Intelligent monitoring reverse osmosis filtration system and fault diagnosis method thereof

Abstract

The invention relates to the field of filtering fault diagnosis, and discloses an intelligent monitoring reverse osmosis filtering system and a fault diagnosis method thereof, which are used for solving the problem of inaccurate fault judgment caused by sensor drifting during reverse osmosis filtering fault diagnosis, and comprises the following steps: if the system is judged to be abnormal, acquiring a sensor drifting parameter; the sensor drifting parameters comprise a static output value, response time and periodic stable data, a sensor drifting index is obtained through evaluation according to the sensor drifting parameters, whether sensor drifting exists currently is judged according to the sensor drifting index, and if it is judged that sensor drifting exists currently, the sensor drifting index is judged. If yes, performing drift correction on the initial operation parameter according to the sensor drift index to obtain a corrected operation parameter, and transmitting the corrected operation parameter to a feature calculation module and an anomaly identification module for secondary anomaly judgment, thereby effectively improving the accuracy of filtering diagnosis.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of filter fault diagnosis, and more particularly to an intelligent monitoring reverse osmosis filtration system and a fault diagnosis method thereof. BACKGROUND

[0002] As a core application form of membrane separation technology, reverse osmosis filtration systems are widely used in industrial pure water preparation, drinking water purification, seawater desalination, wastewater reuse and other scenarios to effectively remove dissolved salts, organic pollutants and microbial impurities in water. In typical applications, reverse osmosis systems are usually configured with multiple membrane components and are equipped with multiple sensors such as water inlet pressure, water flow, conductivity and temperature to monitor system operation in real time, guide operation adjustment and maintenance intervention, and thus ensure water quality and system efficiency.

[0003] In existing systems, reverse osmosis equipment generally uses methods of comparing running parameter thresholds and modeling characteristic indicators for fault recognition and health state assessment. Specifically, the system acquires sensor data in real time, extracts state characteristic indicators such as recovery rate, desalination rate, differential pressure change rate, and compares them with a preset threshold library to identify common operating abnormalities such as membrane fouling, concentrated water blockage, and decreased water production efficiency. Some technologies further introduce expert rules or state clustering models to improve diagnostic coverage and classification accuracy. Such methods usually involve multiple steps such as sensor data acquisition, data preprocessing, indicator calculation, and fault determination, which constitute the main technical path of current reverse osmosis system intelligent monitoring.

[0004] However, the above-mentioned technology at least has the following technical problems: During long-term operation of the reverse osmosis system, some sensors in high-temperature, high-pressure or fluctuating water source environments will gradually degrade in performance, causing their output signals to slowly drift over time, i.e., sensor drift. Such drift usually does not manifest as data anomalies in a short period of time, but the cumulative error will cause systematic deviation in the initial collected parameters. If the system directly uses the offset data to calculate the state characteristic indicators without identifying sensor drift, it may cause misjudgment of results such as abnormal desalination rate and differential pressure rise, thus triggering false alarms or intervention operations, seriously affecting the operation stability and diagnostic reliability of the system. SUMMARY

[0005] To overcome the above-mentioned defects of the prior art, the present application provides an intelligent monitoring reverse osmosis filtration system and a fault diagnosis method thereof to solve the problems in the background art.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions: The application discloses an intelligent monitoring reverse osmosis filtering system, which comprises a parameter acquisition module, an abnormality identification module, a sensor drift judgment module, a drift correction module, a second abnormality judgment module and a fault classification module.

[0007] Preferably, the state characteristic index acquisition step comprises the following steps: acquiring the water production flow and the water inlet flow, calculating the water production flow and the water inlet flow to obtain the recovery rate; acquiring the water inlet conductivity and the water production conductivity, calculating the water inlet conductivity and the water production conductivity to obtain the desalination rate; and acquiring the water inlet pressure and the water production pressure, calculating the water inlet pressure minus the water production pressure and the water inlet pressure to obtain the differential pressure change rate.

