A method and system for evaluating hydrogen conductivity detection

Abstract

The present application belongs to the field of water and steam chemistry of thermal power plants, and relates to a hydrogen conductivity detection and evaluation method and system, the method comprising S1: continuously collecting hydrogen conductivity time series data in a hydrogen ion exchange column through a hydrogen conductivity instrument, simultaneously collecting resin color time series data, correlating and labeling the resin color time series data with the hydrogen conductivity time series data to form joint feature data; S2: constructing a neural network model based on time series, establishing a mapping relationship between hydrogen conductivity change trend and resin replacement timing and salt accumulation; S3: inputting the joint feature data into the trained neural network model based on time series, outputting resin failure remaining time and salt accumulation risk probability, judging the output warning type, and outputting operation guidance information according to the warning type; S4: after receiving the warning type and operation guidance information, performing corresponding operations.

Description

Technical Field

[0001] This invention relates to the field of water vapor chemistry in thermal power plants, specifically to a method and system for detecting and evaluating hydrogen conductivity. Background Technology

[0002] In the water-steam circulation system of a power plant, hydrogen conductivity is a key indicator for measuring the purity of water and steam and determining whether corrosive anions exist in the thermal system. To achieve accurate measurement of hydrogen conductivity, the water sample must first flow through a hydrogen ion exchange column filled with a strong acidic cation exchange resin.

[0003] Existing technologies employ various methods for detecting and evaluating hydrogen conductivity, including online monitoring of outlet hydrogen conductivity combined with a fixed threshold for failure alarms, or independent prediction of the remaining lifespan of the resin based on data-driven models such as neural networks. However, most of these methods rely on only one detection dimension, either simply analyzing the changing trend of hydrogen conductivity time-series data or evaluating the performance status of the ion exchange resin in isolation. They fail to organically combine the actual performance degradation process of the resin exchange column with the dynamic response of the hydrogen conductivity measurement, making it difficult for the evaluation results to truly reflect the coupling relationship between the two.

[0004] In view of this, it is very necessary for the present invention to provide a method and system for detecting and evaluating hydrogen conductivity in order to solve the above-mentioned defects in the prior art. Summary of the Invention

[0005] The purpose of this invention is to solve the technical problem in the prior art of hydrogen conductivity detection that it fails to organically combine the actual performance degradation process of the resin exchange column with the dynamic response of the hydrogen conductivity measurement value, making it difficult for the evaluation results to truly reflect the coupling relationship between the two. The invention provides a hydrogen conductivity detection and evaluation method and system to solve the technical problems existing in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for detecting and evaluating hydrogen conductivity, comprising the following steps: Step S1: Continuously collect hydrogen conductivity time series data in the hydrogen ion exchange column using a hydrogen conductivity meter, and simultaneously collect resin color time series data. Correlate and label the resin color time series data with the hydrogen conductivity time series data to form joint feature data, which includes hydrogen conductivity time series and resin color time series data. Step S2: Construct a time-series-based neural network model to establish a mapping relationship between the trend of hydrogen conductivity change and the timing of resin replacement and salt accumulation; Step S3: Input the joint feature data into the trained time series-based neural network model, output the predicted remaining time of resin failure and the probability of salt accumulation risk, determine the output warning type based on the preset resin failure threshold and the preset salt accumulation risk warning value, and output operation guidance information according to the warning type. Step S4: After receiving the warning type and operation guidance information, perform the corresponding operation; and after the operation is completed, store the joint feature data, replacement records and prediction results of this operation locally and upload them to the cloud server.

[0007] Furthermore, the present invention also provides a hydrogen conductivity detection and evaluation system, comprising: The joint feature generation module contains: Time-series data of hydrogen conductivity in a hydrogen ion exchange column were continuously collected using a hydrogen conductivity meter. The resin color change images are continuously acquired by an image acquisition device set on the outside of the transparent PC column, and resin color time series data is generated. The resin color time series data and the hydrogen conductivity time series data are associated and labeled with the same timestamp to form joint feature data; The temporal neural network modeling module contains: Construct a time-series-based neural network model; Using historical joint feature data as training samples, a time-series-based neural network model was trained to establish a mapping relationship between the trend of hydrogen conductivity change and the timing of resin replacement and salt accumulation. The early warning output module contains: The real-time collected joint feature data is input into the trained time series-based neural network model, which outputs the predicted remaining time of resin failure and the probability of salt accumulation risk. Based on the preset resin failure threshold and the preset salt accumulation risk warning value, determine the type of warning to be output, and match and output the corresponding operation guidance information according to the warning type. The operation execution module contains: After receiving the warning type and operation guidance information, perform the corresponding operation; After the operation is completed, record the operation, store it locally in a structured format, and upload it to the cloud server.

