Search

CN-122006476-A - Intelligent operation and maintenance system and method for reverse osmosis system based on physical-AI hybrid drive

CN122006476ACN 122006476 ACN122006476 ACN 122006476ACN-122006476-A

Abstract

The invention discloses a reverse osmosis system intelligent operation and maintenance system and method based on physical-AI hybrid drive, which relate to the technical field of water treatment and membrane separation engineering and comprise the following steps: the method comprises the steps of collecting the change data of high-pressure pump frequency, water inlet pressure and water production flow along with time in the operation process of a reverse osmosis system, uniformly mapping all the operation data to the same time scale, and establishing an operation rhythm reference time zone covering the whole reverse osmosis device. The invention realizes synchronous alignment and fusion of the physical model and the artificial intelligent model by establishing the operation rhythm reference time zone with uniform time scales, ensures the consistency of the operation and maintenance judgment time sequence and reliable decision, builds a single decision chain and a ripple type time regulation and control ring in a high risk period, ensures continuous and smooth output of control instructions, and improves the operation stability and intelligent regulation and control capability of the reverse osmosis system.

Inventors

  • CUI HAIYING

Assignees

  • 济宁市鲁泉水处理有限公司

Dates

Publication Date
20260512
Application Date
20260308

Claims (10)

  1. 1. The intelligent operation and maintenance method of the reverse osmosis system based on the physical-AI hybrid drive is characterized by comprising the following steps: Collecting the change data of the frequency of a high-pressure pump, the water inlet pressure and the water production flow rate along with time in the operation process of the reverse osmosis system, uniformly mapping all the operation data to the same time scale, and establishing an operation rhythm reference time zone covering the whole reverse osmosis device; Based on the operation rhythm reference time band, carrying out point-by-point alignment on real-time operation data output by the physical model and a trend prediction result output by the artificial intelligent model according to time scales to generate a time differential map; extracting extreme points of the offset segments along the time differential spectrum to form a rhythm conflict list, and binding the rhythm conflict list with the load variation amplitude of the corresponding time period; Based on the rhythm conflict list, arranging a physical model judgment result and an artificial intelligent model judgment result into priority channels according to the risk degree, and rearranging the execution sequence of control instructions in a high-risk decision period to construct a single decision chain; and a ripple time regulation ring is arranged around the single decision chain, the effective time of the control instruction is regulated in the high risk decision period, a short buffer time window is inserted, and the rhythm fluctuation is smoothed, so that the single decision chain keeps a continuous output track on a time axis.
  2. 2. The intelligent operation and maintenance method of reverse osmosis system based on physical-AI hybrid driving according to claim 1, wherein the operation rhythm reference time band generation step is as follows: Continuously acquiring real-time change data of the frequency of the high-pressure pump, the water inlet pressure and the water production flow rate in the operation process of the reverse osmosis device, and arranging the data according to the sampling time sequence to form an original time sequence; taking a sampling time sequence of the high-pressure pump frequency as a reference main time axis, synchronously expanding time records of water inlet pressure and water production flow, and redistributing operation parameters of different sources to uniform time nodes; combining the time mapped values of the high-pressure pump frequency, the water inlet pressure and the water production flow point by point according to a time sequence to generate an operation rhythm reference time zone covering the whole device; And recording the time nodes and the change trend of each operation parameter in the same time frame by taking the operation rhythm reference time band as a unified reference to form an information structure containing a complete time index.
  3. 3. The intelligent operation and maintenance method of reverse osmosis system based on physical-AI hybrid driving according to claim 2, wherein in the process of generating the operation rhythm reference time zone, the time sequence of high-pressure pump frequency is used as a core reference, and the continuous numerical sequence of each operation parameter is formed on a unified time node by carrying out prolongation and weighting distribution on the time records of water inlet pressure and water outlet flow, and the operation rhythm reference time zone is established in time sequence.
  4. 4. The intelligent operation and maintenance method of reverse osmosis system based on physical-AI hybrid driving according to claim 2, wherein the step of aligning the physical model output with the artificial intelligent model output point by point based on the operation rhythm reference time band is as follows: selecting the frequency of a high-pressure pump, the water inlet pressure and the water production flow as alignment references, and enabling real-time operation data output by a physical model to correspond to trend prediction results output by an artificial intelligent model one by one according to time nodes of an operation rhythm reference time zone; Correspondingly analyzing the physical model output and the artificial intelligent model output in each time node, extracting the operation variable value and forming a continuous differential data set; arranging the differential data sets according to a time sequence to form a time differential sequence, and mapping the time differential sequence onto an operation rhythm reference time band to generate a time differential map; And carrying out extension analysis on the offset trend in the time differential spectrum to form a time offset band reflecting the model judgment difference change trend.
  5. 5. The intelligent operation and maintenance method of reverse osmosis system based on physical-AI hybrid driving of claim 4, wherein in the process of generating the time differential spectrum, differential values of adjacent time nodes in the time differential sequence are connected in a continuous time mapping manner, so that the time differential spectrum forms a continuously distributed differential track band on a time axis, and extreme points of offset amplitude variation are used as key feature points.
