Intelligent workshop equipment operation and maintenance management system
Patent Information
- Application Number
- CN202411147151.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-20
- Publication Date
- 2024-11-26
AI Technical Summary
During the water purification process, membrane filtration equipment is susceptible to factors such as contamination, pressure differential changes and chemical erosion, resulting in equipment failure and performance degradation. The existing fault prediction methods cannot fully consider these complex factors, and the prediction accuracy is not high and cannot be achieved. Real-time monitoring and early warning.
An intelligent workshop equipment operation and maintenance management system was designed, using data transmission module, data processing module, membrane filter equipment fault prediction module and real-time early warning module. Data was collected through flow sensors, high-frequency pressure differential sensors and chemical composition sensors, and using dual The Gigabit network and edge computing nodes are localized, combined with deep learning models (such as LSTM models) to predict soiling, pressure differential changes and chemical erosion, and send early warning information through the IPv6 network.
Real-time, high-frequency data acquisition and rapid transmission of membrane filtration equipment is realized, the timeliness and accuracy of data is improved, the accuracy and reliability of fault prediction is significantly improved, and the operation and maintenance personnel are notified in a timely manner for maintenance, reducing the equipment Failure rate and maintenance cost.
Smart Images

Figure CN119024699A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of workshop equipment operation and maintenance management, and specifically to an intelligent workshop equipment operation and maintenance management system. Background Art
[0002] With the rapid development of information technology and the continuous growth of the global economy, enterprises have increasingly higher requirements for production efficiency and quality. In recent years, the application trend of intelligence and automation in industrial production has become increasingly prominent, especially the introduction of dual-gigabit + IPv6 technology, which has brought rapid development to fields such as intelligent manufacturing and the Internet of Things.
[0003] In the water purification industry, the application of dual-gigabit + IPv6 technology in the operation and maintenance management of smart workshop equipment is particularly effective. Through the high bandwidth and low latency characteristics of the dual-gigabit + IPv6 network, all equipment in the smart workshop can achieve real-time data transmission and efficient collaborative operation, which not only greatly improves the efficiency of equipment operation and maintenance management, but also significantly reduces equipment failure rate and maintenance costs.
[0004] At present, membrane filtration equipment plays a core role in the water purification process. However, due to its complex working environment and operating parameters, it is easily affected by factors such as fouling, pressure difference changes, and chemical erosion, which lead to equipment failure and performance degradation. The existing fault prediction method cannot fully consider these complex factors, the prediction accuracy is not high, and it is impossible to monitor and warn in real time; therefore, there is an urgent need for an intelligent workshop equipment operation and maintenance management system to solve such problems. Summary of the invention
[0005] In view of the shortcomings of the prior art, the present invention provides an intelligent workshop equipment operation and maintenance management system to solve the problems in the prior art that the equipment is easily affected by factors such as pollution blockage, pressure difference changes, chemical corrosion, etc., which lead to equipment failure and performance degradation. The existing fault prediction method cannot fully consider these complex factors, the prediction accuracy is not high, and real-time monitoring and early warning are impossible.
[0006] To achieve the above objectives, on the one hand, the present invention provides an intelligent workshop equipment operation and maintenance management system, comprising:
[0007] The data transmission module is responsible for collecting parameter data from the membrane filtration equipment and transmitting it to the data processing terminal through the dual Gigabit network;
[0008] The data processing module uses edge computing nodes to perform local data processing, including feature extraction and preliminary analysis, integrates the processed data and analysis results into structured data packets, and transmits them to the central management system;
[0009] Membrane filtration equipment fault prediction module, used to predict the fouling, pressure difference change and chemical erosion of membrane filtration equipment;
[0010] Real-time warning module: When the deep learning model predicts a failure risk, it sends warning information to the central management system via the IPv6 network to remind operation and maintenance personnel to perform timely maintenance.