[0008] The beneficial effects of this invention are as follows: This invention achieves a joint feature data analysis by continuously collecting time-series data on hydrogen conductivity and resin color of hydrogen ion exchange columns. This analysis integrates the measured hydrogen conductivity with the actual performance degradation process of the resin, allowing the resin color change process to correspond to the dynamic fluctuations in hydrogen conductivity. This enhances the characterization of the resin's exchange performance degradation and improves the accuracy of resin replacement timing. Furthermore, by training a time-series-based neural network model on the historical joint feature data, a mapping relationship is established between the hydrogen conductivity change trend and the remaining time of resin failure and the probability of salt accumulation risk. This enables the prediction of the remaining time of resin failure and the probability of salt accumulation risk, reducing water vapor buildup caused by resin failure or abnormal salt accumulation. Quality fluctuation risk; based on prediction results and preset thresholds, the type of early warning is determined, and targeted operational guidance information is provided, forming a closed loop between hydrogen conductivity testing and resin operation management, improving the authenticity and reliability of the evaluation results for the actual operating status; by storing the hydrogen conductivity data after operation, resin running time, replacement records, and prediction results locally and uploading them to the cloud, data traceability and remote management capabilities are achieved, providing a foundation for subsequent operation optimization; by introducing resin color time series data into the hydrogen conductivity testing and evaluation process, a visualized and quantitative analysis of the resin operating status is achieved, avoiding the misjudgment problem caused by relying solely on a single hydrogen conductivity index, and improving the stability and reliability of the testing and evaluation results.

[0009] Therefore, it is evident that the present invention has outstanding substantive features and significant progress compared with the prior art, and the beneficial effects of its implementation are also obvious. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0011] Figure 1 This is a flowchart of a method for detecting and evaluating hydrogen conductivity. Figure 2 This is a block diagram of a hydrogen conductivity detection and evaluation system. Detailed Implementation

[0012] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. The following embodiments are explanations of the present invention, but the present invention is not limited to the following implementation methods.

[0013] Example 1: like Figure 1As shown, this embodiment provides a method for detecting and evaluating hydrogen conductivity, which includes the following steps: Step S1: Continuously collect hydrogen conductivity time series data in the hydrogen ion exchange column using a hydrogen conductivity meter, and simultaneously collect resin color time series data. Correlate and label the resin color time series data with the hydrogen conductivity time series data to form joint feature data, which includes hydrogen conductivity time series and resin color time series data. Step S1 specifically includes: S11. Continuously measure the outflow water sample in the hydrogen ion exchange column using a hydrogen conductivity meter, collect the measurement results in real time or periodically, and record the corresponding collection time information simultaneously to form a time series of hydrogen conductivity data arranged in chronological order, including hydrogen conductivity measurement value, water sample temperature data, and water sample flow rate data.

[0014] S12. The image acquisition device set on the outside of the transparent PC column continuously acquires images of the internal resin filling state, and acquires resin color change images in real time or periodically during operation to generate resin color time series data. The resin color change mentioned in step S12 includes the location where the resin gradually changes from light to dark, the color change range, whether the resin exhibits delamination or local failure, the resin color change nodes, and the cumulative running time of the resin.

[0015] S13. Match and label the hydrogen conductivity time series data with the resin color time series data to form joint feature data, and transmit it to the cloud server to analyze the resin's exchange performance degradation process and determine the resin failure process during operation.

[0016] It should be noted that step S1, through continuous acquisition of hydrogen conductivity, water sample temperature, and water sample flow rate data, combined with automatic image acquisition of resin color changes, achieves the fusion of operational data and physical state information, comprehensively reflecting the resin's working status and performance change trends. The joint feature data formed through time matching and associated annotation provides a reliable basis for accurate judgment of the resin failure process, offering high-quality data support for subsequent analysis and optimization, and improving operational stability and predictive capabilities.

[0017] Step S2: Construct a time-series-based neural network model to establish a mapping relationship between the trend of hydrogen conductivity change and the timing of resin replacement and salt accumulation; Step S2 specifically includes: S21. Construct a time-series-based neural network model in a cloud server. The time-series-based neural network model adopts a sequence modeling structure to perform time-series learning and nonlinear mapping on the trend of hydrogen conductivity change. The time-series-based neural network model includes an input layer, a time-series feature extraction layer, and an output layer.

