynomial is greater than 0.7; (b) the coefficient of determination or Akaike Information Criterion (AIC) of dP for a quadratic polynomial is greater than 0.7; (c) the difference between the coefficient of determination or Akaike Information Criterion (AIC) values ​​of the quadratic polynomial and the linear polynomial fitted to dP is less than 0.3; (d) the coefficient of determination or Akaike Information Criterion (AIC) of Sp for a linear polynomial is greater than 0.7; (e) the coefficient of determination or Akaike Information Criterion (AIC) of Sp for a quadratic polynomial is greater than 0.7; (f) the difference between the coefficient of determination or Akaike Information Criterion (AIC) of the quadratic polynomial and the linear polynomial fitted to Sp (c) The difference in values ​​is less than 0.3, (g) The Akaike Information Criterion (AIC) of Pf for a linear polynomial is greater than 0.7, (h) The coefficient of determination or Akaike Information Criterion (AIC) of Pf for a quadratic polynomial is greater than 0.7, (i) The difference in the coefficient of determination or Akaike Information Criterion (AIC) values ​​for the quadratic polynomial and the linear polynomial fitted to Pf is less than 0.3, (j) dP increases over time in the data segment, (k) Pf decreases over time in the data segment, and / or (l) Sp increases over time in the data segment.

[0018] In some embodiments, with respect to NF filtration systems and RO filtration systems, a filtration module failure mode in the data segment can be identified as particulate fouling if at least one of the following conditions is met: (a) the coefficient of determination or Akaike Information Criterion (AIC) of dP for a linear polynomial is greater than 0.7; (b) the coefficient of determination or Akaike Information Criterion (AIC) of dP for a quadratic polynomial is greater than 0.7; (c) the difference between the coefficient of determination or Akaike Information Criterion (AIC) values ​​of the quadratic polynomial and the linear polynomial that fit dP is less than 0.3; (d) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a linear polynomial is greater than 0.7; (e) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a quadratic polynomial is greater than 0.7 (f) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial fitted to Pf is less than 0.3, (g) The coefficient of determination or Akaike Information Criterion (AIC) of Sp for the linear polynomial is greater than 0.7, (h) The coefficient of determination or Akaike Information Criterion (AIC) of Sp for the quadratic polynomial is greater than 0.7, (i) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial fitted to Sp is less than 0.3, (j) dP increases rapidly over time in the data segment, (k) Pf decreases rapidly over time in the data segment, and / or (l) Sp increases rapidly over time in the data segment.

[0019] In some embodiments, with respect to NF filtration systems and RO filtration systems, a filtration module failure mode in the data segment can be identified as biofouling, i.e., biofouling, if at least one of the following conditions is met: (a) the coefficient of determination or Akaike Information Criterion (AIC) of dP for a linear polynomial is greater than 0.4; (b) the coefficient of determination or Akaike Information Criterion (AIC) of dP for a quadratic polynomial is greater than 0.7; (c) the difference between the coefficient of determination or Akaike Information Criterion (AIC) values ​​for the quadratic polynomial and the linear polynomial is greater than 0.3; (d) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a linear polynomial is greater than 0.4; (e) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a quadratic polynomial is greater than 0.4 (f) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial fitted to Pf is greater than 0.3, (g) The coefficient of determination or Akaike Information Criterion (AIC) of Sp for the linear polynomial is greater than 0.4, (h) The coefficient of determination or Akaike Information Criterion (AIC) of Sp for the quadratic polynomial is greater than 0.7, (i) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial fitted to Sp is greater than 0.3, (j) dP increases over time in the data segment, (k) Pf decreases over time in the data segment, and / or (l) Sp increases over time in the data segment.

[0020] In some embodiments, with respect to NF filtration systems and RO filtration systems, a filtration module failure mode in the data segment can be identified as chemical decomposition (e.g., by oxidation or similar) if at least one of the following conditions is met: (a) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a linear polynomial is greater than 0.7; (b) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a quadratic polynomial is greater than 0.7; (c) the difference between the coefficient of determination or Akaike Information Criterion (AIC) values ​​of the quadratic and linear polynomials fitted to Pf is less than 0.3; (d) the coefficient of determination or Akaike Information Criterion (AIC) of Sp for a linear polynomial is greater than 0.7; (e) the coefficient of determination or (f) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial fitted to Sp is less than 0.3. (g) The coefficient of determination or Akaike Information Criterion (AIC) of dP for the linear polynomial is greater than 0.7. (h) The coefficient of determination or Akaike Information Criterion (AIC) of dP for the quadratic polynomial is greater than 0.7. (i) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial fitted to dP is less than 0.3. (j) dP remains constant over time in the data segment. (k) Pf increases over time in the data segment, and / or (l) Sp increases over time in the data segment.

[0021] In some embodiments, with respect to NF filtration systems and RO filtration systems, a filtration module failure mode in the data segment can be identified as a physical integrity failure (e.g., due to membrane damage, mechanical reasons, physical damage, etc.) if at least one of the following conditions is met: (a) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a linear polynomial is greater than 0.7; (b) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a quadratic polynomial is greater than 0.7; (c) the difference in the coefficient of determination or Akaike Information Criterion (AIC) values ​​of the quadratic polynomial and the linear polynomial that fit Pf is less than 0.3; (d) the coefficient of determination or Akaike Information Criterion (AIC) of Sp for a linear polynomial is greater than 0.7; (e) the determination of Sp for a quadratic polynomial (f) The coefficient or Akaike Information Criterion (AIC) is greater than 0.7, (g) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic and linear polynomials fitted to Sp is less than 0.3, (h) The coefficient of determination or Akaike Information Criterion (AIC) for dP for the linear polynomial is greater than 0.7, (i) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for dP for the quadratic polynomial is less than 0.3, (j) dP remains constant over time in the data segment, (k) Pf increases rapidly over time in the data segment, and / or (l) Sp increases rapidly over time in the data segment.

[0022] In some embodiments, with respect to NF filtration systems and RO filtration systems, a filtration module failure mode in the data segment can be identified as organic fouling if at least one of the following conditions is met: (a) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a linear polynomial is greater than 0.4, (b) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a quadratic polynomial is greater than 0.7, (c) the difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial fitted to Pf is greater than 0.3, (d) the coefficient of determination or Akaike Information Criterion (AIC) S of Sp for a linear polynomial is greater than 0.4, (e) the coefficient of determination of Sp for a quadratic polynomial is greater than 0.4 (f) The coefficient of determination or Akaike Information Criterion (AIC) is greater than 0.7, (g) The difference between the coefficient of determination or Akaike Information Criterion (AIC) values ​​of the quadratic polynomial and the linear polynomial fitted to Sp is greater than 0.3, (h) The coefficient of determination or Akaike Information Criterion (AIC) of dP for the linear polynomial is greater than 0.7, (i) The difference between the coefficient of determination or Akaike Information Criterion (AIC) values ​​of dP for the quadratic polynomial is less than 0.3, (j) dP remains constant over time in the data segment, (k) Pf decreases over time in the data segment, and / or (l) Sp decreases over time.

[0023] In some embodiments, the term “lower limit” as used herein may be greater than 0 and less than 10%, or 0 to 8%, 0 to 5%, 0 to 3%, or 0 to 2%. In some embodiments, the lower limit may be, for example, about 0-10% (inclusive), 0-9% (inclusive), 0-8% (inclusive), 0-7% (inclusive), 0-6% (inclusive), 0-5% (inclusive), 0-4% (inclusive), 0-3% (inclusive), 0-2% (inclusive), 0-1% (inclusive), 1-10% (inclusive), 2-10% (inclusive), 3-10% (inclusive), 4- It may be 10% (including both ends), 5-10% (including both ends), 6-10% (including both ends), 7-10% (including both ends), 8-10% (including both ends), 9-10% (including both ends), 1-9% (including both ends), 2-8% (including both ends), 3-7% (including both ends), 4-6% (including both ends), 2-6% (including both ends), 4-8% (including both ends), 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, etc. In some embodiments, the term "difference constant" as used herein may be greater than 0 and less than 10, or from 0 to 8, 0 to 5, 0 to 3, or 0 to 2. In some embodiments, the difference constant can be, for example, about 0-10 (inclusive), 0-9 (inclusive), 0-8 (inclusive), 0-7 (inclusive), 0-6 (inclusive), 0-5 (inclusive), 0-4 (inclusive), 0-3 (inclusive), 0-2 (inclusive), 0-1 (inclusive), 1-10 (inclusive), 2-10 (inclusive), 3-10 (inclusive), 4-10 (inclusive), 5-10 (inclusive), 6-10 (inclusive), 7-10 (inclusive), 8-10 (inclusive), 9-10 (inclusive), 1-9 (inclusive), 2-8 (inclusive), 3-7 (inclusive), 4-6 (inclusive), 2-6 (inclusive), 4-8 (inclusive), 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, etc. In some embodiments, the terms “Membrane Cleaning Event (MCE) Reference Time” or “Cleaning In-Place (CIP) Reference Time” as used herein may be greater than 0 and less than 10 days, or 0 to 8, 0 to 5, 0 to 3, or 0 to 2 days.In some embodiments, the MCE reference time is, for example, approximately 0-10 days (inclusive), 0-9 days (inclusive), 0-8 days (inclusive), 0-7 days (inclusive), 0-6 days (inclusive), 0-5 days (inclusive), 0-4 days (inclusive), 0-3 days (inclusive), 0-2 days (inclusive), 0-1 day (inclusive), 1-10 days (inclusive), 2-10 days (inclusive), 3-10 days (inclusive), 4 The duration may be ~10 days (including both ends), 5~10 days (including both ends), 6~10 days (including both ends), 7~10 days (including both ends), 8~10 days (including both ends), 9~10 days (including both ends), 1~9 days (including both ends), 2~8 days (including both ends), 3~7 days (including both ends), 4~6 days (including both ends), 2~6 days (including both ends), 4~8 days (including both ends), 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 7 days, 8 days, 9 days, 10 days, etc. In some embodiments, MCE includes any type of membrane cleaning event (both chemical and physical) to restore filtration performance, e.g., flow reversal, backwashing, chemical backwashing, and / or stationary chemical cleaning. An exemplary method may include a step of comparing performance parameters before and after cleaning to ensure proper cleaning and / or fouling identification. As such, exemplary systems and methods can be used to identify filtration module failure modes and provide guidance for optimal cleaning to ensure efficient operation of the water filtration system.

[0024] With respect to ultrafiltration (UF) and / or microfiltration (MF), the type of fouling correlates with the feedwater conditions and the details of the operating routine (see, for example, Figure 12C). By improving the operating routine parameters, at least some of the original irreversible types of fouling can be shifted to reversible types. Therefore, the system's evaluation and identification of fouling types provides guidance for optimizing system operation.

[0025] In some embodiments, with respect to UF and / or MF filtration systems, the filtration module failure mode in the data segment may be identified as cake-forming fouling if TMP (TMP_F) increases linearly during filtration, but TMP (TMP_BW) during backwashing and TMP (TMP_CEC) during chemical wash remain constant. In some embodiments, with respect to UF and MF filtration systems, the filtration module failure mode in the data segment may be identified as bio / organic fouling if TMP_F increases, but both TMP_BW and TMP_CEC increase linearly or non-linearly when no oxidizing agent or caustic agent is used in CEC. In some embodiments, with respect to UF and MF filtration systems, the filtration module failure mode in the data segment may be identified as inorganic fouling if TMP_F increases, but both TMP_BW and TMP_CEC increase linearly or non-linearly when no acid is used in CEC. In some embodiments, with respect to UF and MF filtration systems, a filtration module failure mode in the data segment may be identified as chemical decomposition if a gradual (weeks / months) increase in membrane filtrate turbidity is observed. In some embodiments, with respect to UF and MF filtration systems, a filtration module failure mode in the data segment may be identified as a physical integrity problem if a rapid (hours / days) increase in membrane filtrate turbidity is observed. Furthermore, a physical integrity problem may also be confirmed by an increase in the silt density index (SDI), an increase in the pressure decay rate during integrity testing, the presence of bubbles in the filtrate port during integrity testing, an increase in the modified fouling index (MFI), and / or the total suspended solids concentration in the filtrate.

[0026] As used herein, the term “cleanliness” refers to the change in the primary index before and after the i-th MCE, or the change in the primary index (for UF / MF, this may be the normalized TMP), preferably as a percentage change, compared to the change in the primary index before the i-th MCE and after the (i1) MCE, or to the initial value at the start of the overall operation. Higher cleanliness signifies a better recovery of membrane performance and can be used as a secondary variable to help distinguish between types of fouling. CEC cleanliness of 60–95% may indicate cake-forming fouling, while CEC cleanliness of less than 60% may indicate either biological organic or inorganic fouling.

[0027] Embodiments of the present disclosure provide an exemplary method for analyzing the performance of a feedwater filtration system. This method includes collecting performance data from the filtration system, normalizing the dataset to startup conditions, removing physically inconsistent data points from the dataset, removing statistical outliers from the dataset, identifying membrane cleaning events in the dataset, dividing the dataset into a plurality of data segments between the membrane cleaning events, fitting a mathematical function to the data segments, analyzing the coefficient of determination of the mathematical function, calculating the derivative of the mathematical function with respect to time, and determining the type of fouling based on the derivative and the coefficient of determination.

[0028] The filtration system includes at least one filtration module. The filtration module includes one or more of the following: a reverse osmosis element, a nanofiltration element, or an ultrafiltration element. Performance data may be collected by sensors installed within the filtration system. Normalization of the dataset may be performed to take into account variations in the feedwater temperature, the feedwater composition, or the feedwater pressure. Outlier removal in the dataset is performed by removing the previous data point (y t-1 ) data point (y t The steps include calculating the percentage change of the data point (y t ) Next data point (y t+1The steps are: to calculate the percentage change of ), to compare the previous percentage change with the lower limit value, and if the percentage change exceeds the lower limit value, the data point (y t This can be carried out by statistical means, which includes the step of identifying ) as an outlier.

[0029] The removal of physically inconsistent data points in the aforementioned dataset may be performed by applying a set of criteria including one or more of the following: the feed flow must be greater than the concentrate flow; the feed flow must be greater than the permeate flow; the feed pressure must be greater than the concentrate pressure; the feed pressure must be greater than the permeate pressure; the feed pressure must be greater than the pressure drop across the filtration module; the permeate conductivity must be less than the feed conductivity; the pressure drop across the filtration module must be less than or equal to the difference between the feed pressure and the concentrate pressure; and / or the osmotic pressure of the feed flow must be less than the feed pressure.