[0011] In another aspect, the present invention provides a method for operation and maintenance management of intelligent workshop equipment, comprising:
[0012] Step S1. Data collection and transmission,
[0013] Use flow sensors and high-frequency differential pressure sensors to monitor flow changes in membrane filtration equipment and the pressure difference on both sides of the membrane;
[0014] Use chemical composition sensors to detect the concentration of chemicals in water to monitor the impact of chemical corrosion on equipment;
[0015] A dual Gigabit network is used for sensor detection data transmission, and each sensor used for data collection is equipped with a unique IPv6 address;
[0016] Step S2. Data processing and edge computing,
[0017] Local data processing is performed by deploying edge computing nodes near the membrane filtration equipment;
[0018] Localization processing is feature extraction and preliminary analysis;
[0019] Deploy edge computing nodes near the equipment to achieve localized data processing and preliminary analysis;
[0020] Step S3. Deep learning model prediction,
[0021] The deep learning model is deployed in the central management system, and the deep learning model is used to predict the fouling, pressure difference change and chemical erosion of membrane filtration equipment;
[0022] The long short-term memory network (LSTM) model is used to accurately predict fouling, pressure difference changes, and chemical erosion through time series data analysis and fault prediction.
[0023] Step S4. Real-time warning,
[0024] When the deep learning model predicts a risk of failure, it immediately sends warning information to the central management system via the IPv6 network;
[0025] The present invention is further configured such that the optical fiber access device is deployed near the membrane filtration device and connected to the data acquisition terminal of the sensor;
[0026] Arrange fiber optic lines to cover all sensor installation locations;
[0027] Assign a unique IPv6 address to each sensor and data acquisition terminal and perform IPv6 address configuration;
[0028] The sensor collects flow, pressure difference and chemical composition data in real time and sends it to the data collection terminal via IPv6 address;
[0029] The data acquisition terminal performs preliminary processing on the sensor data, including denoising and data standardization;
[0030] The data collection terminal transmits the processed data to the edge computing node in real time through a dual gigabit network (gigabit optical network and 5G network);
[0031] By using flow sensors, high-frequency differential pressure sensors and chemical composition sensors, combined with dual gigabit networks and unique IPv6 addresses, real-time, high-frequency data collection and rapid transmission are achieved, greatly improving the timeliness and accuracy of data;
[0032] The present invention is further configured such that the edge computing device is deployed near the membrane filtration device, and the edge computing device is connected to the sensor and the data acquisition terminal via a dual gigabit network;
[0033] Edge computing nodes receive real-time data transmitted from sensors and data acquisition terminals;
[0034] A storage module is configured on the edge computing device to temporarily store the received data;
[0035] Calculate key characteristic parameters in real time at the edge computing node, namely the flow rate change rate, pressure difference fluctuation amplitude, and characteristic parameters of chemical component concentration change, including sliding window average, standard deviation, and trend change;
[0036] The present invention is further configured such that the preliminary analysis on the edge computing node includes:
[0037] Identify potential anomalies and integrate processed data and preliminary analysis results into structured data packages;
[0038] Abnormal conditions include: sudden drop in flow, sudden increase in pressure difference, and excessive chemical composition;
[0039] The locally processed data is transmitted to the central management system via a dual-gigabit network (gigabit optical network and 5G network) for further analysis and prediction;
[0040] The present invention is further configured such that the model input includes the structured data packet in step 2;
[0041] The long short-term memory network LSTM model is used for time series data analysis and fault prediction. The model prediction method is:
[0042] Predict fouling through the changing trends of flow and differential pressure data, and identify and predict fouling of membrane filtration equipment;
[0043] And analyze the pressure difference sensor data, detect and predict the pressure difference changes in real time, and warn of possible failures;
[0044] Monitor and predict the impact of chemical erosion on membrane filtration equipment through chemical composition sensor data, and provide early warning;
[0045] The present invention is further configured to predict fouling through the change trend of flow rate and pressure difference data in the following manner:
[0046] Extract features from structured data packets: flow data F(t), pressure difference data P(t), temperature data T(t), chemical composition data C(t);
[0047] Calculate the flow rate change rate ΔF / Δt: Where F(t) represents the flow rate at the current moment, F(t-1) represents the flow rate at the previous moment, and Δt represents the time interval;
[0048] Calculate the pressure difference change rate ΔP / Δt: Where P(t) represents the pressure difference at the current moment, and P(t-1) represents the pressure difference at the previous moment;
[0049] Calculate the weighted combination of flow and pressure difference WF,P: Among them, α and β are weight coefficients, indicating the importance of flow rate change and pressure difference change to fouling prediction;
[0050] Normalized feature X norm : Here X is the original feature, μ X represents the characteristic mean, σ X is the characteristic standard deviation;
[0051] Build a time series data set, define the time window T, and divide the standardized feature data into multiple time windows: X T =[X(t),X(t-1),...,X(t-T+1)], where X T Represents the feature data within the time window, T is the time window size;
[0052] Use the standardized time series data set to train the LSTM model. The model training process includes forward propagation and back propagation, and the model parameters are adjusted by minimizing the loss function.