  6. 6. The intelligent operation and maintenance method of reverse osmosis system based on physical-AI hybrid driving according to claim 4, wherein the steps of extracting extreme points of offset segments along a time differential map and forming a rhythm conflict list are as follows: Continuously scanning offset data distributed along a time axis of the time differential spectrum, and identifying offset fragments with differences between a physical model and an artificial intelligent model judgment result; Extracting extreme points of the change of the offset amplitude along the time direction of each offset segment, and recording the time positions of the extreme points and the corresponding offset amplitude; Constructing a rhythm conflict list by taking extreme points as key indexes, and recording time positions, offset amplitudes, offset directions and duration corresponding to the extreme points as structured data sequences; Binding the rhythmic conflict list with the load change amplitude of the reverse osmosis device in the corresponding time period, determining the high risk decision time period and identifying on the operation time axis.
  7. 7. The intelligent operation and maintenance method of reverse osmosis system based on physical-AI hybrid driving according to claim 6, wherein the step of binding the rhythmic conflict list with the load variation amplitude is further defined as determining the association relation between the load variation amplitude and the offset amplitude by comparing the inflow pressure variation, the produced water flow variation and the high-pressure pump frequency variation in the adjacent time periods before and after the extreme point, and using the association relation as the basis for determining the high-risk decision period.
  8. 8. The intelligent operation and maintenance method of reverse osmosis system based on physical-AI hybrid driving according to claim 6, wherein the step of constructing a single decision chain based on a list of rhythms conflicts is as follows: Performing risk degree analysis on conflict events recorded in the rhythm conflict list, determining a risk level according to the offset amplitude, the duration and the load variation amplitude, and extracting judgment results of a physical model and an artificial intelligent model; layering and arranging the physical model judgment result and the artificial intelligent model judgment result according to the risk level to form a risk priority channel, and maintaining the mapping relation between the time sequence and the risk level; Rearranging the control instructions in a high-risk decision period, and arranging the control instructions of different models into a continuous execution sequence according to the time sequence and the risk level; and mapping the recombined control logic to an operation time axis to form a single decision chain with continuous time and unified logic, so that the judgment result at the previous time is continuously imported into the decision process at the later time.
  9. 9. The intelligent operation and maintenance method of reverse osmosis system based on physical-AI hybrid driving according to claim 8, wherein the step of setting ripple time regulation loop around single decision chain is as follows: continuously scanning an operation time axis covered by a single decision chain, identifying a time interval containing a high-risk decision period, and establishing a time regulation interval; A ripple type time regulation and control ring is arranged around a high-risk decision period, and a buffer zone with time attenuation characteristic is formed before and after a control instruction effective point, so that the instruction execution process has flexible time span; Reassigning the effective moment of the control instruction according to the risk level and the execution priority in the ripple time regulation ring to form a time layered structure which spreads from the core to the periphery; and mapping the adjusted effective time of the instruction and the buffer time window into a single-decision chain, so that the control instruction forms a continuous output track on a time axis and keeps a stable running rhythm.
  10. 10. The intelligent operation and maintenance system of the reverse osmosis system based on the physical-AI hybrid driving is used for realizing the intelligent operation and maintenance method of the reverse osmosis system based on the physical-AI hybrid driving as set forth in any one of the claims 1 to 9, and is characterized by comprising a time reference construction module, a time sequence alignment analysis module, a conflict identification extraction module, a decision chain construction module and a time sequence regulation and smoothing module: The time reference construction module is used for collecting the change data of the frequency of the high-pressure pump, the water inlet pressure and the water production flow rate along with time in the operation process of the reverse osmosis system, uniformly mapping all the operation data to the same time scale, and establishing an operation rhythm reference time zone for covering the whole reverse osmosis device; the time sequence alignment analysis module is used for aligning the real-time operation data output by the physical model with the trend prediction result output by the artificial intelligent model point by point according to the time scale based on the operation rhythm reference time band to generate a time differential map; the conflict identification extraction module is used for extracting extreme points of the offset segments along the time differential spectrum to form a rhythm conflict list, and binding the rhythm conflict list with the load change amplitude of the corresponding time period; The decision chain construction module is used for arranging the judgment result of the physical model and the judgment result of the artificial intelligent model into priority channels according to the risk degree based on the rhythm conflict list, and rearranging the execution sequence of the control instruction in the high-risk decision period to construct a single decision chain; The time sequence regulating and smoothing module is used for setting a ripple type time regulating and controlling ring around the single decision chain, slightly regulating the effective time of the control instruction in the high risk decision period, inserting a short buffer time window and smoothing the rhythm fluctuation, so that the single decision chain keeps a continuous output track on a time axis.