[0030] The dataset may include normalized pressure drop (dP) before and after the filtration module, normalized permeate flow (Pf) from the filtration module, and normalized salt passage (Sp) of the filtration module. Membrane washing events can be identified using a difference sequence method of dP and Pf, where the difference sequence method is defined as the data point (y t ) and the previous data point (y t-1The steps may include: calculating the difference between dP and Pf to create the difference series of dP and Pf; calculating the mean and standard deviation of the difference series of dP and Pf; comparing each value in the difference series of dP or Pf with the sum of the product of the difference constant and the corresponding standard deviation of the difference series and the corresponding difference series; checking whether there are gaps in the dataset; and checking whether the gaps are wider than the Clean In Place (CIP) reference time (also referred herein as the Membrane Cleaning Event (MCE) reference time). The difference constant is a positive integer, preferably a number between 1 and 10 (inclusive); more preferably a number between 3 and 7 (inclusive); and even more preferably 5. The CIP reference time is a time interval defined by the user as representing a typical duration of CIP for the system under consideration, and can range from a few minutes to several days. Exemplary values ​​for the CIP reference time may be, for example, 120 minutes, 1 day, etc. Preferred CIP reference times are 5 minutes to 2 weeks, with 1 day being a more preferred CIP reference time.

[0031] The system provided herein includes a database configured to electronically store data containing historical performance data of one or more water filtration elements. The system includes a processing unit that communicates with the database. The processing unit is configured to receive historical performance data as input and to identify one or more data points in the historical performance data that indicate cleaning events of the water filtration elements. The processing unit is configured to divide the historical performance data into individual data segments, each of which represents data in the historical performance data between each identified cleaning event. The processing unit is configured to fit a mathematical function to each of the individual data segments. Based on the mathematical function fitted to each individual data segment, the processing unit is configured to identify the type of fouling that occurs between each of the individual data segments.

[0032] In some embodiments, the apparatus may be configured to normalize historical performance data to the startup conditions of one or more water filtration elements before identifying a cleaning event. In some embodiments, when normalizing historical performance data, the apparatus may take into account variations in at least one of the following: feedwater temperature, feedwater composition, or feedwater pressure.

[0033] In some embodiments, the processing apparatus may be configured to identify and remove physically inconsistent data points from past performance data before identifying a cleaning event. With respect to reverse osmosis or nanofiltration, in some embodiments, physically inconsistent data points are removed from past performance data if at least one of the following conditions is not true: feed flow is greater than concentrate flow, feed flow is greater than permeate flow, feed pressure is greater than concentrate pressure, and / or feed pressure is greater than permeate pressure. In some embodiments, physically inconsistent data points are removed from past performance data if at least one of the following conditions is not true: feed flow is greater than concentrate flow, feed flow is greater than permeate flow, feed pressure is greater than concentrate pressure, feed pressure is greater than permeate pressure, feed pressure is greater than the pressure drop across one or more water filtration elements, permeate conductivity is less than the feed conductivity, the pressure drop across one or more water filtration elements is less than or equal to the difference between feed pressure and concentrate pressure, and / or the osmotic pressure of the feed flow is lower than the feed pressure.

[0034] With respect to ultrafiltration or microfiltration, in some embodiments, the removal of physically inconsistent data points in the dataset may be carried out by applying a set of criteria including one or more of the following: feed pressure is greater than or equal to the intermembrane pressure; intermembrane pressure before and after the filtration module is less than or equal to the difference between the feed pressure and the filtrate pressure; and feed turbidity is greater than filtrate turbidity. In some embodiments, the criteria also include: filtrate pressure must be greater than the intermembrane pressure.

[0035] In some embodiments, the processing device may be configured to identify and remove data points of statistical outliers from past performance data before identifying a membrane cleaning event. In such embodiments, identifying and removing data points of statistical outliers from past performance data involves calculating the percentage change of a data point (y t-1 ) from the previous data point (y t ), calculating the percentage change of a data point (y t ) from the next data point (y t+1 ), comparing the previous percentage change with a lower limit value, and if the percentage change exceeds the lower limit, identifying the data point (y t ) as an outlier.

[0036] In some embodiments, identifying a cleaning event may include using a difference series method on the normalized pressure drop (dP) and the normalized permeate flow (Pf) from past performance data. Using the difference series method includes: (i) calculating the difference between a data point (y t ) and the previous data point (y t-1 ) to create a difference series of the normalized pressure drop (dP) and the normalized permeate flow (Pf); (ii) calculating the mean and standard deviation of the difference series of the normalized pressure drop (dP) and the normalized permeate flow (Pf); (iii) comparing each value within the difference series of the normalized pressure drop (dP) and the normalized permeate flow (Pf) with the sum of the product of a difference constant and the corresponding standard deviation of the difference series and the corresponding difference series; (iv) determining whether there is a gap in the data set; and (v) determining whether the gap is wider than the CIP reference time.

[0037] In some embodiments where the system is an ultrafiltration (UF) or microfiltration (MF) system, identifying a wash event may involve using a difference-sequence method for transmembrane pressure drop (TMP), feed turbidity, and permeability from historical performance data. Using a difference-sequence method may involve (i) calculating the difference between the time values ​​of a data point (t) and the previous data point (tx) to create a time-difference series; (ii) calculating the difference between a data point (yt) and the previous data point (yt-x) to create a difference-sequence for transmembrane pressure drop (TMP), feed turbidity, and permeability; (iii) comparing each value in the difference-sequence for time, transmembrane pressure drop (TMP), feed turbidity, and permeability to thresholds for backwash, CEB, and CIP MCE events; (iv) determining whether the comparison matches the backwash MCE criterion; (v) determining whether the comparison matches the CEB MCE criterion; and (vi) determining whether the comparison matches the CIP MCE criterion.

[0038] In some embodiments, the processing unit may be configured to analyze a predictive model, such as the coefficient of determination of a mathematical function or the Akaike Information Criterion (AIC), and to calculate the derivative of the mathematical function with respect to time. In such embodiments, the processing unit may be configured to identify the type of fouling based on the derivative of the coefficient of determination or the Akaike Information Criterion (AIC). In some embodiments, one or more water filtration elements may be at least one of reverse osmosis elements, nanofiltration elements, ultrafiltration elements, ultrafiltration elements, or microfiltration elements. The system may include one or more sensors configured to detect and transmit data about pressure drop, salt permeability, and permeate flow associated with each of the one or more water filtration elements for storage in a database as historical performance data.

[0039] In some embodiments, historical performance data may include the pressure drop, salt permeability, and permeate flow rate of each of one or more water filtration elements. If no changes are detected in the pressure drop, salt permeability, and permeate flow rate, the treatment device can identify a fouling-free condition. If the increase in pressure drop fits a quadratic polynomial function, the treatment device can identify the type of fouling as biofouling. In some examples of biofouling, the salt permeability may increase and the permeate flow rate may decrease. If the decrease in salt permeability fits a quadratic polynomial function, the treatment device can identify the type of fouling as organic fouling. In some examples of organic fouling, the pressure drop may not change and the salt permeability may decrease. If a sharp increase in the pressure drop of the element at the first position (e.g., the tip filtration element in Figure 19) fits a linear polynomial function, the treatment device can identify the type of fouling as particulate fouling. In some examples of particulate fouling, the salt permeability may increase and the permeate flow rate may decrease. In some cases, such effects may be found mainly in the tip element. If the gradual increase in pressure drop across the tail end element (e.g., the tail end filter element in Figure 19) fits a linear polynomial function, the processing unit can identify the type of fouling as scaling. In some cases of scaling, salt permeability may increase and permeate flow may decrease. In some cases, such effects can be found primarily in the tail end element. If the increase in permeate flow fits a linear polynomial function and the increase in salt permeability fits a linear polynomial function, the processing unit can identify the type of fouling as a complete membrane damage event. In some cases of oxidation or membrane damage events, the pressure drop may not change.

[0040] In some embodiments, the processing unit may be configured to receive current performance data from one or more water filter elements as input, such as from one or more sensors (for example, multiple sensors may be installed on the same filter element, and / or sensors may be installed on multiple filter elements in the facility to provide a more comprehensive fouling diagnosis), analyze the current performance data based on a previous analysis of historical performance data and identification of the type of fouling occurring between each of the individual data segments, estimate the type of current fouling occurring in the water filter element, and / or output recommendations for cleaning procedures specific to the type of current fouling occurring in the water filter element. The data collected from the sensors may be filter element specific to identify whether specific conditions are occurring at the front or rear end of the pressure vessel to help estimate the type of fouling occurring. In some embodiments, the current performance data and / or historical performance data can be used to predict potential fouling that may occur in the future (for example, based on the cycle of repeated fouling during a specific time of year, and / or based on environmental conditions).

[0041] Exemplary systems for determining filter conditions can be used by plant operators to improve system performance. The system can analyze past performance to identify past membrane cleaning events (MCEs) and types of filtration module failure modes based on system performance during and between MCE periods. In some embodiments relating to UF or MF systems, MCEs include any type of membrane cleaning event (both chemical and physical) to restore filtration performance, e.g., flow reversal, backwashing, chemical backwashing (CEB), and / or stationary chemical cleaning. In other embodiments relating to NF or RO, MCEs include stationary cleaning. The systems provided herein can be used in ultrafiltration, microfiltration, nanofiltration, and / or reverse osmosis systems to identify fouling, scaling, and / or membrane failure modes. In some embodiments, the system can be used to predict when the next cleaning should / will be performed and to recommend an optimal cleaning procedure.

[0042] Membrane fouling is a complex problem in nanofiltration and reverse osmosis systems. As such, identifying the dominant fouling mechanism during the filtration process can be crucial for predicting the next wash cycle and controlling fouling. Real-time (or substantially real-time) monitoring of water treatment facilities to detect and identify fouling types enables early detection and response by facility operators, preventing excessive fouling and potential adverse effects on the rest of the system. The system provides a precise means for determining the type of fouling occurring, thereby providing accurate recommendations for wash cycles to ensure that appropriate measures are taken to address the fouling event. In some embodiments, the system can be used to predict periodic and / or seasonal fouling, minimizing and / or avoiding fouling events.

[0043] Embodiments of the present disclosure provide an exemplary method for analyzing the performance of a supply water filtration system. This method includes collecting performance data from the filtration system; normalizing the dataset to startup conditions; removing physically inconsistent data points from the dataset; removing statistical outliers from the dataset; identifying membrane cleaning events in the dataset; dividing the dataset into multiple data segments between the membrane cleaning events; fitting a mathematical function to the data segments; analyzing a predictive model, such as the Akaike Information Criterion (AIC) or coefficient of determination of the mathematical function; calculating the derivative of the mathematical function with respect to time; and determining the type of fouling based on the derivative and the coefficient of determination or Akaike Information Criterion (AIC). (See, for example, H. Akaike, “A new look at the statistical model identification,” IEEE Transactions on Automatic Control, vol.19, no.6, pp.716-723 (December 1974)).

[0044] The filtration system includes at least one filtration module. The filtration module includes one or more of the following: an ultrafiltration element, a microfiltration element, a reverse osmosis element, a nanofiltration element, and an ultrafiltration element. Performance data can be collected by sensors installed in the filtration system. Normalization of the data set may be performed to take into account variations in the feedwater temperature, the feedwater composition, and / or feedwater pressure.

[0045] In some embodiments, outlier removal in the dataset, which includes reverse osmosis data or nanofiltration data, is performed by removing the previous data point (y t-1 ) data point (y t The steps include calculating the percentage change of the data point (y t ) Next data point (y t+1The steps are: to calculate the percentage change of ), to compare the previous percentage change with the lower limit value, and if the percentage change exceeds the lower limit value, the data point (y t This can be done by statistical means, which include the step of identifying the value as an outlier. The decline limit is defined here as the number that explains the maximum or minimum difference between two other numbers that are allowed by the system before the measure is triggered. Alternatively, well-known statistical methods such as the Inter-Quantile Range (IQR) method can be used to identify data points that fall outside the acceptable limits (set by or within the system).

[0046] In some embodiments, outlier removal in the dataset, including ultrafiltration or microfiltration data, can be performed by the system using the method described herein. First, the frequency of the data (e.g., per minute, per hour, per day) is compared to general guidelines for at least the following types of membrane cleaning event time scales (air scrubbing (minutes), fluid pressure backwash (minutes), chemical backwash (hours), and in-place cleaning (CIP) (days). If the data is available on the fluid pressure backwash time scale, a second dataset is generated by averaging the dataset so that it is now available on the chemical backwash time scale (hours). A third dataset is generated by averaging the dataset so that it is now available on the in-place cleaning time scale (days). The first round of outlier detection is performed on the previous data point (y t-1 ) data point (y t The steps include calculating the percentage change of the data point (y t ) Next data point (y t+1 The steps are: to calculate the percentage change of ), to compare the previous percentage change with the lower limit value, and if the percentage change exceeds the lower limit value, the data point (y tThis can be performed on a third dataset by statistical means including the step of identifying outliers. Alternatively, well-known statistical methods such as the interquartile range (IQR) method can be used to identify data points that fall outside the acceptable limits.

[0047] The second round of outlier detection involves the previous data point (y t-1 ) data point (y t The steps include calculating the percentage change of the data point (y t ) Next data point (y t+1 The steps are: to calculate the percentage change of ), to compare the previous percentage change with the lower limit value, and if the percentage change exceeds the lower limit value, the data point (y t A third round of outlier detection can be performed on a second dataset by statistical means including the step of identifying the previous data point (y) as an outlier. Alternatively, well-known statistical methods such as the Interquartile Range (IQR) method can be used to identify data points that fall outside the acceptable limits. t-1 ) data point (y t The steps include calculating the percentage change of the data point (y t ) Next data point (y t+1 The steps are: to calculate the percentage change of ), to compare the previous percentage change with the lower limit value, and if the percentage change exceeds the lower limit value, the data point (y t This can be performed on the dataset by a statistical method that includes the step of identifying outliers. Alternatively, well-known statistical methods such as the Interquartile Range (IQR) method can be used to identify data points that fall outside the acceptable limits. Once identified, outliers can be removed from the dataset. At the end of multiple outlier identification rounds, the dataset is expected to now contain virtually no outliers.

[0048] With respect to reverse osmosis and / or nanofiltration, the removal of physically inconsistent data points in the dataset may be performed by applying a set of criteria including one or more of the following: the feed flow must be greater than the concentrate flow; the feed flow must be greater than the permeate flow; the feed pressure must be greater than the concentrate pressure; the feed pressure must be greater than the permeate pressure; the feed pressure must be greater than the pressure drop across the filtration module; the permeate conductivity must be less than the feed conductivity; the pressure drop across the filtration module must be less than or equal to the difference between the feed pressure and the concentrate pressure; and / or the osmotic pressure of the feed flow must be less than the feed pressure.