[0018] The time-series feature extraction layer in step S21 employs a Long Short-Term Memory (LSTM) network structure to model long-term dependencies in historical time-series data. The input layer receives historical joint feature data, and the output layer outputs the predicted results of the remaining resin failure time and salt accumulation risk level. The historical joint feature data includes historical hydrogen conductivity time-series data and historical resin color time-series data.

[0019] By constructing a time-series-based neural network model on a cloud server and using a long short-term memory network for time-series data learning, the long-term dependence of hydrogen conductivity changes and its relationship with the resin failure process can be effectively captured.

[0020] S22. Use historical joint feature data to train a time series-based neural network model and establish a mapping relationship between the trend of hydrogen conductivity change and the timing of resin replacement and salt accumulation.

[0021] Step S22 specifically includes: S221. Perform time alignment processing on the historical joint feature data, and map the historical hydrogen conductivity time series data and historical resin color time series data with different sampling frequencies to the same time scale to construct a time series sample set.

[0022] S222. Perform feature encoding on the time series sample set to construct the input feature vector. ,in, The value of hydrogen conductivity at time t; Let t be the temperature of the water sample at time t; Let t be the water sample flow rate at time t; Let t be the characteristic value of resin color observation at time t.

[0023] The trend of hydrogen conductivity variation was obtained by differential calculation of continuously collected hydrogen conductivity time series data; the hydrogen conductivity data at the current moment and within the previous time window were used to construct a time series. By performing difference calculations or sliding fit processing on this time series, the trend of hydrogen conductivity variation can be obtained. ;in, This represents the hydrogen conductivity value at the current moment, and k represents the length of the time window; S223. Input feature vector The input is fed into a time-series-based neural network model for forward computation to obtain the output result. ,in: The output of the model is f(·); f(·) is the nonlinear mapping function of the neural network. These are the model weight parameters; These are bias parameters; the output includes predicted remaining resin failure time. and the probability of salt accumulation ; S224. Construct a loss function to train and optimize the time-series-based neural network model. The loss function includes a resin failure remaining time prediction error term and a salt accumulation risk probability error term, and its expression is:

[0024] in, To predict the remaining time before resin failure; This represents the actual remaining time before the resin fails. To predict the probability of salt accumulation risk; denoted as the actual probability of salt accumulation risk; CE(·) is the cross-entropy loss function; α and β are weighting coefficients.

[0025] S225. The backpropagation algorithm is used to minimize the loss function and iteratively update the weight parameter W and bias parameter b of the time series-based neural network model so that the output result gradually approximates the real resin failure evolution law. S226. The trained time-series-based neural network model can learn the nonlinear mapping relationship between the hydrogen conductivity change trend and the output result, and then establish a mapping relationship between the hydrogen conductivity change trend and the remaining time of resin failure and the probability of salt accumulation risk, outputting the predicted value of the remaining time of resin failure corresponding to the hydrogen conductivity change trend. and the probability of salt accumulation .

[0026] By training a time-series-based neural network model using historical joint feature data, the model learns the intrinsic correlation between the trend of hydrogen conductivity variation and the remaining time of resin failure and the probability of salt accumulation risk, thus establishing a mapping relationship between the hydrogen conductivity variation trend and the remaining time of resin failure and the probability of salt accumulation risk. The training process enables the time-series-based neural network model to represent complex nonlinear relationships, improving the accuracy and reliability of the predicted remaining resin time and the judgment of the probability of salt accumulation risk.

[0027] It should be noted that step S2, by constructing a time-series-based neural network model and combining historical hydrogen conductivity time-series data and historical resin color time-series data, provides an effective tool for predicting the remaining time of resin failure and the risk of salt accumulation. By using the LSTM model and loss function optimization method, it can accurately model the complex nonlinear relationship between hydrogen conductivity and resin replacement timing, thereby optimizing operation and maintenance decisions, providing early warnings, and extending service life.

[0028] Step S3: Input the joint feature data into the trained time series-based neural network model, output the predicted remaining time of resin failure and the probability of salt accumulation risk, determine the output warning type based on the preset resin failure threshold and the preset salt accumulation risk warning value, and output operation guidance information according to the warning type. Step S3 specifically includes: S31. Input the current joint feature data into the trained time-series-based neural network model, and output the predicted remaining time of resin failure. and the probability of salt accumulation .