[0053] Use LSTM model for prediction;
[0054] The present invention is further configured that the LSTM model prediction step includes: c =Mp (XT), where M p Represents the model prediction process, y c represents the predicted probability of fouling;
[0055] Calculate the fouling risk score R c : Among them, γ and δ are weight coefficients, reflecting the contribution of the prediction results and pressure difference changes to the risk of fouling. c is the predicted probability of fouling, Indicates the rate of change of pressure difference;
[0056] Set the warning threshold θ c , when R c When the threshold is exceeded, an early warning is triggered;
[0057] The present invention is further configured such that the LSTM model structure includes:
[0058] Input tensor: Here n samples represents the number of samples, T is the time window size, n features Indicates feature proficiency;
[0059] Forget Gate: f t is the output vector of the forget gate, W f represents the weight matrix of the forget gate, h t-1 is the hidden state vector at the previous moment, x t is the input vector at the current moment, b f is the bias vector of the forget gate, σ is the Sigmoid activation function;
[0060] Input Gate: i t is the output vector of the input gate, W i is the weight matrix of the input gate, b i is the bias vector of the input gate,
[0061] Candidate memory status: is the candidate memory state vector; W C The weight matrix representing the candidate memory states, b C is the bias vector of the candidate memory state, tanh is the hyperbolic tangent activation function;
[0062] Memory status update: C t is the memory state vector at the current moment, C t-1 is the memory state vector of the previous moment;
[0063] Output Gate: o t is the output vector of the output gate, W o is the weight matrix of the output gate, b o is the bias vector of the output gate,
[0064] Hide status update:h t =o t tanh(C t ), h t Represents the hidden state vector at the current moment;
[0065] In the LSTM model structure, n hldden represents the number of hidden layer units of the LSTM unit, and T is the time window length;
[0066] The present invention is further configured to predict the pressure difference change mode consistent with the fouling condition of the membrane filtration device, and also uses the pre-trained LSTM model for prediction:
[0067] After calculating the pressure difference change rate, pressure difference fluctuation amplitude, and average pressure difference, a time series data set is constructed;
[0068] Then, pressure difference risk assessment and early warning are performed based on the characteristic data within the time window;
[0069] The ways to monitor and predict the impact of chemical attack on membrane filtration equipment are:
[0070] Calculate the chemical composition change rate, maximum change range, and chemical composition concentration;
[0071] Build a time series data set and divide the standardized feature data into multiple time windows;
[0072] Train the LSTM model, preset the chemical erosion risk threshold, and use the LSTM model for prediction:
[0073] The LSTM-based time series analysis model accurately captures the changing trends in equipment operation and provides early warning of possible failures.
[0074] Compared with the prior art, the present invention has the following beneficial effects:
[0075] The present invention uses flow sensors, high-frequency differential pressure sensors and chemical composition sensors, combined with dual gigabit networks and unique IPv6 addresses, to achieve real-time, high-frequency data collection and rapid transmission, greatly improving the timeliness and accuracy of data; deploys edge computing nodes near the equipment to achieve localized data processing and preliminary analysis, and can calculate key characteristic parameters in real time, improving the efficiency and response speed of data processing; adopts a long short-term memory network LSTM model, through time series data analysis and fault prediction, to achieve accurate prediction of fouling, pressure difference changes and chemical erosion, significantly improving the accuracy and reliability of prediction; at the same time, sets an early warning threshold, and immediately sends early warning information to the central management system through the IPv6 network, and warns before a fault occurs, and can promptly notify the operation and maintenance personnel to perform maintenance, improving the timeliness and effectiveness of the early warning; through the combination of dual gigabit networks, high-bandwidth, low-latency, reliable data transmission is achieved, ensuring the stability of data transmission and processing;
[0076] The present invention solves the problems existing in the prior art that the equipment is easily affected by factors such as fouling, pressure difference change, chemical erosion, etc., which lead to equipment failure and performance degradation. The existing fault prediction method cannot fully consider these complex factors, the prediction accuracy is not high, and real-time monitoring and early warning are impossible. BRIEF DESCRIPTION OF THE DRAWINGS
[0077] Figure 1 It is a structural schematic diagram of the intelligent workshop equipment operation and maintenance management system of the present invention. DETAILED DESCRIPTION
[0078] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0079] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0080] See also Figure 1 As shown, it is a schematic diagram of the structure of the intelligent workshop equipment operation and maintenance management system based on this embodiment, and the system includes: a data transmission module, which is responsible for collecting parameter data from the membrane filtration equipment and transmitting it to the data processing terminal through a dual gigabit network;
[0081] The data processing module uses edge computing nodes to perform local data processing, including feature extraction and preliminary analysis, integrates the processed data and analysis results into structured data packets, and transmits them to the central management system;
[0082] Membrane filtration equipment fault prediction module, used to predict the fouling, pressure difference change and chemical erosion of membrane filtration equipment;
[0083] Real-time warning module: When the deep learning model predicts a failure risk, it sends warning information to the central management system via the IPv6 network to remind operation and maintenance personnel to perform timely maintenance.