[0049] In ultrafiltration and / or microfiltration systems, the removal of physically inconsistent data points in the dataset may be carried out by applying a set of criteria including one or more of the following: feed pressure is greater than or equal to the intermembrane pressure; filtrate pressure is greater than or equal to the intermembrane pressure; intermembrane pressure before and after the filtration module is less than or equal to the difference between the feed pressure and the filtrate pressure; and filtrate pressure and feed turbidity are greater than the filtrate turbidity.

[0050] Once outliers and physically inconsistent points have been removed, various membrane cleaning events (MCEs) can be detected through data analysis using methods appropriate for each technology. Such detection and removal of outliers can be performed automatically using the system's processing unit.

[0051] For reverse osmosis and / or nanofiltration, the dataset is expected to include normalized pressure drop (dP) before and after the filtration module, normalized permeate flow (Pf) from the filtration module, and normalized salt permeability (Sp) from the filtration module. Membrane washing events can be identified using a difference sequence method, where the data point (y t ) and the previous data point (y t-1The steps may include: calculating the difference between dP, Sp, and / or Pf to create the difference series; calculating the mean and standard deviation of the difference series for dP, Sp, and / or Pf; comparing each value of the difference series for dP, Sp, or Pf with the sum of the product of the difference constant and the corresponding standard deviation of the difference series and the corresponding difference series; checking whether there are gaps in the dataset; and checking whether the gaps are wider than the reference time for a membrane cleaning event (MCE). The difference constant is a positive integer, preferably a number between 1 and 10 (inclusive); more preferably a number between 3 and 7 (inclusive); and even more preferably 5. The MCE reference time is a time interval defined by the user as indicating the typical duration of an MCE in the system under consideration, and can range from a few minutes to several days. Exemplary values ​​for an MCE reference time may be, for example, 120 minutes, 1 day, etc. A preferred MCE reference time is 5 minutes to 2 weeks, with 1 day being a more preferred MCE reference time.

[0052] A similar approach can be taken for ultrafiltration and / or microfiltration. This dataset is expected to include transmembrane pressure (TMP), change in transmembrane pressure (ΔTMP or dTMP), feedwater temperature (T), filtrate / permeate flow rate (Pf), change in filtrate / permeate flow rate (ΔPf or dPf), and total membrane area (A) for the UF / MF system. Normalized membrane permeability can be calculated from filtrate / permeate flow rate, TMP, and total membrane area. During the filtration process, TMP is expected to increase, while membrane permeability is generally expected to decrease over time during a typical filtration cycle. Membrane washing events can be identified using a difference-sequence method, where the data point (y t ) and the previous data point (y t-1The steps may include: calculating the difference between ) to create the difference series for TMP; calculating the mean and standard deviation of the difference series for TMP; comparing each value of the difference series for TMP with the sum of the product of the difference constant and the corresponding standard deviation of the difference series and the corresponding difference series; checking whether there are gaps in the dataset; and checking whether the gaps are wider than the membrane cleaning event (MCE) reference time. The difference constant is a positive integer, preferably a number between 1 and 10 (inclusive); more preferably a number between 3 and 7 (inclusive); and even more preferably 5. The MCE reference time is a time interval defined by the user as indicating the typical duration of an MCE in the system under consideration, and can range from a few minutes to several days. Exemplary values ​​for the MCE reference time may be, for example, 20 seconds, 20 minutes, or 1 day. A preferred MCE reference time is 5 minutes to 2 weeks, with 1 day being a more preferred MCE reference time.

[0053] The UF or MF cycle includes a filtration step, a backwash (BW) step, a CEC step, and / or a CIP step. Membrane performance is more or less restored in the backwash, CEC, and / or CIP steps, which may result in a negative ΔTMP (or dTMP) for some interval. The membrane cleaning event (MCE) reference time is preferably defined by different operating step intervals, i.e., the data point (y) before the primary and / or secondary index. t-1 ), data point (y t ), and the next data point (y t+1 ) can be defined by different operation step intervals. For example, the previous data point (y t-1 ) is BW t-1 It could be the data point immediately following the data point (y t ) is BW t It could be the data point immediately following (y t+1 ) is BW t+1 It could be the data point immediately following (y t-1 ) is BW t It could be the data point immediately preceding the data point (y t ) is BW tIt could be the data point immediately following (y t+1 ) is BW t+1 It could be the data point immediately preceding it. In yet another example, the previous data point (y t-1 ) is CEC t-1 It could be a data point after the data point (y t ) is CEC t It could be a data point after the next data point (y t+1 ) is CEC t+1 It could be a data point after (y t-1 ) is CEC t It could be the data point immediately preceding the data point (y t ) is CEC t It could be the data point immediately following (y t+1 ) is CEC t+1 It could be the data point immediately preceding it. In yet another example, the previous data point (y t-1 ) is CIP t-1 It could be a data point after the data point (y t ) is CIP t It could be a data point after the next data point (y t+1 ) is CIP t+1 It could be a data point after (y t-1 ) is CIP t It could be the data point immediately preceding the data point (y t ) is CIP t It could be the data point immediately following (y t+1 ) is CIP t+1 It could be the data point immediately preceding it.

[0054] In addition, a membrane cleaning event (MCE) in a data segment can be identified when at least one of the following conditions is met: (a) feed flow is zero, (b) TMP is negative, and / or (c) filtrate flow is positive. If, in a particular normalized TMP (TMP_F) during filtration, there is a decrease of less than 30% in a period of less than 59 minutes accompanied by a linear TMP_F upward trend, the membrane cleaning event mode can be further classified as backwash (TMP_BW). These are the "backwash MCE criteria". Furthermore, if the membrane cleaning event mode is accompanied by a non-linear TMP_F upward trend, there is a decrease of more than 30% and less than 98% of TMP_F in a period of more than 1 day but less than 29 days, the membrane cleaning event mode can be further classified as chemical-assisted cleaning (TMP_CEC). These are the "CEC MCE criteria". Furthermore, membrane cleaning event modes can be further classified as stationary cleaning (TMP_CIP) if there is a non-linear upward trend in TMP_F and a decrease of more than 98% in TMP_F over a period of more than one month. These are the "CIP MCE criteria".

[0055] Once MCEs are identified, the data can be segmented between MCEs, and the type of fouling or scaling that occurred between consecutive MCEs can be determined. The method for determining the type of fouling or scaling depends on the process described below.

[0056] When determining the type of fouling occurring within a filter element, the system analyzes data from the filter element to determine whether various parameters are met. As described herein, some parameters are considered “primary” parameters, depending on the type of fouling involved, and some are considered “secondary” (e.g., arbitrary) parameters. For each type of fouling, there may be one or more primary parameters and one or more secondary parameters. In some cases, the data may satisfy criteria such as, for example, one primary parameter, two or more primary parameters, or one or more secondary parameters. Therefore, the system can first re-examine the data to determine whether one or more of the primary parameters are met. If so, the system can estimate the type of fouling based on the met primary parameters. In some embodiments, the system can estimate the type of fouling based on only one met primary parameter. In some embodiments, the system can rely on additional met primary and / or secondary parameters to solidify, strengthen, or support the original estimate of the type of fouling occurring based on the original met primary parameter. Therefore, the system can first determine whether the primary parameters are met in order to estimate the type of fouling. If the primary parameter is not met, the system can reconsider the secondary parameter and estimate the type of fouling based on the met secondary parameter. In some embodiments, if there is overlap among possible fouling types based on the met primary and / or secondary parameters, the system can perform a ranking operation to determine the likelihood of the occurring fouling type, for example, based on historical data. While some parameters have been discussed as being based on linear or quadratic polynomial functions, in some embodiments, parameters can be identified based on, for example, exponential functions, higher-order polynomial functions, linear functions, nonlinear functions, etc.The fitting of a set of curves to the aforementioned data segments can be performed using statistical means.

[0057] Embodiments of the present disclosure provide an exemplary method for determining a filter state. The method includes receiving historical performance data of one or more water filtration elements as input to a system for determining a filter state. The system for determining a filter state includes a database configured to electronically store the historical performance data and a processing unit that communicates with the database. The method includes identifying one or more data points in the historical performance data that indicate a cleaning event of one or more water filtration elements. The method includes dividing the historical performance data into individual data segments, each of which represents data in the historical performance data between each identified cleaning event. The method includes fitting a mathematical function to each of the individual data segments. The method includes identifying the type of fouling that occurs between each of the individual data segments based on the mathematical function fitted to each of the individual data segments.

[0058] Embodiments of the present disclosure provide an exemplary non-temporary computer-readable medium for storing instructions for determining a filter state that can be executed by a processing unit. Execution of an instruction by the processing unit causes the processing unit to receive historical performance data of one or more water filtration elements as input to a system for determining the filter state. The system for determining the filter state includes a database configured to electronically store the historical performance data and a processing unit that communicates with the database. Execution of an instruction by the processing unit causes the processing unit to identify one or more data points in the historical performance data that indicate a cleaning event for one or more water filtration elements. This can be used to divide the historical performance data into individual data segments, each of which represents data in the historical performance data between each identified cleaning event. Execution of an instruction by the processing unit causes the processing unit to fit a mathematical function to each of the individual data segments. Execution of an instruction by the processing unit causes the processing unit to identify the type of fouling occurring between each of the individual data segments based on the mathematical function fitted to each individual data segment.

[0059] Any combination and / or rearrangement of embodiments is envisioned. Other purposes and features will become apparent from the following detailed description, which will be considered in conjunction with the accompanying drawings. However, it should be understood that the drawings are designed for illustrative purposes only and not to define the limitations of this disclosure.

[0060] Refer to the attached drawings to assist those skilled in the art in the fabrication and use of a system for determining the filter state. [Brief explanation of the drawing]

[0061] [Figure 1] This is a schematic diagram of a traditional spiral-shaped filtration element. [Figure 2]This is a schematic diagram of a typical feed liquid spacer sheet of a traditional filter element, illustrating strand thinning and several geometric relationships. [Figure 3] This is a schematic diagram of a typical separation that occurs in a spiral-shaped reverse osmosis or nanofiltration element. [Figure 4] This is a block diagram of an exemplary system for determining the filter state according to this disclosure. [Figure 5] This is a block diagram of an exemplary computing device for implementing an exemplary system for determining the filter state according to this disclosure. [Figure 6] This is a block diagram of an exemplary system for determining the filter state environment according to this disclosure. [Figures 7A-7B] Figure 7A shows a graph of outlier identification performed by an exemplary system for determining the filter state in a reverse osmosis or nanofiltration system, with outliers included and Figure 7B showing outliers excluded. [Figures 8A-8C] These graphs show the time-dependent changes in the ideal (Figure 8A), expected (Figure 8B), and observed (Figure 8C) pressure drop (dP) profiles of a reverse osmosis or nanofiltration element in operation. [Figure 9] This graph shows membrane cleaning events (MCEs) in the dataset. [Figure 10] This graph shows the difference between unrecovered dP and initial dP for the dataset. [Figure 11] This graph shows the difference between cleanability and initial cleanability for a given dataset. [Figure 12A] Figure 12A shows the types of fouling and important factors, with Figures 12B and 12C showing the important factors based on normalized pressure drop, normalized permeate flow, and normalized salt permeability for NF and RO systems, and Figures 12B and 12C showing the important factors for UF and MF systems. [Figure 12B]Figure 12A shows the types of fouling and important factors, with Figures 12B and 12C showing the important factors based on normalized pressure drop, normalized permeate flow, and normalized salt permeability for NF and RO systems, and Figures 12B and 12C showing the important factors for UF and MF systems. [Figure 12C] Figure 12A shows the types of fouling and important factors, with Figures 12B and 12C showing the important factors based on normalized pressure drop, normalized permeate flow, and normalized salt permeability for NF and RO systems, and Figures 12B and 12C showing the important factors for UF and MF systems. [Figures 13A-13F] The graph shows the dataset analysis for pressure drop, permeate flow, and salt permeability, with specific results indicating particular fouling types, including no fouling (Figure 13A), biofouling (Figure 13B), organic fouling (Figure 13C), particulate fouling (Figure 13D), scaling (Figure 13E), and oxidation or integrity damage (Figure 13F). [Figure 14] This graph shows biofoulings identified based on the dataset. [Figure 15] This graph shows organic fouling identified based on the dataset. [Figure 16] This graph shows identified organic fouling and biofouling overlaying the organic fouling. [Figure 17] This graph shows identified organic fouling and biofouling overlaying the organic fouling. [Figure 18] This graph shows identified organic fouling, particulate contamination, biofouling, and MCE events. [Figure 19] This is a schematic diagram of a filtration element arranged in series with a water flow passing through the filtration element from the tip to the tail end. [Figure 20] This graph shows the MCE of the UF / MF system dataset, including BW, CEC, and the primary indicator. The data are shown on the same axis in Figure 20, with TMP_F data shown in Figure 20A, TMP_BW data in Figure 20B, and TMP_CEC data in Figure 20C. [Figures 20A-20C] This graph shows the MCE of the UF / MF system dataset, including BW, CEC, and the primary indicator. The data are shown on the same axis in Figure 20, with TMP_F data shown in Figure 20A, TMP_BW data in Figure 20B, and TMP_CEC data in Figure 20C. [Figures 21A-21B] These graphs show the time-dependent TMP_BW (Figure 21A), TMP_CEC (Figure 21B), and cleanability (Figure 21C) for each cycle of UF / MF systems identified as experiencing organic / biological fouling and particulate contamination. [Figure 21C] These graphs show the time-dependent TMP_BW (Figure 21A), TMP_CEC (Figure 21B), and cleanability (Figure 21C) for each cycle of UF / MF systems identified as experiencing organic / biological fouling and particulate contamination. [Modes for carrying out the invention]

[0062] The exemplary system for determining filter conditions discussed herein provides accurate monitoring, analysis, and identification of the type of fouling / scaling occurring within the filtration elements of a water treatment facility. The system can be used to determine the type of fouling / scaling occurring and can recommend to the operator when the system should be cleaned to ensure optimal operation. By determining the type of fouling / scaling occurring, the system can recommend the type of cleaning to be performed and the chemicals to be used, providing accurate guidance for effective cleaning, thereby ensuring improved operation of the water treatment facility after cleaning.