[0008] Preferably, the sensor drift index acquisition step comprises the following steps: acquiring the static output value during the zero input or stable standby state of the system, calculating the static output offset coefficient according to the static output value; acquiring the current response time under the standard test trigger condition, calculating the dynamic response drift coefficient according to the current response time; acquiring the stable data sequence in the running period, calculating the long-time sequence stability offset coefficient according to the stable data sequence; and performing normalization processing on the static output offset coefficient, the dynamic response drift coefficient and the long-time sequence stability offset coefficient, and evaluating the sensor drift index according to the normalized static output offset coefficient, the dynamic response drift coefficient and the long-time sequence stability offset coefficient. ; wherein SD represents a sensor drift index, OF represents a normalized static output offset coefficient, RD represents a normalized dynamic response drift coefficient, and SO represents a normalized long-term stability offset coefficient, 、 、 represent weight coefficients of the normalized static output offset coefficient, the normalized dynamic response drift coefficient, and the normalized long-term stability offset coefficient.

[0009] Preferably, the step of obtaining the static output offset coefficient comprises: during a zero-input state or a steady-state standby state of the current system, calling a sensor acquisition interface to continuously record an output value sequence to obtain a static output sequence; based on the acquired static output sequence, calculating a mean value of the sequence to obtain an output mean value; calling a system historical parameter management to extract a historical reference mean value of the sensor in a zero-input state during a factory calibration or a past normal operation period, and calculating the static output offset coefficient according to the output mean value and the historical reference mean value.

[0010] Preferably, the step of obtaining the dynamic response drift coefficient comprises: under a standard test trigger condition, recording a time point when an external input change occurs, denoted as an occurrence time point, and recording a time point when the sensor output reaches a steady-state value, denoted as a stable time point, calculating a difference between the stable time point and the occurrence time point to obtain a current response time; calling a historical reference response time of the current sensor recorded during initial installation or calibration from a system parameter management data; calculating a difference between the current response time and the historical reference response time, and calculating a ratio of the current response time to the historical reference response time to obtain the dynamic response drift coefficient.

[0011] Preferably, the step of obtaining the long-term stability offset coefficient comprises: selecting a stable section of the sensor output data in a current operation period, denoted as a stable data sequence, calculating a mean value of the stable data sequence to obtain a current period steady-state output mean value; calling a system operation record to obtain a corresponding steady-state output mean value sequence of the sensor in k historical periods, calculating a mean value of the steady-state output mean value sequence to obtain a historical reference mean value; calculating an absolute difference between the current period steady-state output mean value and the historical reference mean value, and calculating a ratio of the absolute difference to 1 plus an absolute value of the historical reference mean value to obtain the long-term stability offset coefficient.

[0012] Preferably, the step of determining whether there is a sensor drift according to the sensor drift index comprises: comparing the sensor drift index with a drift threshold value, if the sensor drift index is greater than or equal to the drift threshold value, determining that there is a sensor drift at present; if the sensor drift index is less than the drift threshold value, determining that there is no sensor drift at present, and no operation parameter correction is performed.

[0013] Preferably, the corrected operation parameter obtaining step is: for each initial operation parameter, a preliminary corrected operation parameter is calculated according to the sensor drift index; a time period similar to the current operation condition is selected in the historical operation record, and the historical steady-state average of each operation parameter is calculated; the preliminary corrected operation parameter is added to the historical steady-state average and then divided by 2 to obtain the corrected operation parameter.

[0014] Preferably, an intelligent monitoring reverse osmosis filtration fault diagnosis method comprises the following steps: step 1: during reverse osmosis filtration, initial operation parameters are collected in real time by sensors, the operation parameters including water inlet pressure, concentrated water pressure, water production flow, concentrated water flow, conductivity and temperature; the collected initial operation parameters are preprocessed to obtain preprocessed operation parameters; step 2: based on the preprocessed operation parameters, state characteristic indexes are calculated, the state characteristic indexes including recovery rate, desalination rate and differential pressure change rate; the currently calculated state characteristic indexes are compared with a preset fault threshold library, and whether the system has an abnormality is determined according to the comparison result; step 3: if it is determined that the system has an abnormality, sensor drift parameters are obtained, the sensor drift parameters including static output value, response time and periodic stable data, a sensor drift index is evaluated according to the sensor drift parameters, and whether there is a sensor drift at present is determined according to the sensor drift index; step 4: if it is determined that there is a sensor drift at present, the initial operation parameters are drift-corrected according to the sensor drift index to obtain corrected operation parameters; step 5: the corrected operation parameters are transmitted to a characteristic calculation module and an abnormality identification module for second abnormality determination; step 6: if the second abnormality determination has an abnormality, abnormality characteristics are obtained, a fault type is classified according to a fault rule library corresponding to the abnormality characteristics to obtain a fault classification result, and diagnosis prompt information is generated according to the fault classification result; and step 7: the diagnosis prompt information is output to an operation and maintenance terminal through a man-machine interaction interface.