[0084] The present invention also provides a method for operation and maintenance management of intelligent workshop equipment, the method comprising:
[0085] Step S1. Data collection and transmission,
[0086] Use flow sensors and high-frequency differential pressure sensors to monitor flow changes in membrane filtration equipment and the pressure difference on both sides of the membrane;
[0087] Use chemical composition sensors to detect the concentration of chemicals in water to monitor the impact of chemical corrosion on equipment;
[0088] A dual Gigabit network is used for sensor detection data transmission, and each sensor used for data collection is equipped with a unique IPv6 address;
[0089] The optical fiber access equipment is deployed near the membrane filtration equipment and connected to the data acquisition terminal of the sensor;
[0090] Arrange fiber optic lines to cover all sensor installation locations;
[0091] Assign a unique IPv6 address to each sensor and data acquisition terminal and perform IPv6 address configuration;
[0092] The sensor collects flow, pressure difference and chemical composition data in real time and sends it to the data collection terminal via IPv6 address;
[0093] The data acquisition terminal performs preliminary processing on the sensor data, including denoising and data standardization;
[0094] The data collection terminal transmits the processed data to the edge computing node in real time through a dual gigabit network (gigabit optical network and 5G network);
[0095] By using flow sensors, high-frequency differential pressure sensors and chemical composition sensors, combined with dual gigabit networks and unique IPv6 addresses, real-time, high-frequency data collection and rapid transmission are achieved, greatly improving the timeliness and accuracy of data;
[0096] Specifically, in step 1, the key parameters of the membrane filtration equipment are monitored in real time through flow sensors, high-frequency differential pressure sensors, and chemical composition sensors; each sensor is equipped with a unique IPv6 address, and data is transmitted through a dual-gigabit network to ensure high bandwidth and low latency for data transmission;
[0097] Step S2. Data processing and edge computing,
[0098] Local data processing is performed by deploying edge computing nodes near the membrane filtration equipment;
[0099] Localization processing is feature extraction and preliminary analysis;
[0100] Deploy edge computing nodes near the equipment to achieve localized data processing and preliminary analysis;
[0101] The edge computing device is deployed near the membrane filtration equipment, and the edge computing device is connected to the sensor and data acquisition terminal through a dual gigabit network;
[0102] Edge computing nodes receive real-time data transmitted from sensors and data acquisition terminals;
[0103] A storage module is configured on the edge computing device to temporarily store the received data;
[0104] Calculate key characteristic parameters in real time at the edge computing node, namely the flow rate change rate, pressure difference fluctuation amplitude, and characteristic parameters of chemical component concentration change, including sliding window average, standard deviation, and trend change;
[0105] The initial analysis on the edge computing node includes:
[0106] Identify potential anomalies and integrate processed data and preliminary analysis results into structured data packages;
[0107] Abnormal conditions include: sudden drop in flow, sudden increase in pressure difference, and excessive chemical composition;
[0108] The locally processed data is transmitted to the central management system via a dual-gigabit network (gigabit optical network and 5G network) for further analysis and prediction;
[0109] Specifically, the edge computing nodes deployed in step 2 can quickly identify potential abnormal conditions, including sudden drops in flow, sudden increases in pressure differences, and excessive chemical composition, and integrate the processed data into structured data packets for transmission to the central management system; the localized processing method greatly reduces the delay in data transmission and improves the efficiency and real-time performance of data processing;
[0110] Step S3. Deep learning model prediction,
[0111] The deep learning model is deployed in the central management system, and the deep learning model is used to predict the fouling, pressure difference change and chemical erosion of membrane filtration equipment;
[0112] The long short-term memory network (LSTM) model is used to accurately predict fouling, pressure difference changes, and chemical erosion through time series data analysis and fault prediction.