[0063] Figure 2 is a schematic diagram of a portion of the feed liquid spacer sheet 4 of the filter element, showing strand thinning and some geometric relationships. The feed liquid spacer sheet 4 is a nonwoven polymer net formed by intersecting a first set of substantially parallel filaments 34 with a second set of substantially parallel filaments 36 at angles 38, 40. The two sets of filaments 34, 36 are attached to each other at the intersection 42. The two intersecting sets of substantially parallel filaments 34, 36 form a two-dimensional array of similar parallelograms 44 (shown as dotted lines in Figure 2), the lengths of which define the mesh sizes 46, 48. The parallelograms have both acute angles 38 and obtuse angles 40, except when the two sets of filaments 34, 36 are perpendicular to each other. The acute angle 38 is bisected by a line 50 drawn substantially parallel to the flow direction 30. The angles at which the filaments 34, 36 cross the flow direction 30 are called their transverse angles 52, 54. The strands are spaced 46 and 48 apart, and each strand has a filament width of 62 and 64.

[0064] Figure 3 is a schematic diagram of the separation occurring in a helical reverse osmosis or nanofiltration element 70. The concentrated feed solution 72 can be passed between two or more membranes 74 such that the concentrate (containing waste or by-products 78) is guided in one direction and the diluted solution 76 (e.g., permeate flow or product) passes through the membrane 74 and exits the filtration system. Thus, the semipermeable membrane 74 allows water and small amounts of dissolved salt to pass through. The operational objectives of the water filtration system include maximizing the permeate flow (Pf), minimizing the salt transmission rate (Sp), and minimizing the pressure drop (dP). Feed pressure, feed temperature, feed flow, and feed water quality (dissolved solids) can affect the operation of the system.

[0065] The exemplary system can be used in ultrafiltration, nanofiltration, and / or reverse osmosis systems to identify types of filtration module failure modes (e.g., fouling, scaling, membrane failure, or combinations thereof). The system can help operators improve water treatment system performance by analyzing past performance in more detail. In some embodiments, the system can be used to predict when the next cleaning should be performed and recommend the optimal cleaning procedure.

[0066] The system conveniently does not require the installation of additional external devices dedicated to fouling type identification. Instead, the system determines the type of fouling that has occurred / is occurring and the optimal cleaning to address it, depending on the system's past performance and membrane cleaning events (MCEs) (sometimes referred to herein as clean-in-place (CIP) events). The system can be used with any type of water and does not require any special water treatment applications. Using this system, the time between cleanings (e.g., between MCE events) can be identified and these periods can be analyzed to evaluate the type of fouling and cleaning recommendations. Based on this data, the system can help predict when the next cleaning should be performed and what the most appropriate cleaning type will be. Thus, the system can operate in a substantially automated, independent manner, enabling similar types of operation in water treatment facilities. As discussed herein, the system can use the primary and secondary slopes of the evolution of pressure drop, permeate flow (net driving pressure), and water quality (conductivity or salt permeability), and the evolution of this data is normalized by temperature, pressure, flow, and / or water quality (dissolved solids).

[0067] Therefore, using the exemplary system, the type of fouling occurring within the filtration system and / or elements can be identified without the need to add any additional sensors or devices to the filtration system, simplifying overall use and reducing the operating costs of the system. The system first collects performance data associated with the filtration system, including, for example, differential pressure, permeate flow rate, and salt transmission rate. The system normalizes the data to starting conditions so that normalized pressure drop, normalized permeate flow rate, and normalized salt transmission rate are available and stored electronically in the system. Data normalization can be performed according to industry procedures. (See, for example, FilmTec® Reverse Osmosis Membranes Technical Manual, Water Solutions, DuPont, Table 27. Design equations for projecting RO system performance: Individual element performance, Form No. 45-D01504-en, Rev. 13, p. 110 (October 2022)). Temperature may be sensed and recorded to normalize the data. The system automatically removes contradictory or inconsistent data, such as a supply fluid flow rate less than the concentrated fluid flow rate, thereby eliminating statistical outliers.

[0068] Once the data is normalized and inconsistencies are removed, the system can identify membrane cleaning events (MCEs). Identifying cleaning events can be performed by creating a difference series of TMP, dP, and / or flow (the difference between data points at time points t and t-1), identifying the mean and standard deviation of the difference series, searching for data points where the difference value of TMP, dP is greater than or less than a K* standard deviation threshold, and confirming that there are gaps in the data that exceed a given threshold (e.g., 1 day). As used herein, the term “difference series” means the creation of a series, generally a time series Δx t =x t -x t-1 This can be explained by taking the difference between consecutive occurrences, Δx tIt has a constant mean and variance and can therefore be treated as a stationary series. (See, for example, Pal, A. et al., Practical Time Series Analysis: Master Time Series Data Processing, Visualization, and Modeling Using Python, United Kingdom: Packt Publishing (2017)). This approach can also identify plant shutdowns and restarts. In some cases, the system can be used to distinguish between MCE and plant shutdown-restart. In some embodiments, the absence of changes in dP, flow, and / or salt permeability before and after can be used by the system to indicate a plant shutdown without an MCE effect. In some embodiments, detection by the system of changes in temperature, pH, TMP, and / or flow rate can be associated with the MCE mode compared to plant shutdown / restart.

[0069] The system can segment data between CIPs and analyze each dataset between washes independently. Once a wash cycle is entered, and once a fouling type is identified, the system can use programmed, adapted equations to predict when the next wash will occur. After this next wash is identified, the optimal type of wash is suggested. For biofouling, organic fouling, and particulate fouling, conventional caustic washes may be recommended. For scaling, acid washes may be recommended. Particularly with respect to biofouling, the acceleration of biofouling can be estimated using the second derivative of the slope obtained in the previous and current periods (e.g., by analyzing how quickly the next MCE is needed). This can be calculated to more accurately predict when the next wash will occur due to biofouling, as the interval between washes becomes shorter once biofouling is present. If the system detects and identifies a combination of fouling types, a comprehensive wash consisting of caustic wash followed by acid wash may be recommended. The system can operate autonomously and provide advice on any identified fouling issues, so it can provide guidance to the user regarding corrective actions to be taken, and the user can decide whether the system should autonomously implement these corrective actions.

[0070] Therefore, the system performs data collection and analysis of the water treatment system's historical performance, normalizes the data, cleans the data from outliers to check data consistency, identifies membrane cleaning events, performs first- and second-order polynomial regression analysis and their coefficients of determination or Akaike Information Criterion (AIC), applies programmed logic to identify types of fouling / failure, uses previous information to predict the next type of fouling, and builds an adaptive system that includes fouling considerations. The system may include artificial intelligence and / or machine learning capabilities to improve the identification of fouling types and / or the predictability of fouling events. Input to the artificial intelligence and machine learning capabilities is obtained, preferably based on, feedback from the operator.

[0071] Figure 4 is a block diagram of an exemplary system 100 (hereinafter, "system 100") for determining the filter state. System 100 generally includes one or more water treatment facilities 102, each containing one or more filtration elements 104. The filtration elements 104 can be arranged back-to-back in series within a pressure vessel (for example, filtration elements 400-410 in Figure 19), allowing water to flow into the inlet of filtration element 400 at the front end, sequentially through each of the filtration elements 400-410, and out through filtration element 410 at the rear end. Generally, particles, biofouling, and / or integrity failure typically occur in the filtration elements at or near the front end, scaling and / or integrity failure occur in the filtration elements at or near the rear end, and organic fouling may occur in any of the filtration elements. In some embodiments, the pressure vessel can accommodate 6-8 filtration elements in series. The feedwater can be injected into the pressure vessel at the tip, with the filter element at the tip being the first to be exposed to the feedwater (e.g., the filter element at the first position). The filter element furthest from the pressure vessel inlet defines the filter element at the tail end. The data obtained regarding the filter elements within the pressure vessel may be filter element-specific, and by identifying and distinguishing the data from the tip and tail ends of the pressure vessel, it is possible to accurately estimate the type of fouling.

[0072] System 100 includes sensors 106 installed within the water treatment facility 102 to detect conditions associated with the filtration process, which can be used by System 100 to identify the type of fouling. In some embodiments, the sensors 106 can be used to detect, for example, the pressure drop before and after each filtration element 104, the permeate flow through each filtration element 104 (net driving pressure), the water quality (conductivity) to each filtration element 104, and the temperature of the water passing through the filtration elements 104. In some embodiments, the collected data may pertain to each of the filtration elements 104 to provide a more detailed diagnosis of System 100 (e.g., with at least one sensor 106 associated with each filtration element 104). As an example, smart sensors 106 can be incorporated into System 100 to determine whether the dP increase in the leading element (indicating particulate fouling) or the trailing element (indicating scaling) is more dominant. In some embodiments, the collected data may pertain to the entire filtration system (e.g., with one or more sensors 106 installed at specific locations in System 100). The sensor 106 communicates electronically with the central computing system 122 and / or processing unit 124 of system 100, enabling the collected data to be used to determine the type of fouling and the proposed cleaning.

[0073] System 100 includes one or more databases 108 that electronically store data associated with the operation of Facility 102 and System 100. Data may be transmitted electronically to and / or from the databases 108 via the communication interface 110 of System 100. Database 108 may include historical performance data 112 (e.g., historical data) including, for example, information received from Sensor 106 such as measured or detected conditions, cleaning schedules and activities, facility shutdown events, replacement of filtration elements, and changes in water supply sources. Database 108 may also include current performance data 114 including, for example, measured or detected conditions received from Sensor 106 in real time or substantially real time.

[0074] System 100 may include one or more users and / or one or more user devices 116 that communicate with System 100 via a communication interface 110. Users and / or user devices 116 may, for example, be operators of facility 102, or individuals responsible for scheduling cleaning events. Users 116 can electronically send and receive data to and from the system via a user interface 118, which may have a graphical user interface (GUI) 120 in some embodiments. The GUI 120 may be a display integrated into a user device 116 to enable users 116 to communicate with each other and / or with System 100 via the communication interface 110.

[0075] System 100 may include a central computing system 122 that communicates with each of the users 116 (for example, via the users' user devices) and one or more databases 108 associated with System 100 via a communication interface 110. The communication interface 110 is configured to provide a communication network between components of System 100, thereby enabling components of System 100 to electronically transmit and / or receive data. System 100 may include at least one processing unit 124 having a processor 126 for receiving and processing data stored in System 100.

[0076] During operation, system 100 can first receive historical performance data 112 as input to the normalization module 128, which is run by system 100 and can output normalized data 130 based on temperature or one or more other parameters, such as pressure, flow and / or water quality (dissolved solids). Next, the system runs the processing module 132 to detect and remove outliers from the normalized data 130 and outputs cleaned data 134 for further processing. The membrane cleaning event (MCE) module 136 is run by system 100 and can analyze the cleaned data 134 to identify and mark membrane cleaning events 138. This can be achieved by creating a difference series of TMP, dP, and / or permeate flow (the difference between data points at time t and t-1), identifying the mean and standard deviation of the difference series, searching for data points where the difference values ​​of TMP, dP, and / or flow are below or above a K* standard deviation threshold, and confirming that there is a gap in the data that exceeds a given threshold (e.g., 1 day). If a gap exceeding a given threshold exists in the data, system 100 can identify this point as a membrane cleaning event 138.

[0077] System 100 segments the data between each identified membrane cleaning event 138 to analyze each dataset independently between cleanings. In particular, it is understood that the baseline associated with the operation of Facility 102 and / or filtration element 104 may change as the filtration element 104 is used over a long period. For example, a new filtration element 104 will have a different performance baseline compared to a filtration element that has been in operation for six months and has been cleaned multiple times. As such, System 100 segments the data and analyzes the database independently between each cleaning to ensure accurate identification and prediction of fouling types.

[0078] System 100 runs the fouling identification module 140 on each data set associated with the operation between each membrane cleaning event 138 to determine the type of fouling that occurred during the operation of each data set. System 100 fits each data segment to a mathematical function, preferably a first- and second-order polynomial. System 100 analyzes the first- and second-order slopes of the normalized pressure drop and normalized permeate flow to normalize the salt permeability. If all slopes are close to zero, System 100 determines that the facility 102 should continue its normal operation (for example, no action is taken, or System 100 notifies via the graphical user interface 120 that no type of fouling was detected and cleaning is not required). For example, by reviewing past performance data 112, System 100 may determine that no fouling was detected at all, cleaning was unnecessary, and the facility 102 should have continued its normal operation. Upon reviewing the current performance data 114, it can be determined that since no fouling has occurred and cleaning is not required, facility 102 should be operating normally.

[0079] If the increase in pressure drop over time fits well to a quadratic polynomial, system 100 can conclude that biofouling was likely occurring. Biofouling alone (without interference from other types of fouling) is typically characterized by an initial flat normalized pressure drop (dP), followed by an increase in the quadratic polynomial to the pressure drop as biofouling begins to grow on the membrane. Biofouling is typically more pronounced at the leading element. Typically, the normalized permeate flow decreases over time as the net driving pressure decreases during the increase in dP (the quadratic polynomial as well). Salt blocking may remain stable or worsen (hence the increase in salt permeability) as a result of concentration polarization induced by the biofilm. For example, reviewing past performance data 112, system 100 can determine that biofouling occurred during this dataset and that corresponding washing should have been performed. Reviewing current performance data 114, system 100 can determine that biofouling is occurring and that appropriate washing is required.

[0080] In some embodiments, a filtration module failure mode in a data segment may be identified as a biofouling by the system 100 if at least one of the following conditions is met: (a) the coefficient of determination or Akaike Information Criterion (AIC) of dP for a linear polynomial is greater than 0.4; (b) the coefficient of determination or Akaike Information Criterion (AIC) of dP for a quadratic polynomial is greater than 0.7; (c) the difference between the coefficient of determination or Akaike Information Criterion (AIC) values ​​for the quadratic polynomial and the linear polynomial is greater than 0.3; (d) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a linear polynomial is greater than 0.4; (e) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a quadratic polynomial is greater than 0.4 (f) The coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial is greater than 0.3, (g) The coefficient of determination or Akaike Information Criterion (AIC) value for Sp for the linear polynomial is greater than 0.4, (h) The coefficient of determination or Akaike Information Criterion (AIC) value for Sp for the quadratic polynomial is greater than 0.7, (i) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial is greater than 0.3, (j) dP increases over time in the data segment, (k) Pf decreases over time in the data segment, and / or (l) Sp increases over time in the data segment. In some embodiments, biofouling can be identified if at least one of the dP conditions is met, and Pf and Sp provide secondary (optional) support for identification.

[0081] If the decrease in normalized permeate flow over time fits well to a quadratic polynomial, system 100 can be concluded to be likely experiencing organic fouling. Organic fouling alone (without interference from other types of fouling) is typically characterized by a decrease in normalized permeate flow due to the rapid accumulation of organic matter on the unused membrane, but after some operation, it levels off because the amount of organic matter accumulated on the membrane equals the amount of organic matter evaporating as a result of the cross-flow filtration mechanism. Typically, the normalized pressure drop remains flat, as in a pure organic fouling mechanism, and bacteria do not grow to block the feed-concentrate membrane channel. As the organic matter accumulated on the membrane creates "additional" thickness and resistance on the membrane, the salt permeability typically decreases, which typically improves over time. For example, reviewing past performance data 112, system 100 can be determined to have experienced organic fouling during this dataset and that corresponding washing should have been performed. Upon reviewing the current performance data 114, it can be determined that system 100 is experiencing organic fouling and requires proper cleaning.