[0015] Technical effects and advantages of the present application: If it is determined that the system has an abnormality, sensor drift parameters are obtained, the sensor drift parameters including static output value, response time and periodic stable data, a sensor drift index is evaluated according to the sensor drift parameters, and whether there is a sensor drift at present is determined according to the sensor drift index, if it is determined that there is a sensor drift at present, the initial operation parameters are drift-corrected according to the sensor drift index to obtain corrected operation parameters, the corrected operation parameters are transmitted to a characteristic calculation module and an abnormality identification module for second abnormality determination, and the accuracy of filtration diagnosis is effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A structure diagram of an intelligent monitoring reverse osmosis filtration system is provided for the embodiments of the present application.

[0017] Figure 2 An intelligent monitoring reverse osmosis filtration fault diagnosis method flow chart is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0018] The technical solutions in the present application will be described clearly and completely in combination with the drawings in the present application. In addition, the forms of each structure described in the following embodiments are only examples. The intelligent monitoring reverse osmosis filtration system and the fault diagnosis method thereof involved in the present application are not limited to each structure described in the following embodiments. All other embodiments obtained by those skilled in the art without making creative efforts belong to the scope of protection of the present application.

[0019] The present application provides an intelligent monitoring reverse osmosis filtration system, as shown in the accompanying drawings, the system comprises: Figure 1 A parameter acquisition module is configured to acquire initial operation parameters in real time through a sensor when reverse osmosis filtration is performed. The operation parameters include water production flow, water inlet flow, water inlet conductivity, water production conductivity, water inlet pressure, water production pressure and the like, and are used to construct a current operation state data set of the system. The acquired initial operation parameters are preprocessed to obtain preprocessed operation parameters. The preprocessing includes filtering, outlier rejection and time alignment processing, so as to ensure that the input data is complete, continuous and consistent, and to provide a reliable data basis for subsequent diagnosis and analysis. An abnormality identification module is configured to calculate state characteristic indexes based on the preprocessed operation parameters. The state characteristic indexes include recovery rate, desalination rate and differential pressure change rate and the like, and are used to reflect the operation health status of the reverse osmosis system. The current calculated state characteristic indexes are compared with a preset fault threshold library, and it is determined whether the system has an abnormality according to the comparison result. It should be specifically noted that the fault threshold library refers to a set of index determination criteria for various key characteristic indexes in the operation state of the reverse osmosis system, which is used to distinguish between normal operation state and potential abnormal state. The threshold library is set according to system design specifications, historical operation data and expert experience, and contains safety threshold ranges or upper and lower threshold values corresponding to multiple state characteristic indexes such as recovery rate, desalination rate and differential pressure change rate. When the calculated state characteristic indexes exceed the corresponding threshold range, it is determined that the current system operation has an abnormality, which is used to trigger the subsequent fault diagnosis process.

[0020] In the present embodiment, it should be specifically noted that the state characteristic index acquisition step is: ​The water production flow rate and the water inlet flow rate in the running parameters are obtained, and the water production flow rate is calculated by the water inlet flow rate to obtain the recovery rate, which is used to represent how much proportion of the water inlet is converted into the water production, and reflects the running efficiency of the membrane system. The higher the value is, the better the water resource utilization rate is. The water inlet conductivity and the water production conductivity in the running parameters are obtained, and the desalination rate is calculated according to the water inlet conductivity and the water production conductivity. The specific obtaining steps are as follows: . In the formula, S represents the desalination rate, which is used to measure the removal ability of the system to the dissolved ions, represents the water production conductivity, represents the water inlet conductivity. The higher the desalination rate is, the purer the water quality is. If the index decreases, it may indicate that the membrane is polluted or penetrated. The water inlet pressure and the water production pressure in the running parameters are obtained, and the differential pressure change rate is calculated by the ratio of the water inlet pressure minus the water production pressure to the water inlet pressure. The larger the differential pressure change rate is, generally indicating that the membrane resistance increases, which may be caused by phenomena such as particle blockage, organic pollution or biological pollution blockage.