[0029] S32. Calculate the preset threshold for resin failure. Compared with the preset salt accumulation risk warning value The resin failure preset threshold is calculated as follows:

[0030] in: A preset threshold is set for resin failure. λ represents the historical average failure time of the resin; λ is the safety factor, 0 < λ < 1; The probability of salt accumulation risk is calculated as follows:

[0031] Where: σ is the Sigmoid function. For real-time hydrogen conductivity, The rate of change of hydrogen conductivity, For the water sample temperature, The water sample flow rate is, For weight parameters, .

[0032] S33. An alert is triggered when any of the following conditions are met: Triggering resin replacement warning or Triggering a salt accumulation warning; among which: To preset a warning value for the risk of salt accumulation; Common settings .

[0033] S34. Output operational guidance information according to the warning type. The operational guidance information includes corresponding measures for resin replacement warnings and corresponding measures for salt accumulation risk. The corresponding measures for resin replacement warning include: closing the inlet valve of the sampling pipeline; preparing to replace the hydrogen-type cation exchange resin; removing the lower end cap and replacing the resin using the threaded locking ring; and restoring the operating status.

[0034] Countermeasures to address the risk of salt accumulation include: checking water quality parameters of the water vapor system; adjusting water sampling flow rate or sampling stability; optimizing the operating parameters of the upstream water treatment system; and, if necessary, increasing the monitoring frequency to track the trend of hydrogen conductivity changes.

[0035] It should be noted that by jointly analyzing real-time hydrogen conductivity data and historical trend data, and combining the resin failure time prediction model and the salt accumulation risk probability model, the resin operating status and the risk of salt accumulation in the water vapor system can be predicted in advance, thereby enabling early warning-driven maintenance decisions.

[0036] Step S4: After receiving the warning type and operation guidance information, perform the corresponding operation; and after the operation is completed, store the joint feature data, replacement records and prediction results of this operation locally, and upload them to the cloud server for optimization of the time series-based neural network model and water vapor quality analysis.

[0037] Example 2: like Figure 2 As shown, this embodiment provides a hydrogen conductivity detection and evaluation system, comprising: The joint feature generation module contains: Time-series data of hydrogen conductivity in a hydrogen ion exchange column were continuously collected using a hydrogen conductivity meter. The resin color change images are continuously acquired by an image acquisition device set on the outside of the transparent PC column, and resin color time series data is generated. The resin color time series data and the hydrogen conductivity time series data are associated and labeled with the same timestamp to form joint feature data; The hydrogen conductivity time series data acquisition process in the joint feature generation module is as follows: The hydrogen conductivity meter is used to continuously measure the effluent water sample from the hydrogen ion exchange column online. The hydrogen conductivity measurement value, water sample temperature data and water sample flow rate data are recorded in real time or periodically, and the acquisition timestamp is added simultaneously to form hydrogen conductivity time series data. The acquisition frequency is set to be adjustable from once per minute to once per hour to ensure that short-term fluctuations and long-term trends of hydrogen conductivity can be captured.

[0038] The resin color time series data acquisition process in the joint feature generation module is as follows: The resin state inside the transparent PC column is continuously captured by an image acquisition device. Images of resin color changes are acquired in real time or periodically, and acquisition timestamps are added simultaneously to form resin color time series data. The resin color time series data includes the location of resin color change, color change range, resin stratification state, local failure area, and cumulative resin running time, which is used to reflect the change process of resin exchange performance.

[0039] The temporal neural network modeling module contains: Construct a time-series-based neural network model; Using historical joint feature data as training samples, a time-series-based neural network model was trained to establish a mapping relationship between the trend of hydrogen conductivity change and the timing of resin replacement and salt accumulation. The time-series-based neural network modeling module's construction process is as follows: A time-series-based neural network model was constructed on a cloud server. LSTM was used as the time-series feature extraction layer. The input layer received joint feature data after time alignment and feature encoding, and the output layer output the remaining time of resin failure and the probability of salt accumulation risk. The dimension of the time-series feature extraction layer was set to 64-128, and the time window length was set to 24-72 hours to adapt to the dynamic change cycle of the power plant's water-steam system.