[0082] In some embodiments, a filtration module failure mode in a data segment may be identified by system 100 as organic fouling if at least one of the following conditions is met: (a) the coefficient of determination or Akaike information criterion (AIC) of Pf for a linear polynomial is greater than 0.4; (b) the coefficient of determination or Akaike information criterion (AIC) of Pf for a quadratic polynomial is greater than 0.7; (c) the difference between the coefficient of determination or Akaike information criterion (AIC) values ​​for the quadratic polynomial and the linear polynomial fitted to Pf is greater than 0.3; (d) the coefficient of determination or Akaike information criterion (AIC) of Sp for a linear polynomial is greater than 0.4; (e) the coefficient of determination or Akaike information criterion (AIC) of Sp for a quadratic polynomial is greater than 0.7 (f) The coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic and linear polynomials that fit Sp is greater than 0.3, (g) The coefficient of determination or Akaike Information Criterion (AIC) for dP for the linear polynomial is greater than 0.7, (h) The coefficient of determination or Akaike Information Criterion (AIC) for dP for the quadratic polynomial is greater than 0.7, (i) The difference in the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic and linear polynomials that fit dP is less than 0.3, (j) dP remains constant over time in the data segment, (k) Pf decreases over time in the data segment, and / or (l) Sp decreases over time. In some embodiments, organic fouling can be identified if at least one of the Pf conditions is met, and dP and Sp provide secondary (optional) support for identification.

[0083] If the increase in normalized pressure drop fits well to a linear polynomial, it can be concluded that system 100 is experiencing either scaling or particulate fouling. Scaling is typically associated with a slower rate of increase in pressure drop compared to particulate fouling, as scaling precipitates typically take a little more time to crystallize and form. On the other hand, particulate fouling is typically a faster process, for example, when ultrafiltration is fouled and an increase in transmembrane pressure (TMP) is observed during the backwash cycle. Scaling typically occurs with or in conjunction with an increase in salt permeability over time, as scaling precipitated on the membrane increases concentration polarization on the membrane boundary layer. Scaling can also be associated with a decrease in normalized permeate flow over time, as osmotic pressure increases accordingly. Particulate fouling occurs when particulate matter clogs the membrane, which can result in a decrease in normalized permeate flow as the net driving pressure decreases over time. For example, by reviewing past performance data 112, system 100 can determine that scaling or particulate contamination occurred during this dataset and that corresponding cleaning should have been performed. By reviewing current performance data 114, system 100 can determine that scaling or particulate contamination has occurred and that appropriate cleaning is required.

[0084] In some embodiments, a filtration module failure mode in a data segment may be identified as scaling by system 100 if at least one of the following conditions is met: (a) the coefficient of determination or Akaike information criterion (AIC) of dP for a linear polynomial is greater than 0.7; (b) the coefficient of determination or Akaike information criterion (AIC) of dP for a quadratic polynomial is greater than 0.7; (c) the difference between the coefficient of determination or Akaike information criterion (AIC) values ​​of the quadratic polynomial and the linear polynomial that fit dP is less than 0.3; (d) the coefficient of determination or Akaike information criterion (AIC) of Sp for a linear polynomial is greater than 0.7; (e) the coefficient of determination or Akaike information criterion (AIC) of Sp for a quadratic polynomial is greater than 0.7. (f) The difference in the coefficient of determination or Akaike Information Criterion (AIC) value for a quadratic polynomial and a linear polynomial fitted to Sp is less than 0.3; (g) The coefficient of determination or Akaike Information Criterion (AIC) of Pf for a linear polynomial is greater than 0.7; (h) The coefficient of determination or Akaike Information Criterion (AIC) of Pf for a quadratic polynomial is greater than 0.7; (i) The difference in the coefficient of determination or Akaike Information Criterion (AIC) value for a quadratic polynomial and a linear polynomial fitted to Pf is less than 0.3; (j) dP increases preferentially in the tail element over time in the data segment; (k) Pf decreases over time in the data segment; and / or (l) Sp increases over time in the data segment. In some embodiments, scaling can be identified when at least one of the dP conditions is met and the dP condition distribution within the system is such that dP preferentially increases in the tail element, with Pf and Sp providing secondary (optional) assistance for identification.

[0085] In some embodiments, a filtration module failure mode in a data segment may be identified by system 100 as particulate contamination if at least one of the following conditions is met: (a) the coefficient of determination or Akaike Information Criterion (AIC) of dP for a linear polynomial is greater than 0.7; (b) the coefficient of determination or Akaike Information Criterion (AIC) of dP for a quadratic polynomial is greater than 0.7; (c) the difference between the coefficient of determination or Akaike Information Criterion (AIC) values ​​for the quadratic polynomial and the linear polynomial fitted to dP is less than 0.3; (d) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a linear polynomial is greater than 0.7; (e) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a quadratic polynomial is greater than 0.7; (f) Pf is fitted (g) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial is less than 0.3, (h) The coefficient of determination or Akaike Information Criterion (AIC) of Sp for the linear polynomial is greater than 0.7, (i) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for Sp for the quadratic polynomial is greater than 0.7, (j) dP increases rapidly over time, preferentially in the leading element of the data segment, (k) Pf decreases rapidly over time in the data segment, and / or (l) Sp increases rapidly over time, preferentially in the trailing element of the data segment. In some embodiments, particulate contamination can be identified when at least one of the dP conditions is met and the dP condition distribution within the system (preferably, dP increases at the leading element), with Pf and Sp providing secondary (optional) support for identification.

[0086] If the increase in normalized permeate flow fits a linear polynomial, it can be concluded that system 100 may experience potential problems with membrane integrity (such as chemical decomposition or halogenation of the membrane, or problems with the physical integrity of the membrane elements). Membrane integrity failures typically cause an increase in salt permeability over time. Membrane integrity failures can be classified as physical or chemical. In particular, problems with physical integrity cause a rapid increase in salt permeability over time, while chemical decomposition or halogenation of the membrane causes a gradual increase in salt permeability. For example, reviewing past performance data 112, it can be determined that system 100 experienced potential problems with membrane integrity and that appropriate measures should have been taken. Reviewing current performance data 114, it can be determined that system 100 may be experiencing problems with membrane integrity and that appropriate measures should be taken.

[0087] In some embodiments, a filtration module failure mode in a data segment may be identified by system 100 as chemical decomposition if at least one of the following conditions is met: (a) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a linear polynomial is greater than 0.7; (b) the coefficient of determination or Akaike Information Criterion (AIC) of Pf for a quadratic polynomial is greater than 0.7; (c) the difference between the coefficient of determination or Akaike Information Criterion (AIC) values ​​of the quadratic polynomial and the linear polynomial that fit Pf is less than 0.3; (d) the coefficient of determination or Akaike Information Criterion (AIC) of Sp for a linear polynomial is greater than 0.7; (e) the coefficient of determination or Akaike Information Criterion (AIC) of Sp for a quadratic polynomial is greater than 0.7 (f) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic and linear polynomials fitted to Sp is less than 0.3, (g) The coefficient of determination or Akaike Information Criterion (AIC) of dP for a monopolynomial is greater than 0.7, (h) The coefficient of determination or Akaike Information Criterion (AIC) of dP for a quadratic polynomial is greater than 0.7, (i) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic and linear polynomials fitted to dP is less than 0.3, (j) dP remains constant over time in the data segment, (k) Pf increases over time in the data segment, and / or (l) Sp increases over time in the data segment. In some embodiments, membrane integrity failures can be identified based on an increase in permeate flow accompanied by a simultaneous increase in salt permeability, and dP provides secondary (optional) support for identification.

[0088] In some embodiments, a filtration module failure mode in a data segment may be identified by system 100 as a physical integrity problem if at least one of the following conditions is met: (a) the coefficient of determination or Akaike information criterion (AIC) of Pf for a linear polynomial is greater than 0.7, (b) the coefficient of determination or Akaike information criterion (AIC) of Pf for a quadratic polynomial is greater than 0.7, (c) the difference between the coefficient of determination or Akaike information criterion (AIC) values ​​of the quadratic polynomial and the linear polynomial that fit Pf is less than 0.3, (d) the coefficient of determination or Akaike information criterion (AIC) of Sp for a linear polynomial is greater than 0.7, (e) the coefficient of determination or Akaike information criterion (AIC) of Sp for a quadratic polynomial is (f) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial fitted to Sp is less than 0.3, (g) The coefficient of determination or Akaike Information Criterion (AIC) of dP for the linear polynomial is greater than 0.7, (h) The coefficient of determination or Akaike Information Criterion (AIC) of dP for the quadratic polynomial is greater than 0.7, (i) The difference between the coefficient of determination or Akaike Information Criterion (AIC) value for the quadratic polynomial and the linear polynomial fitted to dP is less than 0.3, (j) dP remains constant over time in the data segment, (k) Pf increases rapidly over time in the data segment, and / or (l) Sp increases rapidly over time in the data segment. In some embodiments, membrane integrity failures can be identified based on a rapid increase in permeate flow accompanied by a rapid increase in salt permeability (preferably in the leading or trailing element), and dP provides secondary (optional) support for identification. If the system 100 concludes that the data indicates a particular type of fouling event, the system 100 can output a notification via the graphical user interface 120, along with the detected fouling event, supporting data, and recommendations for a cleaning strategy to address the fouling.If previous analyses detect that multiple fouling types are occurring simultaneously, such as organic fouling and subsequent biofouling within the same period or dataset, a combined cleaning strategy is required, and system 100 can provide recommendations on how to implement such a strategy. Thus, historical performance data can be used to identify different types of fouling and / or combinations of fouling that may occur based on the performance of facility 102, and the correlation of such fouling determinations is electronically stored as fouling type data 142.

[0089] In some embodiments, the current performance data 114 can be evaluated using fouling type data 142 and historical performance data 112 to detect one or more fouling types occurring in facility 102 and provide recommendations for cleaning operations. In some embodiments, the current performance data 114 can be evaluated using fouling type data 142 and historical performance data 112 to predict potential foulings that may occur. For example, system 100 can run a prediction module 144, which receives current performance data 114 as input, and predict the operational trajectory of facility 102 by estimating the types of fouling that may occur in the near future. In some embodiments, system 100 can run a prediction module 144, which receives historical performance data 112 as input, and estimate when a particular type of fouling may occur in facility 102, for example, a repeating pattern of a particular fouling type, such as the beginning of each summer or a specific month. In such cases, system 100 can provide user 116 with notification of potential fouling types that may occur within a specific time window, enabling user 116 to plan cleaning operations in advance. Therefore, the system 100 can provide accurate identification of the fouling type, recommend fouling type-specific cleaning procedures that focus on addressing the actual problem occurring at facility 102, ensure that the filter element 104 is properly cleaned, and extend the overall lifespan of the element 104.

[0090] Figure 5 is a block diagram of a computing device 200 according to an exemplary embodiment of the present disclosure. The computing device 200 includes one or more non-temporary computer-readable media for storing one or more computer-executable instructions or software for carrying out the exemplary embodiment. The non-temporary computer-readable media may include, but are not limited to, one or more types of hardware memory, non-temporary tangible media (e.g., one or more magnetic storage disks, one or more optical disks, one or more flash drives), etc. For example, memory 206 included in the computing device 200 may store computer-readable and computer-executable instructions or software for carrying out the exemplary embodiment of the present disclosure (e.g., instructions for operating a normalization module, instructions for operating a cleaning module, instructions for operating a film cleaning event (MCE) module, instructions for operating a fouling identification module, instructions for operating a prediction module, instructions for operating a processing unit, instructions for operating a communication interface, instructions for operating a user interface, instructions for operating a central computing system, and combinations thereof). The computing device 200 may also include a configurable and / or programmable processor 202 and associated core 204, as well as one or more additional configurable and / or programmable processors 202' and associated core 204', for executing computer-readable and computer-executable instructions or software stored in memory 206, as well as other programs for controlling system hardware (for example, in a computer system with multiple processors / cores). Processors 202 and 202' may be single-core processors or multi-core (204 and 204') processors, respectively.

[0091] Virtualization can be employed within the computing device 200 to enable the dynamic sharing of infrastructure and resources within the computing device 200. A virtual machine 214 may be provided to handle processes running on multiple processors, so that the process appears to be using only one computing resource rather than multiple computing resources. Multiple virtual machines may be used on a single processor. Memory 206 may include computer system memory or random access memory, such as DRAM, SRAM, EDO RAM, etc. Memory 206 may also include other types of memory, or combinations thereof.

[0092] A user may interact with the computing device 200 through a visual display device 218 (e.g., a personal computer, a mobile smart device, etc.), such as a computer monitor, and the computer monitor may display at least one user interface 220 (e.g., a graphical user interface) which may be provided according to an exemplary embodiment. The computing device 200 may include other I / O devices for receiving input from the user, such as a camera, keyboard, microphone, or any suitable multipoint touch interface 208, pointing device 210 (e.g., a mouse). The keyboard 208 and pointing device 210 may be coupled to the visual display device 218. The computing device 200 may include other suitable conventional I / O peripherals.

[0093] The computing device 200 may also include at least one storage device 224, such as a hard drive, CD-ROM, eMMC (MultiMediaCard), SD (Secure Digital) card, flash drive, non-volatile storage medium, or other computer-readable medium, for storing data and computer-readable instructions and / or software for implementing exemplary embodiments of the system described herein. The exemplary storage device 224 may also store at least one database 226 for storing any appropriate information necessary for implementing the exemplary embodiment. For example, the exemplary storage device 224 may store at least one database 226 for storing information such as historical performance data, current performance data, normalized data, cleaned data, membrane cleaning events, fouling type data, and combinations thereof, as well as computer-readable instructions and / or software for implementing the exemplary embodiment described herein. The database 226 may be updated manually or automatically at any appropriate time to add, delete, and / or update one or more items in the database.