[0021] The sensor drift determination module determines that the system is abnormal, and obtains the sensor drift parameters, including the static output value, the response time and the periodic stable data. The sensor drift index is evaluated according to the sensor drift parameters, and whether the sensor drift exists at present is determined according to the sensor drift index. In this embodiment, it needs to be specifically explained that the obtaining steps of the sensor drift index are as follows: The static output value during the zero input or stable standby state of the system is obtained, and the static output offset coefficient is calculated according to the static output value; The current response time is obtained under the standard test triggering condition, and the dynamic response drift coefficient is calculated according to the current response time; The stable data sequence in the running period is obtained, and the long-time sequence stability offset coefficient is calculated according to the stable data sequence; The static output offset coefficient, the dynamic response drift coefficient and the long-time sequence stability offset coefficient are normalized, and the sensor drift index is evaluated according to the normalized static output offset coefficient, the dynamic response drift coefficient and the long-time sequence stability offset coefficient. The specific obtaining steps are as follows: ; In the formula, SD represents a sensor drift index, OF represents a normalized static output offset coefficient, the greater the output offset of the sensor in a zero-input state, that is, the higher the value of the static output offset coefficient, the more serious the overall drift of the sensor. The coefficient reflects the systematic deviation of the sensor body output under no load or steady state conditions, so its trend can be directly used to represent the influence of the degradation of the static performance of the sensor on the overall drift level. The greater the coefficient value, the more significant the contribution to the sensor drift index. RD represents a normalized dynamic response drift coefficient. The more significant the delay of the current response time compared with the historical reference value, the greater the value of the coefficient, indicating that the dynamic response performance of the sensor has degraded, and the stronger the influence on the overall drift trend. The coefficient is used to measure the tracking ability of the sensor to rapidly changing signals in actual operation, and the change in its value can directly reflect the dynamic contribution of the hysteresis degree of the sensor to the drift level. SO represents a normalized long-term stability offset coefficient. The greater the deviation between the current period output value and the historical multiple period reference mean value, the higher the value of the coefficient, indicating that the sensor has a persistent and progressive drift trend. The coefficient reflects the change in the output stability of the sensor during long-term operation, and is representative of the medium and long-term influence on the overall drift level. The greater the value, the more significant the weight contribution in the sensor drift index, 、 、 OF, RD and SO are weight coefficients of the normalized static output offset coefficient, the normalized dynamic response drift coefficient and the normalized long-term stability offset coefficient, respectively, and , 、 、 are obtained by the analytic hierarchy process. The analytic hierarchy process is a multi-criteria decision-making method. By constructing a hierarchical structure model between the decision target, evaluation index and scheme, combining expert judgment or system evaluation to compare the importance between each index, constructing a judgment matrix, and then calculating the weight of each index relative to the overall target. This method combines the advantages of subjective judgment and mathematical derivation, and is suitable for weight distribution of multiple influencing factors in a complex system, ensuring that each influence coefficient has a reasonable evaluation weight in the calculation of the sensor drift index.

[0022] In this embodiment, it needs to be specifically pointed out that the acquisition step of the static output offset coefficient is: During the period when the current system is in a zero-input state or a steady-state standby state, a sensor acquisition interface is called to continuously record the output value sequence, and a static output sequence is obtained. It needs to be specifically pointed out that the zero-input state refers to a state in which the system runs without load, flow or signal input, and theoretically the sensor should output zero. Based on the collected static output sequence, the mean value of the sequence is calculated to obtain an output mean value for representing a systematic offset trend; The system history parameter management is called to extract a historical reference mean value of the sensor in a zero input state during factory calibration or past normal operation period, and a static output offset coefficient is calculated according to the output mean value and the historical reference mean value. The specific obtaining steps are as follows: In the formula, OF represents the static output offset coefficient, represents the output mean value, represents the historical reference mean value, and the denominator plus 1 is used to avoid the case that the denominator is zero. The larger the static output offset coefficient is, the more serious the static drift is, and it is used to reflect the static degradation trend of the sensor.