[0040] The early warning output module contains: The real-time collected joint feature data is input into the trained time series-based neural network model, which outputs the predicted remaining time of resin failure and the probability of salt accumulation risk. Based on the preset resin failure threshold and the preset salt accumulation risk warning value, determine the type of warning to be output, and match and output the corresponding operation guidance information according to the warning type. The operation execution module contains: After receiving the warning type and operation guidance information, perform the corresponding operation; After the operation is completed, record the operation, store it locally in a structured format, and upload it to the cloud server.

[0041] Those skilled in the art will clearly understand that the techniques in the embodiments of the present invention can be implemented using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a medium such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, or any other medium capable of storing program code. It includes several instructions to cause a computer device (which may be a personal computer, a server, or a second device, network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0042] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

[0043] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or modules may be electrical, mechanical, or other forms.

[0044] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0045] In addition, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0046] The above-disclosed embodiments are merely preferred embodiments of the present invention, but the present invention is not limited thereto. Any non-creative variations that can be conceived by those skilled in the art, as well as any improvements and modifications made without departing from the principles of the present invention, should fall within the protection scope of the present invention.

Claims

1. A method for detecting and evaluating hydrogen conductivity, characterized in that, Includes the following steps: Step S1: Continuously collect hydrogen conductivity time series data in the hydrogen ion exchange column using a hydrogen conductivity meter, and simultaneously collect resin color time series data. Correlate and label the resin color time series data with the hydrogen conductivity time series data to form joint feature data, which includes hydrogen conductivity time series and resin color time series data. Step S2: Construct a time-series-based neural network model to establish a mapping relationship between the trend of hydrogen conductivity change and the timing of resin replacement and salt accumulation; Step S3: Input the joint feature data into the trained time series-based neural network model, output the predicted remaining time of resin failure and the probability of salt accumulation risk, determine the output warning type based on the preset resin failure threshold and the preset salt accumulation risk warning value, and output operation guidance information according to the warning type. Step S4: After receiving the warning type and operation guidance information, perform the corresponding operation; After the operation is completed, the joint feature data, replacement records and prediction results of this operation are stored locally and then uploaded to the cloud server.

2. The method for detecting and evaluating hydrogen conductivity according to claim 1, characterized in that, Step S1 specifically includes: S11. Continuously measure the effluent water sample in the hydrogen ion exchange column using a hydrogen conductivity detector, collect the measurement results in real time or periodically, and record the corresponding collection time information synchronously to form hydrogen conductivity time series data arranged in chronological order; S12. The image acquisition device set on the outside of the transparent PC column continuously acquires images of the internal resin filling state, and acquires resin color change images in real time or periodically during operation to generate resin color time series data. S13. Match and label the hydrogen conductivity time series data with the resin color time series data to form joint feature data, and transmit it to the cloud server.

3. The method for detecting and evaluating hydrogen conductivity according to claim 2, characterized in that, The hydrogen conductivity time series data includes hydrogen conductivity measurements, water sample temperature data, and water sample flow rate data.

4. The method for detecting and evaluating hydrogen conductivity according to claim 3, characterized in that, Step S2 specifically includes: S21. Construct a time-series-based neural network model in a cloud server. The time-series-based neural network model adopts a sequence modeling structure to perform time-series learning and nonlinear mapping on the trend of hydrogen conductivity change. The time-series-based neural network model includes an input layer, a time-series feature extraction layer, and an output layer. S22. Use historical joint feature data to train a time series-based neural network model and establish a mapping relationship between the trend of hydrogen conductivity change and the timing of resin replacement and salt accumulation. The historical joint feature data includes historical hydrogen conductivity time series data and historical resin color time series data.

5. The method for detecting and evaluating hydrogen conductivity according to claim 4, characterized in that, The time-series feature extraction layer adopts a Long Short-Term Memory (LSTM) network structure to model long-term dependencies in historical time-series data; the input layer is used to receive historical joint feature data; and the output layer is used to output the prediction results of the remaining resin failure time and salt accumulation risk level.