[0094] The computing device 200 may include a network interface 212 configured to interface with one or more networks, such as a local area network (LAN), a wide area network (WAN), or the internet, via at least one network device 222, but not limited to a variety of connections including standard telephone lines, LAN or WAN links (e.g., 802.11, T1, T3, 56kb, X.25), broadband connections (e.g., ISDN, Frame Relay, ATM), wireless connections, controller area networks (CAN), or any combination of any or all of the above. The network interface 212 may include a built-in network adapter, a network interface card, a PCMCIA network card, a PaCI / PCIe network adapter, an SD adapter, a Bluetooth adapter, a CardBus network adapter, a wireless network adapter, a USB network adapter, a modem, or any other device suitable for interface the computing device 200 to any type of network it can communicate with and for performing the operations described herein. Furthermore, the computing device 200 may be any computer system, such as a workstation, desktop computer, server, laptop, handheld computer, tablet computer (e.g., tablet computer), mobile computing or communication device (e.g., smartphone communication device), embedded computing platform, or other form of computing or telecommunications device that is communicative and has sufficient processor power and memory capacity to perform the operations described herein.

[0095] The computing device 200 may run any operating system 216, for example, any version of the Microsoft® Windows® operating system, different releases of the Unix and Linux operating systems, any version of macOS® for Macintosh computers, any embedded operating system, any real-time operating system, any open-source operating system, any proprietary operating system, or any other operating system that can run on the computing device and perform the operations described herein. In exemplary embodiments, the operating system 216 may run in native mode or emulated mode. In exemplary embodiments, the operating system 216 may run on one or more cloud machine instances.

[0096] Figure 6 is a block diagram of an exemplary system environment 300 for determining filter conditions according to an exemplary embodiment of the present disclosure. The environment 300 may include servers 302, 304 configured to communicate with at least one water treatment facility 306, at least one sensor 308, at least one system 310, at least one processing unit 312, at least one user interface 314, and a central computing system 318 via a communication platform 324, the communication platform being any network capable of transmitting information between devices communicably coupled to the network. For example, the communication platform 324 could be the Internet, an intranet, a virtual private network (VPN), a wide area network (WAN), a local area network (LAN), etc. In some embodiments, the communication platform 324 may be part of a cloud environment.

[0097] Environment 300 may include repositories or databases 320, 322, which can communicate via a communication platform 324 with servers 302, 304, as well as a water treatment facility 306, a sensor 308, a system 310, at least one processing unit 312, at least one user interface 314, and a central computing system 318. In exemplary embodiments, servers 302, 304, the water treatment facility 306, the sensor 308, the system 310, at least one processing unit 312, at least one user interface 314, and the central computing system 318 can be implemented as computing devices (e.g., computing device 200). Those skilled in the art will recognize that databases 320, 322 can be incorporated into at least one of servers 302, 304. In some embodiments, databases 320, 322 can store data such as historical performance data, current performance data, normalized data, cleaned data, membrane cleaning events (MCEs), fouling type data, and combinations thereof, and such data can be distributed across multiple databases 320, 322.

[0098] As described above, the exemplary system collects historical performance data, normalizes the data, cleans the data to remove outliers and for data consistency, identifies MCEs, performs linear and quadratic polynomial regression and predictive models, e.g., the coefficient of determination or the Akaike Information Criterion (AIC), applies logic to identify the type of fouling / failure, and uses the previous information to predict the next type of fouling using an (optional) adaptive system that includes fouling considerations. In some embodiments, data smoothing can be performed after the data cleaning step. In some embodiments, three different methods can be used to remove outliers, e.g., removal of data points with abrupt changes, manual, and kernel smoothing, or a combination thereof. In some embodiments, the system can perform a cleanability calculation to determine when the cleaning procedure should be performed based on the data curve / direction of the current performance data.

[0099] Regarding the cleaning or outlier removal step, the system can implement a relative percentage change method, piecewise kernel smoothing, or a combination thereof. Consistency checks may include the following characteristics: feed pressure exceeds concentrate pressure, feed pressure exceeds permeate pressure, feed flow exceeds concentrate flow, feed flow exceeds permeate flow, and feed pressure exceeds pressure change (dP). Figures 7A and 7B are graphs of outlier identification performed by systems using the methods discussed, with Figure 7A including outliers and Figure 7B excluding outliers. In this method, data outliers commonly occur due to process upsets, measurement errors, etc. Outliers result in excessive "false positive" MCEs, and Equation 1 for identifying / indicating process stoppages below governs one type or relative percentage change analysis.

number

[0100] In some embodiments, outlier identification can be performed by the system based on piecewise kernel smoothing. In such embodiments, the entire data is divided into different segments, for example, segments. If the time gap between two segments exceeds one day, for each segment, a kernel smoother (e.g., a locally weighted scatterplot smoother, i.e., LOWESS) is deployed. A kernel smoother generates a curve formed by repeatedly finding a locally weighted fit of a simple curve at sampled points within the domain. The default settings of the system are linear (lambda), tri-cube (weight function), 0.5 (alpha), and 0 (sampling delta). Depending on the data, the system can automatically identify the optimal alpha value. In some embodiments, JMP (trademark) kernel smoothing can be used.

[0101] After the dataset has been cleaned, system 100 can review and analyze the data to identify membrane cleaning events (MCEs). System 100 can segment the data around an MCE to identify the start and end points of the MCE for further processing to identify the fouling types that occur between MCEs. Generally, an MCE is an effort to "reset" or improve the water treatment process. An MCE also represents wasted production time and chemical cost expenditures. System 100 may rely on more than one day of missing data before and after changes in several factors indicating an MCE, such as changes in KPIs (permeate flow (Pf), salt permeability (Sp), and element pressure drop (dP), combinations thereof, etc.). Some potential issues for identifying MCEs can include, for example, outliers in the data (if not properly removed), inconsistent time intervals between MCEs, the fact that KPIs do not necessarily change significantly due to an MCE, the use of process stops as "low-cost" MCEs, or combinations thereof. Generally, the processes described with respect to RO data in FIGS. 7A and 7B identify and remove outliers in the data. If the KPIs do not change significantly, such data can indicate (instead of an MCE) a process stop, which can help the system to identify. If a process stop functions as a "low-cost" cleaning, the system can consider such a "low-cost" cleaning as an MCE.

[0102] The MCE identification process is based on the difference series (∇y t-1 ) of the data point (y t ) from the previous data point (y t , shown as; ∇y t = y t - y t-1This can be done by creating (y t ) is identified as the restart value after MCE. This method is based on TMP, dP, and / or Extreme Values ​​in Differencing Series (EVDS) of the flow difference series. This method checks whether the change in the data point deviates far from the expected variation. dP and Pf may be the optimal KPIs for identifying the MCE point. The default value of k is 5, but it can be changed by the user. The MCE identification process can be expressed by Equation 2 below. If(∇dP t ≤ Column mean (∇dP) - K· Standard deviation of a column (∇dP) or (∇Flow t ≥ Column mean (∇Flow) + K· Column standard deviation (∇Flow) and (∇Date t >1) (2) In this case, the current observation corresponds to the restart time after MCE. The corresponding time in the previous observation indicates the MCE time.

[0103] Figures 8A–8C are graphs showing the ideal (Figure 8A), expected (Figure 8B), and observed (Figure 8C) pressure drop (dP) in RO or NF processes. The purpose of MCE (here, CIP) is to restore the system performance to its initial state (ideally). In the ideal scenario in Figure 8A, dP returns to its initial value after cleaning. In the expected scenario in Figure 8B, dP decreases with each cleaning but steadily increases. In the observed scenario in Figure 8C, dP changes inconsistently. dP is generally used to check the cleaning effect (other KPIs may be used in other embodiments). Indices have been developed over the years to measure cleaning effect, such as static indices (uncoated dP, cleanability, etc.) compared to dP after commissioning, and dynamic indices (dynamic cleanability) compared to dP after the previous CIP.

[0104] In some embodiments, the system can identify the MCE start and end points, calculate the average of three dP values ​​at the start and end of each MCE, and calculate the associated cleansing effectiveness index for each MCE. A similar process can be used with other KPIs, either in addition to or instead of dP. Other cleansing indices can be calculated. The number of data points used to calculate the average can be modified by the system user or operator. For example, in some embodiments, the average can be taken over a period of at least two hours to ensure stable readings. However, in some embodiments, an average taken over more than or less than two hours can be used. Figure 9 is a graph showing the MCEs of a dataset, Figure 10 is a graph showing unrecovered dP vs. initial for the dataset, and Figure 11 is a graph showing cleansing effectiveness vs. initial for the dataset. Figure 10 shows the difference between dP measured after the MCE and the initial dP (e.g., dP measured at startup even before the first MCE is performed). In a complete system as described in Figure 8A, Figure 10 is a horizontal line at 0 (i.e., no unrecovered dP). The phrase “against initial” at the top of Figure 10 indicates that the plots are relative to the initial dP measured at startup before the first MCE is performed, compared to other dP values ​​after the first, second, third, etc. MCE. Similarly, with respect to Figure 11, cleanliness can be defined as the percentage change in the primary index (for RO / NF, such primary index may be dP, Pf, or Sp; for UF / MF, such primary index may be normalized TMP) before and after the MCE, compared to a baseline condition. Such cleanliness captures the extent to which the dP increase was removed by the MCE, compared to the ideal case where all dP increases from startup are removed by the MCE. In a perfect system as described in Figure 8A, Figure 11 is a horizontal line of straight lines at 100% (i.e., the elements are cleaned to startup conditions each time). In Figure 9, a i represents the dP value after the i-th MCE, and b i This represents the dP value before the i-th MCE. These values ​​are given by the following equations 3 and 4:

number

number

[0105] In some cases, MCE in RO or NF processes can be identified by detecting a lower pressure drop, an increase in normalized flux, or a salt permeability restored to its starting point. When the system starts / stops, the permeate flow and salt permeability oscillate until they stabilize. Once MCE is identified, the system can analyze the data between MCEs to determine the type of fouling occurring based on the system's operating values. Figure 12A is a chart of indices / parameters based on fouling types in RO / NF processes, as well as normalized pressure drop, normalized permeate flow, and normalized salt permeability. The information in Figure 12A, combined with the information in Figure 13, can be used by the system to identify a specific type of fouling occurring. For some types of fouling, the indices / parameters are re-examined based on the location of the filter element, e.g., tip, tail, or both. Since some types of fouling can occur in any filter element, the indices / parameters are re-examined at any location on the filter element. Normalized pressure drop can serve as a primary indicator of fouling-free, biofouling, particulate contamination, and scaling. Normalized permeate flow can serve as a primary indicator of fouling-free, organic fouling, particulate contamination, and mechanical element damage (e.g., chemical decomposition, physical integrity failure). Normalized salt permeate can serve as a primary indicator of scaling and element damage. Therefore, these important factors can be re-examined by the system for each dataset corresponding to the inter-MCE period to determine the type of fouling occurring.

[0106] In UF or MF processes, normalized membrane pressure (TMP) can serve as a primary indicator of no fouling, biofouling, particulate contamination, scaling, and no scaling. As shown in Figures 12B and 12C, other fouling types and indicators / parameters include flux and permeability. Different types of fouling may be present depending on the feedwater. The type of fouling can be identified by 1) analysis of the feedwater, or 2) membrane recovery after different types of washing. For example, if acidic chemical washing works well in restoring membrane TMP or flux, it is mainly inorganic fouling or scaling; if acidic chemical washing does not work well, but alkaline and / or oxidizing chemical washing works well in restoring membrane TMP or flux, it is mainly biofouling or organic fouling. Irreversible fouling is generally present after each washing (including air scrubbing, backwashing, and chemical washing) because partial pores or internal pore channels in UF or MF are blocked by fouling. Irreversible fouling is characterized by a decrease in cleanability over time, during successive MCE (e.g., backwashing steps), or an increase in TMP after each MCE.

[0107] As described herein, the system can first determine whether there is a match based on at least one primary indicator / parameter for identifying the type of fouling. Such identification may be reinforced by additional primary and / or secondary indicators / parameters that match the dataset. However, the system only requires a match on a single primary indicator / parameter for fouling type estimation. In some cases, two or more types of fouling may occur, and the system can identify such a scenario, for example, based on two or more different primary parameters that are met. In some cases, if no match on a primary indicator / parameter is identified, the system can determine the type of fouling based on a match on a secondary parameter.

[0108] In some embodiments, the “decision tree” performed by the system may include the following steps: (i) identifying the slope of a primary parameter, (ii) identifying the fit tendency of the primary parameter, and (iii) determining the position of the filter element for the identified performance change. Based on the match of at least one primary parameter with slope versus time (or slope versus time and data fit; or slope versus time, data fit, and position of the filter element), a fouling type can be identified. If no match is found for at least one primary parameter, the system's “decision tree” may identify the slope and fit tendency of a secondary parameter to identify a fouling type based on the match of the secondary parameter. If the match of the primary and / or secondary parameter identifies two or more potential fouling types, the system may provide the user with potential matches, and the user can review the data to determine which of the potential matches is the actual fouling type in which the fouling is occurring. In some cases, two or more types of fouling may occur within the filter element, and the system may identify multiple fouling types based on the match of the primary and / or secondary parameter. Identified foulings can be used to identify and analyze current performance data (e.g., in real time or substantially real time) to estimate the types of foulings that may be occurring and / or are likely to occur soon, and to recommend more specific and accurate cleaning procedures to address the foulings. Identified foulings can also be used to predict when foulings may occur, based on the detection of recurring fouling trends occurring during specific weeks or months, and / or taking into account specific water conditions (e.g., temperature).

[0109] In particular, focusing on Figure 12A, which is specific to RO systems and relates to fouling-free systems, data can be used from any filtration element position, and dP, Pf, and Sp are each considered primary indicators / parameters. With respect to the fouling-free systems to be identified, any of the dP, Pf, and / or Sp curves should fit a linear function with a slope of approximately zero versus time (e.g., less than 5% change over the entire operating period of the dataset).

[0110] Continuing to refer to Figure 12A, for biofouling, the data are reconsidered with respect to a tip filter element which may include only a first filter element (or, in some embodiments, may include first and second filter elements at the tip of the pressure vessel). dP is a first-order parameter of biofouling, where the data fits a nonlinear function and the slope-to-time is positive (both of these must occur to satisfy the first-order parameter). Pf is a second-order parameter of biofouling, where the data fits a nonlinear function and the slope-to-time is negative (both of these must occur to satisfy the second-order parameter). Sp is another second-order parameter of biofouling, where the data fits a nonlinear function and the slope-to-time is positive (both of these must occur to satisfy the second-order parameter). Thus, in any case discussed herein, both function fit and slope-to-time must be satisfied to satisfy the first-order or second-order parameter. If the first-order parameter condition is satisfied, the system can identify the fouling type as biofouling even if the second-order parameter is not satisfied. However, if one or more of the quadratic parameters are met, the system can use this data to enhance the original identification based on the primary parameters.