[0023] In the embodiment, it needs to be specifically explained that the obtaining steps of the dynamic response drift coefficient are as follows: Under the standard test triggering condition (such as starting the pump, starting the valve or flow mutation), the time point when the external input change occurs (such as the flow starts to change) of the target sensor is recorded, which is recorded as the occurrence time point, and the time point when the sensor output reaches the steady state value is recorded, which is recorded as the stable time point. The stable time point and the occurrence time point are difference calculated to obtain the current response time; The historical reference response time recorded by the current sensor at the initial installation or calibration is called from the system parameter management data, which is used as the ideal or initial steady state response time of the sensor under the same excitation condition, with the unit of seconds, as the comparison reference of the current response time; The difference calculation between the current response time and the historical reference response time and the ratio calculation between the historical reference response time are performed to obtain the dynamic response drift coefficient.

[0024] In the embodiment, it needs to be specifically explained that the obtaining steps of the long-term sequence stability offset coefficient are as follows: In the current operation period, the stable section of the sensor output data is selected, which is recorded as a stable data sequence, and the mean value of the stable data sequence is calculated to obtain a current period steady state output mean value; It needs to be specifically explained that the operation period refers to the continuous stable operation stage of the reverse osmosis system in a complete start-stop process, a membrane section continuous filtration process or a set time window. Usually, the system is in a stable working condition, the sensor continuously collects data and there is no start-stop fluctuation as the determination basis. The operation period is used as a time division unit for sensor performance evaluation, which is used to capture the representative sensor output characteristics in each period, so as to facilitate the cross-period consistency analysis and long-term stability evaluation.

[0025] ​The stable section of data refers to a continuous sampling time period in the sensor output data within a running cycle, the fluctuation amplitude of which is small, and the change trend is stable, usually meeting the condition that the output change rate is close to zero or fluctuates within a set threshold range. The stable section reflects the true measurement performance of the system under undisturbed, steady-state load or constant environmental conditions, and is the basis for calculating the steady-state output mean value in the calculation period, which can eliminate transient influencing factors such as start-up disturbance and fluctuation adjustment, and improve the accuracy of drift evaluation.

[0026] The system running record is called to obtain the corresponding steady-state output mean value sequence of the sensor in the k historical cycles, the mean value of the steady-state output mean value sequence is calculated, and the historical reference mean value is obtained; The absolute difference between the current cycle steady-state output mean value and the historical reference mean value is calculated, and the ratio with the absolute value of the historical reference mean value plus 1 is calculated to obtain the long-term sequence stability offset coefficient. The larger the long-term sequence stability offset coefficient is, the more significant the current output deviates from the historical long-term characteristics, indicating that the sensor has a long-term drift trend.

[0027] In this embodiment, it needs to be specifically explained that the step of determining whether there is sensor drift at present according to the sensor drift index is: The sensor drift index is compared with the drift threshold value. If the sensor drift index is greater than or equal to the drift threshold value, it is determined that there is sensor drift at present. If the sensor drift index is less than the drift threshold value, it is determined that there is no sensor drift at present, and no running parameter correction is performed. The drift index is obtained by an adaptive threshold method, which is a threshold determination method based on the dynamic change characteristics of running data. By monitoring the historical distribution characteristics and fluctuation trend of the sensor drift index in real time, combined with the current environmental parameters, running state and statistical deviation, the drift determination threshold value is dynamically adjusted, so that the threshold value can be automatically updated with the system running state. Avoid the misalignment problem of fixed threshold value under different working conditions, and adaptively reflect the drift sensitivity of the sensor under different temperatures, pressures and aging degrees, thereby improving the accuracy and environmental adaptability of drift determination.