[0113] The model input includes the structured data package in step 2;
[0114] The long short-term memory network LSTM model is used for time series data analysis and fault prediction. The model prediction method is:
[0115] Predict fouling through the changing trends of flow and differential pressure data, and identify and predict fouling of membrane filtration equipment;
[0116] And analyze the pressure difference sensor data, detect and predict the pressure difference changes in real time, and warn of possible failures;
[0117] Monitor and predict the impact of chemical erosion on membrane filtration equipment through chemical composition sensor data, and provide early warning;
[0118] The method of predicting fouling through the changing trend of flow and pressure difference data is as follows:
[0119] Extract features from structured data packets: flow data F(t), pressure difference data P(t), temperature data T(t), chemical composition data C(t);
[0120] Calculate the flow rate change rate ΔF / Δt: Where F(t) represents the flow rate at the current moment, F(t-1) represents the flow rate at the previous moment, and Δt represents the time interval;
[0121] Calculate the pressure difference change rate ΔP / Δt: Where P(t) represents the pressure difference at the current moment, and P(t-1) represents the pressure difference at the previous moment;
[0122] Calculate the weighted combination of flow and pressure difference W F,P : Among them, α and β are weight coefficients, indicating the importance of flow rate change and pressure difference change to fouling prediction;
[0123] Normalized feature X norm : Here X is the original feature, μ X represents the characteristic mean, σ X is the characteristic standard deviation;
[0124] Build a time series data set, define the time window T, and divide the standardized feature data into multiple time windows: X T =[X(t),X(t-1),...,X(t-T+1)], where X T Represents the feature data within the time window, T is the time window size;
[0125] Use the standardized time series data set to train the LSTM model. The model training process includes forward propagation and back propagation, and the model parameters are adjusted by minimizing the loss function.
[0126] Use LSTM model for prediction;
[0127] The LSTM model prediction steps include: c =M p (X T ), where M p Represents the model prediction process, y c represents the predicted probability of fouling;
[0128] Calculate the fouling risk score R c : Among them, γ and δ are weight coefficients, reflecting the contribution of the prediction results and pressure difference changes to the risk of fouling. c is the predicted probability of fouling, Indicates the rate of change of pressure difference;
[0129] Set the warning threshold θ c , when R c When the threshold is exceeded, an early warning is triggered;
[0130] The LSTM model structure includes:
[0131] Input tensor: Here n samples represents the number of samples, T is the time window size, n features Indicates feature proficiency;
[0132] Forget Gate: f t is the output vector of the forget gate, W f represents the weight matrix of the forget gate, h t-1 is the hidden state vector at the previous moment, xt is the input vector at the current moment, b f is the bias vector of the forget gate, σ is the Sigmoid activation function;
[0133] Input Gate: i t is the output vector of the input gate, W i is the weight matrix of the input gate, b i is the bias vector of the input gate,
[0134] Candidate memory status: is the candidate memory state vector; W C The weight matrix representing the candidate memory states, b C is the bias vector of the candidate memory state, tanh is the hyperbolic tangent activation function;
[0135] Memory status update: C t is the memory state vector at the current moment, C t-1 is the memory state vector of the previous moment;
[0136] Output Gate: o t is the output vector of the output gate, W o is the weight matrix of the output gate, b o is the bias vector of the output gate,
[0137] Hide status update:h t =o t tanh(C t ), h t Represents the hidden state vector at the current moment;
[0138] In the LSTM model structure, n hldden represents the number of hidden layer units of the LSTM unit, and T is the length of the time window;
[0139] The predicted pressure difference change pattern is consistent with the fouling of the membrane filtration equipment, and the pre-trained LSTM model is also used for prediction:
[0140] After calculating the pressure difference change rate, pressure difference fluctuation amplitude, and average pressure difference, a time series data set is constructed;
[0141] Then, pressure difference risk assessment and early warning are performed based on the characteristic data within the time window;
[0142] The ways to monitor and predict the impact of chemical attack on membrane filtration equipment are:
[0143] Calculate the chemical composition change rate, maximum change range, and chemical composition concentration;
[0144] Build a time series data set and divide the standardized feature data into multiple time windows;
[0145] Train the LSTM model, preset the chemical erosion risk threshold, and use the LSTM model for prediction:
[0146] The LSTM-based time series analysis model accurately captures the changing trends in equipment operation and warns of possible failures in advance;
[0147] Specifically, in step 3, the LSTM model deployed in the central management system effectively predicted the fouling, pressure difference change and chemical erosion of the membrane filtration equipment; extracted key features from the structured data packet to perform time series data analysis and fault prediction; effectively reduced the failure rate and maintenance cost of the equipment; especially in terms of fouling prediction, pressure difference change prediction and chemical erosion prediction;
[0148] Step S4. Real-time warning,
[0149] When the deep learning model predicts a risk of failure, it immediately sends warning information to the central management system via the IPv6 network;
[0150] Specifically, when the deep learning model predicts a failure risk in step 4, the system can immediately send warning information to the central management system through the IPv6 network. Operation and maintenance personnel can receive the warning information in a timely manner and take maintenance measures, which improves the timeliness of fault response and effectively avoids sudden failures of equipment.