[0111] For organic fouling in RO / NF systems, the data is re-examined with respect to any of the filter elements (e.g., not limited to the tip or tail filter elements). Pf is a first-order parameter of organic fouling, where the data fits a nonlinear function and the slope-to-time is negative. dP is a second-order parameter of organic fouling, where the data fits a linear function and the slope-to-time is approximately zero. Sp is another second-order parameter of organic fouling, where the data fits a nonlinear function and the slope-to-time is negative. If the first-order parameter condition is met, the system can identify the fouling type as organic fouling even if the second-order parameters are not met. However, if one or more of the second-order parameters are met, the system can use this data to enhance the original identification based on the first-order parameter.

[0112] For particulate fouling in RO / NF systems, the data is re-examined with respect to the tip filtration element. dP is a primary parameter of particulate fouling, where the data fits a linear function and the slope-to-time is positive. Pf is a quadratic parameter of particulate fouling, where the data fits a linear function and the slope-to-time is negative. Sp is another quadratic parameter of particulate fouling, where the data fits a linear function and the slope-to-time is positive. If the primary parameter conditions are met, the system can identify the fouling type as particulate fouling even if the quadratic parameters are not met. However, if one or more of the quadratic parameters are met, the system can use this data to enhance the original identification based on the primary parameter.

[0113] For scaling in RO / NF systems, data is reconsidered with respect to a tail end filter element that may include a final filter element at the tail end of the pressure vessel (or, in some embodiments, the final and second-to-last filter elements). dP is a primary parameter of scaling, where the data fits a linear function and the slope-to-time is positive. Pf is a quadratic parameter of scaling, where the data fits a linear function and the slope-to-time is negative. Sp is another quadratic parameter of scaling, where the data fits a linear function and the slope-to-time is positive. If the primary parameter condition is met, the system can identify the fouling type as scaling, even if the quadratic parameters are not met. However, if one or more of the quadratic parameters are met, the system can use this data to enhance the original identification based on the primary parameter.

[0114] For completeness failures in RO / NF systems, data is re-examined with respect to both the front and rear filtration elements, which may include only the first, i.e., first and last filtration elements at the front and rear of the pressure vessel (or, in some embodiments, two filtration elements at the front and two at the rear). Pf is one primary parameter of completeness failure, where the data fits a linear function and the slope versus time is positive. Sp is another primary parameter of completeness failure, where the data fits a linear function and the slope versus time is positive. dP is a quadratic parameter of completeness failure, where the data fits a linear function and the slope versus time is approximately zero. If the primary parameter condition is met for only one primary parameter, the system can identify the fouling type as completeness failure even if both the primary and / or quadratic parameters are not met. However, if both primary and / or quadratic parameters are met, the system can use this data to enhance the original identification based on the primary parameter.

[0115] Referring to Figures 12B and 12C, in relation to UF or MF systems, the term "normalized TMP" describes the force required to maintain a constant flux in the system, taking temperature effects into account. Another similar term, "normalized permeability," can also be used, calculated by dividing the actual normalized flux by the TMP. Normalized resistance is related to the reciprocal of normalized permeability. In general, since UF or MF systems operate in a constant flow or flux mode, normalized TMP can be used, and normalized TMP functions similarly to normalized permeability or normalized resistance when characterizing the film filter state and fouling. For simplicity, normalized TMP will be used in this description.

[0116] When discussing UF or MF systems, the backwash (BW) cleaning cycle may include multiple steps, such as upper BW and bottom BW, depending on the direction of flow through the membrane module relative to the train. This can also be integrated with other cleaning enhancement operations such as AS (Air Scrubbing), FF (Forward Flash), or combinations thereof.

[0117] When discussing UF or MF systems, chemically charged cleaning (CEC) refers to periodic cleaning with chemicals added to the cleaning solution. The introduction of cleaning agents into the membrane module can be carried out through different flows. For example, cleaning agents can be introduced through backwash flows (referring to CEB - chemically charged backwash), feed flows, etc. In addition, additional tanks can be introduced into the UF system to facilitate cleaning. For example, the cleaning solution can be recirculated between the membrane rack / skid and the tank. If necessary, the concentration of the cleaning solution, the pH, value, and type of cleaning agent, as well as the cleaning temperature, can be adjusted depending on the type of water and the type of membrane fouling.

[0118] When discussing UF or MF systems, clean-in-place (CIP) generally refers to chemical cleaning with higher chemical concentrations, longer immersion and recirculation times, higher cross-flow, and potentially higher cleaning solution temperatures compared to CEC. For UF or MF systems, cleaning agents typically introduced into membrane modules during a membrane cleaning event may include, for example, sodium hydroxide, sodium hypochlorite, hydrochloric acid, oxalic acid, citric acid, or combinations of two or more of these. Other chemicals typically introduced into the feed solution during filtration, such as coagulants, are not considered cleaning agents for the purposes of this disclosure.

[0119] Figure 12B shows the type of washing cycle and its effect on the primary indicator of the UF or MF system. For filtration, TMP_F has a linear fit in the range of approximately 0, with a duration of less than 60 minutes and no washing agent used. For backwashing, the TMP_F fit is linear with a positive slope, with a washing performance of less than 30%, a duration of less than 300 seconds, a frequency of less than every 59 minutes, and a low concentration of washing agent (e.g., oxidizing agent or other washing agent in the range of 1-10 ppm). For chemical washing, the TMP_F fit is nonlinear with a positive slope, with a washing performance of 30-98%, a duration of 1-60 minutes, a frequency of 1-29 days, and a medium concentration of washing agent (e.g., oxidizing agent or other washing agent in the range of 10-100 ppm). For stationary cleaning, TMP_F compliance is a positive-slope nonlinearity, with cleaning efficiency exceeding 98%, a duration exceeding 1 hour, a frequency exceeding 1 month, and high concentrations of cleaning agents (e.g., oxidizing agents or other cleaning agents in the range of 100-1000 ppm, depending on the membrane material and the specifications of the UF / MF supplier). Typically, stationary cleaning also uses temperatures higher than the ambient temperature.

[0120] Due to the complexity of the feedwater to the UF or MF, the slope of the primary data is classified into different types of feedwater in the historical data. The historical data includes at least several operating cycles, each operating cycle including a filtration process and a washing process (including one or more steps from the group of air scrubbing, backwashing, flow reversal, chemical washing, chemical backwashing, and stationary chemical washing). The changes in TMP and its pattern in each cycle reflect the state of the membrane under the specific operating process, including more detailed information such as the duration and frequency of each step. Figures 20, 20A, 20B, and 20C show an example of normalized TMP over several days including several filtration and backwash (F / BW) cycles and three CEC membrane washing events.

[0121] In practice, identifying the membrane state of UF / MF within a CIP cycle is of greater interest. It provides feedback on whether the current membrane and operating procedures are suitable for the current feedwater, thus meeting treatment targets and fouling control levels. Similar to RO / NF, there are parameters that can be used to identify membrane state and fouling. The primary indicators used in UF / MF systems are the normalized TMP at a specific point in time within each cycle, including TMP_F (TMP during filtration), TMP_BW (initial TMP at the start of the filtration cycle after BW), and TMP_CEC (initial TMP at the start of the filtration cycle after CEC). Figures 20, 20A, 20B, and 20C also show some examples of TMP_F, TMP_BW, and TMP_CEC in a typical UF / MF TMP profile. In addition to these indicators, the washability percentage of each MCE can also be used to assist in identifying the type of fouling in UF / MF. The water quality of the permeate, most often turbidity, can be used as a secondary indicator in addition to the primary indicators mentioned above to help identify other membrane problems such as chemical degradation or physical integrity. Washability is defined as the difference between the normalized TMP before MCE and the initial value, in particular as the change in normalized TMP before and after MCE.

[0122] Fouling in UF processes is understood as a phenomenon caused by the deposition or adsorption of contaminants present in the feedwater flow onto the membrane surface or into the internal structure "within the pores." Membrane fouling can occur and be caused by insufficient pretreatment, excessive coagulation or changes in upstream processes, and the slow accumulation of biological, organic, or inorganic contaminants or precipitates in the feedwater. In particular, with respect to membrane fouling, in ideal conditions treating water completely free of contaminants, the membrane will not foul at all. However, in reality, UF / MF systems will almost always experience various types of fouling. Fouling occurs when the membrane blocks and removes contaminants during the filtration step. Subsequent MCE steps, such as BW or CEC, will generally remove the fouling to varying degrees, depending on the type and portion of the foulant. Over time, the accumulated net fouling will lead to higher TMP, as it increases filtration resistance in addition to intrinsic membrane resistance.

[0123] Generally, there are different types of fouling, including particulate or cake-forming fouling, biofouling or organic fouling, and inorganic fouling, i.e., scaling. The first type, particulate fouling, can be called rapid removal fouling or reversible fouling, as the majority of this type of accumulated fouling can be removed during the subsequent BW step. The actual contaminants of this type of fouling can be parts, suspended matter, and / or colloids, all of which can be partially characterized by turbidity. Biological fouling is caused by the adhesion and growth of microorganisms on the membrane, which can also lead to the formation of viscous biofilms. Organic fouling is one of the main causes of fouling in the UF process and is caused by organic matter (silt, organic acids, humic substances) adsorbed onto the membrane. Inorganic fouling, or scaling, is caused by the precipitation of inorganic matter on the membrane. Biofouling, organic fouling, and inorganic fouling can also be called slow removal or irreversible fouling.

[0124] Referring to Figures 12B and 12C, which are specific to UF or MF systems, the combined primary indicators provide insights into membrane fouling and the effectiveness of cleaning to maintain stable operation of the membrane system. For example, in an ideal case with no fouling, such as when treating ideal water completely free of contaminants, TMP_F is constant (i.e., linear) over time and represents the initial state of the membrane, and the same can be expected for TMP_BW and TMP_CEC. In such a case, the cleaning efficiency is approximately 100%, and the filtrate turbidity is less than 1 NTU.

[0125] For particulate or cake-forming type fouling, TMP_F increases linearly during the filtration cycle. For the TMP_BW data series, it can be linearly fitted, with a statistically low slope, generally close to 0. The TMP_CEC data series has similar characteristics to the TMP_BW data series. Washability can range from approximately 40% or 60% to approximately 95%, and turbidity is less than 1 NTU.

[0126] In the case of organic / biological fouling, TMP_F can increase linearly or non-linearly during the filtration cycle. However, unlike reversible fouling, statistically, TMP_BW either grows linearly over time with a relatively large positive slope greater than the slope of a particular TMP, or grows non-linearly like a quadratic polynomial or other exponential function. The slope of a particular TMP depends on the membrane fouling and washing control targets. For illustrative purposes, the slope of a particular TMP could be ~0.033 bar / day for TMP_BW in a UF / MF system, with the total duration of filtration and BW accounting for >90% of the time in a day. This specification aims for a particular normalized TMP to increase by only 1 bar per month before the TMP exceeds the recommended general operating limit and each CIP is performed. TMP_CEC can increase linearly or non-linearly with a positive slope. A specific slope of TMP_CEC can be ~0.033 bar / day, for example, TMP_BW. Washability is generally less than 60%, and turbidity is less than 1 NTU. Inorganic fouling or scaling may have similar indicators to organic / biofouling. However, while oxidizing agent / caustic agent CEC may be used to identify organic / biological fouling, acidic CEC and their respective TMP_CEC may be considered for inorganic fouling or scaling. Figures 21A-21C are graphs showing TMP_BW, TMP_CEC, and washability over time or over their respective cycles for specific UF / MF systems identified as suffering from organic / biological fouling and particulate fouling. For chemical decomposition failures, a positive gradual change is used as an indicator of filtrate turbidity. For physical integrity failures, a positive rapid change is used as an indicator of filtrate turbidity.

[0127] Figures 13A–13F are graphs showing dataset analysis for pressure drop, permeate flow, and salt permeability, with specific results indicating specific fouling types in RO / NF systems, including no fouling (Figure 13A), biofouling (Figure 13B), organic (Figure 13C), particulate (Figure 13D), scaling (Figure 13E), and integrity (Figure 13F). For Figures 13A–13F and Figures 21A, 21B, and 21C, and throughout this disclosure, the terms “rapid,” “step,” or “gradual” change refer to a change of at least 5% of a given parameter over an operating period of less than 24 hours; the terms “abrupt” or “gradual” change refer to a change of at least 5% of a given parameter over an operating period of 24 hours or more; and the terms “zero” or “no change” refer to a change of less than 5% over the entire operating period of the dataset.

[0128] As shown in Figure 13A, no fouling is generally represented by no change in any factor. In these results, washing is unnecessary. As shown in Figure 13B, biofouling is generally represented by an increase in pressure drop that fits a quadratic polynomial (with an arbitrary decrease in permeate flow and an increase in salt permeability). Washing with a specific compound for biofouling would be recommended. As shown in Figure 13C, organic fouling is generally represented by a decrease in permeate flow that fits a quadratic polynomial (with an arbitrary decrease in salt permeability and a constant pressure drop). Washing with a specific compound for organic fouling would be recommended.

[0129] As shown in Figure 13D, particulate fouling is generally represented by a rapid increase in pressure drop (e.g., a stepwise change) that fits a linear polynomial (accompanied by any abrupt decrease in permeate flow (e.g., a stepwise change) and a rapid increase in salt permeability (e.g., a stepwise change) that fits a linear polynomial). Washing with specific compounds for particulate fouling would be recommended. Troubleshooting pretreatment would be recommended. The increase may be linear, but generally such an increase is represented as a ramp. As shown in Figure 13E, scaling is generally represented by an increase in pressure drop (accompanied by any decrease in permeate flow and an increase in salt permeability that fits a linear polynomial). Washing with specific compounds for scaling would be recommended. Reducing overall system recovery would be recommended. Biofouling and scaling can have similar effects, but with respect to biofouling, the increase in dP is more dominant in the tip element, while this effect is observed in the tail element with respect to scaling. As shown in Figure 13F, oxidation or integrity damage is generally represented by an increase in permeate flow conforming to a linear polynomial and an increase in salt permeability conforming to a linear polynomial (with any constant pressure drop). Replacing the current filter element would be recommended.

[0130] Figure 14 is a graph showing biofoulings identified based on a dataset, and Figure 15 is a graph showing organic foulings identified based on a different dataset. The identification of these foulings was performed by the system by re-examining and analyzing the data to determine key factor conditions and aligning these key factor conditions with the guidelines described above. Based on these guidelines, the system developed specific features of the datasets and identified specific types of foulings occurring. The distinction of organic foulings was reproduced with excellent results based on experiments with various datasets.

[0131] Figures 16 and 17 are graphs showing identified organic fouling and biofouling overlaid on organic fouling in an RO system. Excellent identification and reproducibility of results were achieved using the system during the experiment. Figure 18 is a graph showing identified organic fouling, particulate fouling, biofouling, and CIP events in an RO system based on different datasets.