[0028] The drift correction module, if it is determined that there is sensor drift at present, drift correction is performed on the initial running parameters according to the sensor drift index to obtain the corrected running parameters; In this embodiment, it needs to be specifically explained that the corrected running parameters are obtained by the following steps: For each initial running parameter, the preliminary corrected running parameter is calculated according to the sensor drift index, and the specific obtaining steps are as follows: ; In the formula, is the preliminary corrected running parameter, is the initial running parameter, and SD is the sensor drift index. In the historical operation record, a time period similar to the current operation condition is selected, and the historical steady-state mean value of each operation parameter is calculated, denoted as , which is used to reflect the reference output level of the system in the state without drift; the mean value can be calculated based on a sliding time window or the average value in the last N normal cycles, and all the mean values are derived from the system operation log; The preliminary corrected operation parameter is added to the historical steady-state mean value and divided by 2 to obtain the corrected operation parameter.

[0029] The second abnormality determination module is used to transmit the corrected operation parameter to the feature calculation module and the abnormality identification module for the second abnormality determination; After completing the drift correction of the initial operation parameter, the state feature calculation and abnormality identification process are re-executed based on the corrected parameter, avoiding the misdiagnosis or missed judgment problem caused by sensor drift. By re-inputting the corrected high-credibility data into the abnormality identification module for the second judgment, the pseudo-abnormal results caused by sensor errors can be effectively eliminated, the accuracy and stability of fault determination are improved, more reliable abnormal feature input is provided for the subsequent fault classification module, and the anti-interference ability and diagnostic reliability of the entire diagnostic system in complex environment are enhanced.

[0030] The fault classification module obtains the abnormal feature if the second abnormality determination exists, classifies the fault type according to the fault rule library corresponding to the abnormal feature, obtains the fault classification result, and generates the diagnostic prompt information containing the fault type, the influence range and the processing suggestion according to the fault classification result; It should be specifically pointed out that the fault type classification according to the fault rule library corresponding to the abnormal feature, the fault classification result, and the diagnostic prompt information containing the fault type, the influence range and the processing suggestion according to the fault classification result are prior art, and the specific steps are not described in detail.

[0031] It should be specifically pointed out that the abnormal feature refers to the index item in the state feature index that is judged to be out of the preset normal range, which is used to represent the specific parameter feature of abnormal fluctuation or performance degradation in the current system operation state.

[0032] It needs to be specifically pointed out that the fault rule base refers to a set of knowledge rules for realizing fault type recognition and classification judgment, which is constructed based on the mapping relationship between abnormal features and historical fault modes. The rule base contains the association rules, logical expressions or classification conditions between various typical abnormal feature combinations and fault types, which can support the automatic identification of various reverse osmosis system faults such as membrane pollution, abnormal water inlet, and pressure difference mutation. After identifying the abnormal features, the system matches and compares with the fault rule base to determine the fault type that best matches the current state, and outputs the corresponding diagnostic information and intervention suggestions accordingly.

[0033] The information prompting module is used to output the diagnostic prompt information to the operation and maintenance terminal through the human-computer interaction interface to assist the user in quickly locating and processing the fault problem.

[0034] In this embodiment, it needs to be specifically pointed out that, as shown in Figure 2 An intelligent monitoring reverse osmosis filtration fault diagnosis method, comprising the following steps: Step 1: During reverse osmosis filtration, real-time acquisition of initial operating parameters is performed through sensors, and the operating parameters include water inlet pressure, concentrated water pressure, water production flow, concentrated water flow, electrical conductivity and temperature. The acquired initial operating parameters are preprocessed to obtain preprocessed operating parameters; Step 2: Based on the preprocessed operating parameters, state feature indexes are calculated, including recovery rate, desalination rate and pressure difference change rate; the current calculated state feature indexes are compared with the preset fault threshold library, and whether the system has an abnormality is determined according to the comparison result; Step 3: If it is determined that the system has an abnormality, sensor drift parameters are obtained, including static output value, response time and periodic stable data, a sensor drift index is evaluated according to the sensor drift parameters, and whether there is a sensor drift at present is determined according to the sensor drift index; Step 4: If it is determined that there is a sensor drift at present, the initial operating parameters are drift-corrected according to the sensor drift index to obtain corrected operating parameters; Step 5: The corrected operating parameters are transmitted to the feature calculation module and the abnormality identification module for second abnormality determination; Step 6: If the second abnormality determination has an abnormality, abnormal features are obtained, a fault type classification is performed according to the fault rule base corresponding to the abnormal features, a fault classification result is obtained, and diagnostic prompt information is generated according to the fault classification result; Step 7: The diagnostic prompt information is output to the operation and maintenance terminal through the human-computer interaction interface.