Description

Intelligent operation and maintenance system and method for reverse osmosis system based on physical-AI hybrid drive Technical Field The invention relates to the technical field of water treatment and membrane separation engineering, in particular to an intelligent operation and maintenance system and method of a reverse osmosis system based on physical-AI hybrid driving. Background The intelligent operation and maintenance of the reverse osmosis system based on the physical-AI hybrid drive refers to an operation and maintenance technical mode of introducing an artificial intelligent algorithm to perform collaborative modeling and decision-making based on the physical mechanism and engineering constraint of the reverse osmosis process without purely relying on artificial experience or a black-box artificial intelligent model in the operation and management process of a Reverse Osmosis (RO) system and an industrial control system. The method is characterized by comprising the steps of firstly carrying out physical consistency check and standardization treatment on collected operation data such as pressure, flow, conductivity and the like by utilizing engineering standards such as mass conservation, hydraulic relations and ASTMD4516, ensuring that the data entering a model truly reflects the performance state of the membrane, forming an interpretable performance datum line by using a physical mechanism model on the basis, and carrying out trend correction and short-term prediction on nonlinear and time-varying characteristics such as membrane pollution and the like by using a deep learning model, thereby realizing intelligent decision on cleaning time, energy consumption control and medicament addition. The hybrid driving mode essentially uses a physical law to constrain the reasoning boundary of AI, uses AI to make up the problem of insufficient response of the traditional mechanism model to complex pollution evolution and operation disturbance, and converts the operation and maintenance post-operation response and experience judgment of the reverse osmosis system and the industrial control system into a prospective, verifiable and interpretable intelligent operation and maintenance system based on the real physical state and the prediction result. The prior art has the following defects: In the prior art, when the operation and maintenance decision of the reverse osmosis system is carried out based on the cooperation of the physical model and the artificial intelligent model, the physical model is continuously updated according to the operation parameters such as pressure, flow, conductivity and the like acquired in real time and in a fixed time step or event triggering mode, so as to reflect the instant operation state of the system, and the artificial intelligent model outputs a trend judgment result according to a preset reasoning period based on historical time sequence data. Under the dynamic operation scene of rapid load change of the reverse osmosis system, such as frequent start-stop of a high-pressure pump, adjustment of recovery rate or abrupt change of water quality of inflow, the inconsistency of a physical model and an artificial intelligent model in update frequency and reasoning rhythm easily causes multiple sets of mutually deviated state judgment results in the system at the same moment. When the outputs of the different models are used for generating operation control or operation and maintenance suggestions at the same time without uniform time sequence alignment and consistency constraint, the situation that control decisions are mutually contradictory easily occurs in the prior art, and further the problems of frequent switching, execution conflict and even misleading of manual intervention of operation and maintenance instructions are caused, so that the stability of the operation of the reverse osmosis system and the reliability of operation and maintenance decisions are affected. The above information disclosed in the background section is only for enhancement of understanding of the background of the disclosure and therefore it may include information that does not form the prior art that is already known to a person of ordinary skill in the art. Disclosure of Invention The invention aims to provide an intelligent operation and maintenance system and method of a reverse osmosis system based on physical-AI hybrid drive, so as to solve the problems in the background technology. In order to achieve the purpose, the invention provides the following technical scheme that the reverse osmosis system intelligent operation and maintenance method based on physical-AI hybrid drive comprises the following steps: Collecting the change data of the high-pressure pump frequency, the water inlet pressure and the water production flow rate along with time in the operation process of the reverse osmosis system, uniformly mapping all the operation data to the same time scale, and establishing an operat