[0132] In some cases, the system may follow specific rules for analyzing data and detecting fouling types. The system can first detect cycles between MCEs and, for each cycle, perform linear and quadratic polynomial regressions on dP, salt permeability, and permeate flow. The system can be used to detect CIP events and fouling types in both RO (single-pass) and CCRO (closed-circuit multi-pass) systems.

[0133] While exemplary embodiments have been described herein, it should be explicitly noted that these embodiments should not be construed as limiting, and rather additions and modifications to those expressly described herein are also included within the scope of the invention. Furthermore, it should be understood that the features of the various embodiments described herein are not mutually exclusive, and various combinations or rearrangements may exist, even if not expressed herein, without departing from the spirit and scope of the invention.

Claims

1. A system for determining the filter state, A database configured to electronically store data including past performance data of one or more water filtration elements, A processing device that communicates with the aforementioned database, The aforementioned past performance data is received as input, Identify one or more data points in the past performance data that indicate a cleaning event of one or more of the water filtration elements, The aforementioned past performance data is divided into individual data segments, and each of the individual data segments represents the data of the past performance data between each identified cleaning event. A mathematical function is fitted to each of the aforementioned data segments, and Based on the mathematical function adapted to each of the individual data segments, the type of fouling occurring between each of the individual data segments is identified. A processing apparatus configured as follows A system that includes this.

2. The system according to claim 1, wherein the processing apparatus is further configured to normalize the past performance data to the startup conditions of one or more water filtration elements before identifying the cleaning event.

3. The system according to claim 1 or 2, wherein the processing apparatus is further configured to take into account fluctuations in at least one of the feedwater temperature, feedwater composition, or feedwater pressure when normalizing the past performance data.

4. The system according to any one of claims 1 to 3, wherein the processing apparatus is further configured to identify and remove physically inconsistent data points from past performance data before identifying the cleaning event.

5. The system is a reverse osmosis (RO) or nanofiltration (NF) system, and the physically contradictory data points are under the following conditions: The supply liquid flow is greater than the concentrated liquid flow. The supply liquid flow is greater than the permeate liquid flow. The supply liquid pressure is higher than the concentrate liquid pressure, The supply fluid pressure is higher than the permeate fluid pressure. If one or more of the following conditions are not met, the data is removed from the historical performance data, or if the system is an ultrafiltration (UF) or microfiltration (MF) system, the physically inconsistent data points are subject to the following conditions: The supply fluid pressure is higher than the intermembrane pressure. The filtrate pressure is higher than the intermembrane pressure, and The turbidity of the supply liquid exceeds the turbidity of the filtrate. The system according to claim 4, wherein if one or more of the above conditions are not met, the data is removed from the past performance data.

6. The system according to any one of claims 1 to 5, wherein the processing apparatus is further configured to identify and remove statistically outlier data points from past performance data before identifying the cleaning event.

7. Identifying and removing statistical outlier data points from the aforementioned past performance data is: Previous data point (y t-1 ) Data point (y t ) calculate the percentage change, The aforementioned data point (y t ) Next data point (y t+1 ) calculate the percentage change, Comparing the previous percentage change with the lower limit value, If the percentage changes before and after exceed the aforementioned decrease limit, the data point (y t Identifying ) as an outlier and The system according to claim 6, including the system described in claim 6.

8. Identifying the aforementioned cleaning event means The difference series method is used to obtain the normalized pressure drop (dP) and normalized permeate flow (Pf) from the aforementioned past performance data. Including the use of the difference series method, The aforementioned data point (y t ) and the aforementioned previous data point (y t-1 The difference between the two is calculated to create a difference series of the normalized pressure drop (dP) and the normalized permeate flow (Pf), The mean and standard deviation of the difference series of the normalized pressure drop (dP) and the normalized permeate flow (Pf) are calculated, The values ​​of the difference series for the normalized pressure drop (dP) and the normalized permeate flow (Pf) are compared with the sum of the product of the difference constant and the corresponding standard deviation of the difference series and the corresponding value of the difference series. To determine whether gaps exist in the aforementioned dataset, To determine whether the aforementioned gap is wider than the MCE reference time. The system according to any one of claims 1 to 7, including the system described in any one of claims 1 to 7.

9. The system is the ultrafiltration (UF) or microfiltration (MF) system, and the cleaning event is identified by The difference series method is used for the transmembrane pressure drop (TMP), feed turbidity, and permeability from the aforementioned past performance data. Including the use of the difference series method, The difference in time values ​​between the aforementioned data point (t) and the previous data point (t-x) is calculated to create a time difference series. The difference between the said data point (y t ) and the previous data point (y t-x ) is calculated to create the said difference series for transmembrane pressure drop (TMP), feed solution turbidity, and permeability. The values ​​in the difference series of time, transmembrane pressure drop (TMP), feed turbidity, and permeability are compared with the thresholds for MCE events of backwashing, CEB, and CIP; To determine whether the comparison conforms to the backwash MCE standard, To determine whether the comparison conforms to the CEB MCE standard, To determine whether the comparison conforms to the CIP MCE criteria and The system according to any one of claims 1 to 8, including the system described in any one of claims 1 to 8.

10. The system according to any one of claims 1 to 9, wherein the processing device is further configured to analyze the coefficient of determination or the Akaike Information Criterion (AIC) of the mathematical function and to calculate the derivative of the mathematical function with respect to time, or the processing device is further configured to identify the type of fouling based on the coefficient of determination or the derivative of the Akaike Information Criterion (AIC).

11. The system according to any one of claims 1 to 10, wherein the one or more water filtration elements include at least one element selected from the group consisting of an ultrafiltration element, a microfiltration element, a reverse osmosis element, a nanofiltration element, and an ultrafiltration element.

12. The system according to any one of claims 1 to 11, further comprising one or more sensors configured to detect and transmit data relating to one or more of the pressure drop, salt permeability, and permeate flow associated with each of the one or more water filtration elements, for storage in the database as past performance data.

13. The system is a reverse osmosis (RO) or nanofiltration (NF) system. The aforementioned past performance data includes the pressure drop, salt permeability, and permeate flow rate of each of the one or more water filtration elements. The system according to any one of claims 1 to 12, wherein the apparatus identifies a fouling-free condition when no change in the pressure drop, salt permeability, and permeate flow is detected.

14. The system is the reverse osmosis (RO) or nanofiltration (NF) system, The aforementioned past performance data includes the pressure drop, salt permeability, and permeate flow rate of each of the one or more water filtration elements, and The system according to any one of claims 1 to 13, wherein the processing device identifies the type of fouling as biofouling when the increase in pressure drop fits a quadratic polynomial function.

15. The system is the reverse osmosis (RO) or nanofiltration (NF) system, The aforementioned past performance data includes the pressure drop, salt permeability, and permeate flow rate of each of the one or more water filtration elements, and The system according to any one of claims 1 to 14, wherein the apparatus identifies the type of fouling as organic fouling when the decrease in salt permeability fits a quadratic polynomial function.

16. The system is the reverse osmosis (RO) or nanofiltration (NF) system, The aforementioned past performance data includes the pressure drop, salt permeability, and permeate flow rate of each of the one or more water filtration elements, and The system according to any one of claims 1 to 15, wherein the processing apparatus defines the type of fouling as particulate fouling if the rapid increase in the pressure drop of the element at the first position fits a linear polynomial function.

17. The system is the reverse osmosis (RO) or nanofiltration (NF) system, The aforementioned past performance data includes the pressure drop, salt permeability, and permeate flow rate of each of the one or more water filtration elements, and The system according to any one of claims 1 to 16, wherein the processing device identifies the type of fouling as scaling if the gradual increase in the pressure drop of the element at the tail end position fits a linear polynomial function.

18. The system is the reverse osmosis (RO) or nanofiltration (NF) system, The aforementioned past performance data includes the pressure drop, salt permeability, and permeate flow rate of each of the one or more water filtration elements, and The system according to any one of claims 1 to 17, wherein if the rapid increase in the permeate flow fits a linear polynomial function and the rapid increase in the salt permeability fits a linear polynomial function, the apparatus identifies the type of fouling as a physical integrity failure.

19. The system is an ultrafiltration (UF) or microfiltration (MF) system, and the historical performance data includes one or more of the feed pressure, feed turbidity, intermembrane pressure drop, filtrate pressure, and filtrate turbidity, and the physically inconsistent data points are defined as follows: The supply fluid pressure is greater than or equal to the intermembrane pressure drop. The transmembrane pressure drop is less than or equal to the difference between the supply fluid pressure and the filtrate pressure, and The feed liquid turbidity is greater than the filtrate turbidity. The system according to any one of claims 1 to 11, wherein if one or more of the above conditions are not met, the data is removed from the past performance data.

20. The system is the ultrafiltration (UF) or microfiltration (MF) system, and the historical performance data includes one or more of the feed pressure, feed turbidity, intermembrane pressure drop, filtrate pressure, and filtrate turbidity, and the processing apparatus is subject to the following conditions: The increase in intermembrane pressure during filtration fits a linear polynomial function; the intermembrane pressure remains constant during backwashing; the intermembrane pressure remains constant during backwashing and CEC; the filtrate turbidity remains constant or remains below 1 NTU; and the washability remains constant or remains between 60% and 95%. The system according to any one of claims 1 to 11 or 19, which identifies the type of fouling as rapid removal or cake-forming fouling when one or more of the following conditions are met.

21. The system according to any one of claims 1 to 11, 19, or 20, wherein the system is the ultrafiltration (UF) or microfiltration (MF) system, and the historical performance data includes one or more of feed pressure, feed turbidity, intermembrane pressure drop, filtrate pressure, and filtrate turbidity, and the apparatus identifies the type of fouling as slow removal, biological or organic fouling when one or more of the following conditions are met: the increase in intermembrane pressure during filtration conforms to a linear function, a quadratic polynomial function, or another nonlinear function; the increase in intermembrane pressure during backwashing and CEC conforms to a linear or nonlinear function when no oxidizing agent or caustic agent is used; the filtrate turbidity remains constant or remains less than 1 NTU; and the cleanliness remains less than 60% or conforms to a linear polynomial.

22. The system is the ultrafiltration (UF) or microfiltration (MF) system, and the historical performance data includes one or more of the feed pressure, feed turbidity, intermembrane pressure drop, filtrate pressure, and filtrate turbidity, and the processing apparatus is subject to the following conditions: The increase in intermembrane pressure during filtration conforms to a linear function, a quadratic polynomial function, or another nonlinear function; the increase in intermembrane pressure during backwash conforms to a linear or nonlinear function; the increase in intermembrane pressure during CEC conforms to a linear or nonlinear function when no acid is used; the filtrate turbidity remains constant or less than 1 NTU; and the washability conforms to a quadratic polynomial or remains less than 60%. The system according to any one of claims 1 to 11, 19, 20, or 21, which identifies the type of fouling as irreversible or inorganic fouling when one or more of the following conditions are met.

23. The system is the ultrafiltration (UF) or microfiltration (MF) system, and the past performance data includes one or more of the feed liquid pressure, feed liquid turbidity, intermembrane pressure drop, filtrate pressure, and filtrate turbidity. The processing apparatus is as follows: The intermembrane pressure during filtration remains constant; the washing performance remains constant; the filtrate turbidity increases linearly; or the filtrate turbidity increases gradually. The system according to any one of claims 1 to 11 or 19 to 22, which identifies the type of failure as chemical decomposition when one or more of the following conditions are met.

24. The system is the ultrafiltration (UF) or microfiltration (MF) system, and the past performance data includes one or more of the feed pressure, feed turbidity, intermembrane pressure drop, filtrate pressure, and filtrate turbidity, and the apparatus meets the following conditions: the intermembrane pressure decreases linearly; the cleaning performance remains constant; the filtrate turbidity shows a rapid increase; the filtrate turbidity increases non-linearly. The system according to any one of claims 1 to 11 or 19 to 23, which identifies the type of failure as a physical integrity failure when one or more of the following conditions are met.

25. The aforementioned processing apparatus is The system receives current performance data of one or more water filtration elements, such as data received from one or more sensors, as input. Based on the previous analysis of the past performance data and the identification of the type of fouling occurring between each of the individual data segments, the current performance data is analyzed. To estimate the type of fouling currently occurring in one or more of the aforementioned water filtration elements, and Outputs a recommended cleaning procedure specific to the type of fouling currently occurring in one or more of the aforementioned water filtration elements. The system according to any one of claims 1 to 24, further configured as follows.

26. The system according to any one of claims 1 to 25, wherein the processing apparatus is further configured to detect the type of scaling, or the no-fouling or no-scaling condition.

27. The system according to any one of claims 1 to 26, wherein the data further comprises: the configuration of the membrane filtration unit; the operating conditions and protocol of the system; or both the configuration of the membrane filtration unit and the operating conditions and protocol of the system.

28. The system according to any one of claims 1 to 27, further comprising artificial intelligence, machine learning capabilities, or both, wherein the artificial intelligence or machine learning capabilities operate to improve the identification of fouling types, the predictability of fouling events, or both the identification of fouling types and the predictability of fouling events.

29. A method for determining the filter state, The system for determining the filter state receives past performance data of one or more water filtration elements as input to the system, wherein the system for determining the filter state includes (i) a database configured to electronically store the past performance data, and (ii) a processing device that communicates with the database. Identifying one or more data points in the past performance data that indicate a cleaning event of one or more of the water filtration elements, The process involves dividing the aforementioned past performance data into individual data segments, wherein each of the individual data segments represents the data of the aforementioned past performance data between each identified cleaning event. To fit a mathematical function to each of the aforementioned data segments, Based on the mathematical function adapted to each of the individual data segments, the type of fouling occurring between each of the individual data segments is identified. Methods that include...

30. The method according to claim 29, further comprising identifying the type of scaling, or the no-fouling or no-scaling conditions.

31. A non-temporary computer-readable medium for storing instructions for determining a filter state that can be executed by a processing device, wherein the processing device executes the instructions, The system for determining the filter state receives past performance data of one or more water filtration elements as input to the system, wherein the system for determining the filter state includes (i) a database configured to electronically store the past performance data, and (ii) a processing device that communicates with the database. Identifying one or more data points in the past performance data that indicate a cleaning event of one or more of the water filtration elements, The process involves dividing the aforementioned past performance data into individual data segments, wherein each of the individual data segments represents the data of the aforementioned past performance data between each identified cleaning event. To fit a mathematical function to each of the aforementioned data segments, Based on the mathematical function adapted to each of the individual data segments, the type of fouling occurring between each of the individual data segments is identified. A non-temporary computer-readable medium that enables execution.

32. The system according to claim 31, wherein the processing apparatus is further configured to identify a type of scaling, or a no-fouling or no-scaling condition.

Images