[0151] The above contents are only for explaining the technical idea of the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.
Claims
1. An intelligent workshop equipment operation and maintenance management system, comprising a data transmission module, a data processing module, a membrane filtration equipment fault prediction module and a real-time warning module, characterized in that: The data transmission module is responsible for collecting parameter data from the membrane filtration equipment and transmitting it to the data processing terminal through the dual Gigabit network; The data processing module uses edge computing nodes to perform local data processing, including feature extraction and preliminary analysis, integrates the processed data and analysis results into structured data packets, and transmits them to the central management system; Membrane filtration equipment fault prediction module, used to predict the fouling, pressure difference change and chemical erosion of membrane filtration equipment; Real-time warning module: When the deep learning model predicts a failure risk, it sends warning information to the central management system via the IPv6 network to remind operation and maintenance personnel to perform timely maintenance.
2. A method for operation and maintenance management of intelligent workshop equipment, characterized in that: The intelligent workshop equipment operation and maintenance management system described in claim 1 is used, comprising the following steps: Step S1. Data collection and transmission, Use flow sensors and high-frequency differential pressure sensors to monitor flow changes in membrane filtration equipment and the pressure difference on both sides of the membrane; Use chemical composition sensors to detect the concentration of chemical substances in water; A dual Gigabit network is used for sensor detection data transmission, and each sensor used for data collection is equipped with a unique IPv6 address; Step S2. Data processing and edge computing, Local data processing is performed by deploying edge computing nodes near the membrane filtration equipment; Localization processing is feature extraction and preliminary analysis; Step S3. Deep learning model prediction, The deep learning model is deployed in the central management system, and the deep learning model is used to predict the fouling, pressure difference change and chemical erosion of membrane filtration equipment; Step S4. Real-time warning, When the deep learning model predicts a failure risk, it immediately sends an early warning message to the central management system via the IPv6 network.
3. The intelligent workshop equipment operation and maintenance management method according to claim 2 is characterized in that: The optical fiber access equipment is deployed near the membrane filtration equipment and connected to the data acquisition terminal of the sensor; Arrange fiber optic lines to cover all sensor installation locations; Assign a unique IPv6 address to each sensor and data acquisition terminal and perform IPv6 address configuration; The sensor collects flow, pressure difference and chemical composition data in real time and sends it to the data collection terminal via IPv6 address; The data acquisition terminal performs preliminary processing on the sensor data, including denoising and data standardization; The data acquisition terminal transmits the processed data to the edge computing node in real time through a dual gigabit network.
4. The intelligent workshop equipment operation and maintenance management method according to claim 3 is characterized in that: The edge computing device is deployed near the membrane filtration equipment, and the edge computing device is connected to the sensor and data acquisition terminal through a dual gigabit network; Edge computing nodes receive real-time data transmitted from sensors and data acquisition terminals; A storage module is configured on the edge computing device to temporarily store the received data; The key characteristic parameters, namely the flow rate change rate, pressure difference fluctuation amplitude, and characteristic parameters of chemical component concentration change, are calculated in real time at the edge computing node, including the sliding window average, standard deviation, and trend change.
5. The intelligent workshop equipment operation and maintenance management method according to claim 4 is characterized in that: The initial analysis on the edge computing node includes: Identify potential anomalies and integrate processed data and preliminary analysis results into structured data packages; Abnormal conditions include: sudden drop in flow, sudden increase in pressure difference, and excessive chemical composition; The locally processed data is transmitted to the central management system via a dual-gigabit network for further analysis and prediction.