[0035] Finally: the above only for the preferred embodiments of the present application, and not for limiting the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application, should be included in the scope of protection of the present application.

[0036] The above description is merely a specific implementation of the present application, but the protection scope of the present application is not limited thereto, any modification, equivalent replacement, improvement, etc. made within the technical range disclosed by the present application, should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An intelligent monitoring reverse osmosis filtration system, characterized in that, The system comprises: A parameter acquisition module, configured to acquire initial operation parameters in real time by a sensor during reverse osmosis filtration, the operation parameters including water inlet pressure, concentrated water pressure, water production flow, concentrated water flow, conductivity and temperature; and to preprocess the acquired initial operation parameters to obtain preprocessed operation parameters; An anomaly identification module, configured to calculate state characteristic indexes based on the preprocessed operation parameters, the state characteristic indexes including recovery rate, desalination rate and differential pressure change rate; To compare the currently calculated state characteristic indexes with a preset fault threshold library, and to determine whether the system has an anomaly according to the comparison result; A sensor drift determination module, configured to acquire sensor drift parameters including static output value, response time and periodic stable data if it is determined that the system has an anomaly, to evaluate a sensor drift index based on the sensor drift parameters, and to determine whether there is sensor drift at present based on the sensor drift index; A drift correction module, configured to perform drift correction on the initial operation parameters based on the sensor drift index to obtain corrected operation parameters if it is determined that there is sensor drift at present; A second anomaly determination module, configured to transmit the corrected operation parameters to the feature calculation module and the anomaly identification module for second anomaly determination; A fault classification module, configured to acquire anomaly features if the second anomaly determination has an anomaly, to perform fault type classification based on a fault rule library corresponding to the anomaly features to obtain a fault classification result, and to generate a diagnosis prompt information based on the fault classification result; An information prompt module, configured to output the diagnosis prompt information to an operation and maintenance terminal through a man-machine interaction interface.

2. The intelligent monitoring reverse osmosis filtration system according to claim 1, wherein: The state characteristic index acquisition step is: The water production flow and the water inlet flow in the operation parameters are acquired, the water production flow and the water inlet flow are calculated by ratio, and the recovery rate is obtained; The water inlet conductivity and the water production conductivity in the operation parameters are acquired, and the desalination rate is calculated based on the water inlet conductivity and the water production conductivity; The water inlet pressure and the water production pressure in the operation parameters are acquired, the water inlet pressure is subtracted from the water production pressure, and the water inlet pressure is calculated by ratio, and the differential pressure change rate is obtained.

3. The intelligent monitoring reverse osmosis filtration system of claim 1, wherein, The sensor drift index acquisition step is: The static output value during the zero input or steady-state standby state of the system is acquired, and the static output offset coefficient is calculated based on the static output value; The current response time is acquired under the standard test triggering condition, and the dynamic response drift coefficient is calculated based on the current response time; The stable data sequence in the operation period is acquired, and the long-time sequence stability offset coefficient is calculated based on the stable data sequence; The static output offset coefficient, the dynamic response drift coefficient and the long-time sequence stability offset coefficient are normalized, and the sensor drift index is evaluated based on the normalized static output offset coefficient, the dynamic response drift coefficient and the long-time sequence stability offset coefficient, and the specific acquisition step is: ； In the formula, SD represents a sensor drift index, OF represents a normalized static output offset coefficient, RD represents a normalized dynamic response drift coefficient, SO represents a normalized long-term stability offset coefficient, , , represent weight coefficients of the normalized static output offset coefficient, the normalized dynamic response drift coefficient, and the normalized long-term stability offset coefficient.