6. The method for detecting and evaluating hydrogen conductivity according to claim 5, characterized in that, Step S22 specifically includes: S221. Perform time alignment processing on the historical joint feature data, and map the historical hydrogen conductivity time series data and historical resin color time series data with different sampling frequencies to the same time scale to construct a time series sample set; S222. Perform feature encoding on the time series sample set to construct the input feature vector. ,in, The value of hydrogen conductivity at time t; Let t be the temperature of the water sample at time t; Let t be the water sample flow rate at time t; The characteristic value of resin color observation at time t; S223. Input feature vector The input is fed into a time-series-based neural network model for forward computation to obtain the output result. ,in: The output of the model is f(·); f(·) is the nonlinear mapping function of the neural network. These are the model weight parameters; These are bias parameters; the output includes predicted remaining resin failure time. and the probability of salt accumulation ; S224. Construct a loss function to train and optimize the time-series-based neural network model. The loss function includes a resin failure remaining time prediction error term and a salt accumulation risk probability error term, and its expression is: in, To predict the remaining time before resin failure; This represents the actual remaining time before the resin fails. To predict the probability of salt accumulation risk; The actual risk probability of salt accumulation; CE(·) is the cross-entropy loss function; α and β are weighting coefficients; S225. The backpropagation algorithm is used to minimize the loss function and iteratively update the weight parameter W and bias parameter b of the time series-based neural network model so that the output results gradually approximate the real resin failure evolution law. S226. The trained time-series-based neural network model can establish a mapping relationship between the trend of hydrogen conductivity change and the remaining time of resin failure and the probability of salt accumulation risk, and output the predicted value of the remaining time of resin failure corresponding to the trend of hydrogen conductivity change. and the probability of salt accumulation .

7. The method for detecting and evaluating hydrogen conductivity according to claim 6, characterized in that, The trend of hydrogen conductivity variation is obtained by differential calculation of continuously collected hydrogen conductivity time series data; the hydrogen conductivity data at the current moment and within the previous time window are used to construct a time series: By performing difference calculations or sliding fit processing on this time series, the trend of hydrogen conductivity variation can be obtained. ;in, This represents the hydrogen conductivity value at the current moment, and k represents the length of the time window.

8. The method for detecting and evaluating hydrogen conductivity according to claim 7, characterized in that, Step S3 specifically includes: S31. Input the current joint feature data into the trained time-series-based neural network model, and output the predicted remaining time of resin failure. and the probability of salt accumulation ; S32. Calculate the preset threshold for resin failure. Compared with the preset warning value for salt accumulation risk The resin failure preset threshold is calculated as follows: in: A preset threshold is set for resin failure. λ represents the historical average failure time of the resin; λ is the safety factor, 0 < λ < 1; The probability of salt accumulation risk is calculated as follows: Where: σ is the Sigmoid function. For real-time hydrogen conductivity, The rate of change of hydrogen conductivity, For the water sample temperature, The water sample flow rate is, For weight parameters, ; S33. An alert is triggered when any of the following conditions are met: Triggering resin replacement warning or Triggering a salt accumulation warning; among which: To preset a warning value for the risk of salt accumulation; Common settings ; S34. Output operational guidance information according to the warning type. The operational guidance information includes corresponding measures for resin replacement warnings and corresponding measures for salt accumulation risk.

9. The method for detecting and evaluating hydrogen conductivity according to claim 8, characterized in that, The corresponding measures for the resin replacement early warning include: closing the inlet valve of the sampling pipeline; preparing to replace the hydrogen-form cation exchange resin; removing the lower end cap and replacing the resin using the threaded locking ring; and restoring the operating status. Countermeasures to address the risk of salt accumulation include: checking water quality parameters of the water vapor system; adjusting water sampling flow rate or sampling stability; optimizing the operating parameters of the upstream water treatment system; and increasing monitoring frequency to track the trend of hydrogen conductivity changes.

10. A hydrogen conductivity detection and evaluation system, characterized in that, include: The joint feature generation module contains: Time-series data of hydrogen conductivity in a hydrogen ion exchange column were continuously collected using a hydrogen conductivity meter. The resin color change images are continuously acquired by an image acquisition device set on the outside of the transparent PC column, and resin color time series data is generated. The resin color time series data and the hydrogen conductivity time series data are associated and labeled with the same timestamp to form joint feature data; The temporal neural network modeling module contains: Construct a time-series-based neural network model; Using historical joint feature data as training samples, a time-series-based neural network model was trained to establish a mapping relationship between the trend of hydrogen conductivity change and the timing of resin replacement and salt accumulation. The early warning output module contains: The real-time collected joint feature data is input into the trained time series-based neural network model, which outputs the predicted remaining time of resin failure and the probability of salt accumulation risk. Based on the preset resin failure threshold and the preset salt accumulation risk warning value, determine the type of warning to be output, and match and output the corresponding operation guidance information according to the warning type. The operation execution module contains: After receiving the warning type and operation guidance information, perform the corresponding operation; After the operation is completed, record the operation, store it locally in a structured format, and upload it to the cloud server.

Images