6. A smart workshop equipment operation and maintenance management method according to claim 5, characterized in that: The model input includes the structured data package in step 2; The long short-term memory network LSTM model is used for time series data analysis and fault prediction. The model prediction method is: Predict fouling through the changing trends of flow and differential pressure data, and identify and predict fouling of membrane filtration equipment; And analyze the pressure difference sensor data, detect and predict the pressure difference changes in real time, and warn of possible failures; Through chemical composition sensor data, the impact of chemical erosion on membrane filtration equipment can be monitored and predicted, and early warning can be provided.
7. A smart workshop equipment operation and maintenance management method according to claim 6, characterized in that: The method of predicting fouling through the changing trend of flow and pressure difference data is as follows: Extract features from structured data packets: flow data F(t), pressure difference data P(t), temperature data T(t), chemical composition data C(t); Calculate the flow rate change rate ΔF / Δt: Where F(t) represents the flow rate at the current moment, F(t-1) represents the flow rate at the previous moment, and Δt represents the time interval; Calculate the pressure difference change rate ΔP / Δt: Where P(t) represents the pressure difference at the current moment, and P(t-1) represents the pressure difference at the previous moment; Calculate the weighted combination of flow and pressure difference W F,P : Among them, α and β are weight coefficients, indicating the importance of flow rate change and pressure difference change to fouling prediction; Normalized feature X norm : Here X is the original feature, μ X represents the characteristic mean, σ X is the characteristic standard deviation; Build a time series data set, define the time window T, and divide the standardized feature data into multiple time windows: X T =[X(t),X(t-1),...,X(t-T+1)], where X T Represents the feature data within the time window, T is the time window size; Use the standardized time series data set to train the LSTM model. The model training process includes forward propagation and back propagation, and the model parameters are adjusted by minimizing the loss function. Use LSTM model for prediction.
8. The intelligent workshop equipment operation and maintenance management method according to claim 7 is characterized in that: The LSTM model prediction steps include: c =M p (X T ), where M p Represents the model prediction process, y c represents the predicted probability of fouling; Calculate the fouling risk score R c : Among them, γ and δ are weight coefficients, reflecting the contribution of the prediction results and pressure difference changes to the risk of fouling. c is the predicted probability of fouling, Indicates the rate of change of pressure difference; Set the warning threshold θ c , when R c When the threshold is exceeded, an alert is triggered.
9. The intelligent workshop equipment operation and maintenance management method according to claim 8, characterized in that: The LSTM model structure includes: Input tensor: Here n samples represents the number of samples, T is the time window size, n features Indicates feature proficiency; Forget Gate: f t is the output vector of the forget gate, W f represents the weight matrix of the forget gate, h t-1 is the hidden state vector at the previous moment, x t is the input vector at the current moment, b f is the bias vector of the forget gate, σ is the Sigmoid activation function; Input Gate: i t is the output vector of the input gate, W i is the weight matrix of the input gate, b i is the bias vector of the input gate, Candidate memory status: is the candidate memory state vector; W C The weight matrix representing the candidate memory states, b C is the bias vector of the candidate memory state, tanh is the hyperbolic tangent activation function; Memory status update: C t is the memory state vector at the current moment, C t-1 is the memory state vector of the previous moment; Output Gate: o t is the output vector of the output gate, W o is the weight matrix of the output gate, b o is the bias vector of the output gate, Hide status update:h t =o t tanh(C t ), h t Represents the hidden state vector at the current moment; In the LSTM model structure, n hldden represents the number of hidden layer units of the LSTM unit, and T is the time window length.
10. A smart workshop equipment operation and maintenance management method according to claim 9, characterized in that: The predicted pressure difference change pattern is consistent with the fouling of the membrane filtration equipment, and the pre-trained LSTM model is also used for prediction: After calculating the pressure difference change rate, pressure difference fluctuation amplitude, and average pressure difference, a time series data set is constructed; Then, pressure difference risk assessment and early warning are performed based on the characteristic data within the time window; The ways to monitor and predict the impact of chemical attack on membrane filtration equipment are: Calculate the chemical composition change rate, maximum change range, and chemical composition concentration; Build a time series data set and divide the standardized feature data into multiple time windows; Train the LSTM model, preset the chemical corrosion risk threshold, and use the LSTM model for prediction.