4. The intelligent monitoring reverse osmosis filtration system of claim 3, wherein, The static output offset coefficient acquisition step is: During the zero input state or the steady-state standby state of the current system, the sensor acquisition interface is called to continuously record the output value sequence, and the static output sequence is obtained; Based on the acquired static output sequence, the mean value of the sequence is calculated, and the output mean value is obtained; The system history parameter management is called to extract the historical reference mean value of the sensor in the zero input state during the factory calibration or the past normal operation period, and the static output offset coefficient is calculated according to the output mean value and the historical reference mean value.

5. The intelligent monitoring reverse osmosis filtration system of claim 3, wherein: The acquisition step of the dynamic response drift coefficient is: Under the standard test trigger condition, the time point when the external input change occurs is recorded as the occurrence time point, and the time point when the sensor output reaches the steady-state value is recorded as the stable time point, the stable time point and the occurrence time point are calculated by difference, and the current response time is obtained; The historical reference response time recorded when the current sensor is initially installed or calibrated is called from the system parameter management data; The current response time and the historical reference response time are calculated by difference, and the dynamic response drift coefficient is calculated by ratio.

6. The intelligent monitoring reverse osmosis filtration system of claim 3, wherein: The acquisition step of the long-term sequence stability offset coefficient is: In the current operation period, the stable section of the sensor output data is selected, recorded as the stable data sequence, and the mean value of the stable data sequence is calculated to obtain the current period steady-state output mean value; The system operation record is called to obtain the corresponding steady-state output mean value sequence of the sensor in k historical periods, and the mean value of the steady-state output mean value sequence is calculated to obtain the historical reference mean value; The current period steady-state output mean value and the historical reference mean value are calculated by absolute difference, and the long-term sequence stability offset coefficient is calculated by ratio with 1 plus the absolute value of the historical reference mean value.

7. The intelligent monitoring reverse osmosis filtration system of claim 1, wherein: The step of determining whether the sensor drift exists currently according to the sensor drift index is: The sensor drift index is compared with the drift threshold value, if the sensor drift index is greater than or equal to the drift threshold value, it is determined that the sensor drift exists currently, if the sensor drift index is less than the drift threshold value, it is determined that the sensor drift does not exist currently, and the operation parameter correction is not performed.

8. The intelligent monitoring reverse osmosis filtration system of claim 1, wherein: The step of obtaining the corrected operation parameter is: For each initial operation parameter, the preliminary corrected operation parameter is calculated according to the sensor drift index; In the historical operation record, a time period similar to the current operation condition is selected, and the historical steady-state mean value of each operation parameter is calculated; The preliminary corrected operation parameter is added to the historical steady-state mean value and then divided by 2 to obtain the corrected operation parameter.

9. A method for diagnosing faults of a reverse osmosis filtration system, the method being used to implement the intelligent monitoring reverse osmosis filtration system of claims 1-8, characterized in that: The method comprises the following steps: Step 1: during reverse osmosis filtration, real-time acquisition of initial operation parameters is performed through a sensor, the operation parameters including water inlet pressure, concentrated water pressure, water production flow, concentrated water flow, conductivity and temperature; the acquired initial operation parameters are preprocessed to obtain preprocessed operation parameters; Step 2: based on the preprocessed operation parameters, state characteristic indexes are calculated, the state characteristic indexes including recovery rate, desalination rate and pressure difference change rate; the state characteristic indexes calculated currently are compared with a preset fault threshold value library, and whether the system is abnormal is determined according to the comparison result; Step 3: If it is determined that the system has an anomaly, obtain sensor drift parameters including static output values, response times, and periodic stability data, evaluate a sensor drift index based on the sensor drift parameters, and determine whether there is a sensor drift based on the sensor drift index; Step 4: If it is determined that there is a sensor drift, correct the initial operating parameters based on the sensor drift index to obtain corrected operating parameters; Step 5: Transmit the corrected operating parameters to the feature calculation module and the anomaly identification module for a second anomaly determination; Step 6: If the second anomaly determination has an anomaly, obtain an anomaly feature, classify a fault type based on a fault rule library corresponding to the anomaly feature to obtain a fault classification result, and generate a diagnostic prompt message based on the fault classification result; Step 7: Output the diagnostic prompt message to the operation and maintenance terminal through a human-computer interaction interface.

Images