A method for equipment corrosion assessment and life prediction and its application

Abstract

This invention relates to the field of equipment monitoring technology, and in particular to a method and application for equipment corrosion assessment and lifespan prediction, comprising the following steps: deploying a sensor network in the easily corroded areas of coal chemical equipment to collect multi-dimensional data; preprocessing the collected multi-source data by outlier removal, data compression, time synchronization, and spatiotemporal alignment; extracting image features using a convolutional neural network, analyzing the processed data using an LSTM-attention model, and establishing a fuzzy comprehensive evaluation matrix to assess the corrosion level; constructing a physical model based on Faraday's law of electrolysis and a data-driven model based on a Transformer network, fusing the output through a Bayesian network, and combining it with Monte Carlo simulation to predict the remaining lifespan. This invention, by integrating multi-source data with intelligent algorithms, achieves accurate assessment of equipment corrosion status and accurate prediction of remaining lifespan, ensuring the safe and efficient operation of coal chemical equipment.

Description

Technical Field

[0001] This invention relates to the field of equipment monitoring technology, and in particular to a method and application for equipment corrosion assessment and life prediction. Background Technology

[0002] In the coal chemical industry, equipment operates long-term under high temperatures (200–1500℃), high pressures (5–30 MPa), and environments rich in CO2, H2S, and Cl. - Extreme environments with corrosive media. Taking a coal-water slurry gasifier as an example, its inner wall is subjected to high-temperature erosion exceeding 1400℃ for extended periods. Simultaneously, acidic gases such as H2S and HCN, along with the electrolyte solution formed by condensate, work in conjunction with the mechanical wear of coal ash and slag, resulting in a complex damage process of "high-temperature corrosion + electrochemical corrosion + erosion." In ammonia synthesis plants, the risk of hydrogen corrosion on the synthesis tower walls is particularly prominent. Under high temperature and pressure, H2 molecules penetrate the steel lattice, reacting with carbon to generate methane, leading to decarburization. Simultaneously, pitting corrosion caused by Cl- rust accelerates localized failure.

[0003] Corrosion problems in pipeline systems are more complex. Taking a low-temperature methanol washing unit as an example, in a -40°C environment, acidic gases such as H2S and COS dissolve in the methanol solution to form corrosive media. At pipe bends and diameter changes, the scouring effect intensifies, leading to turbulent corrosion. During shutdowns and maintenance, residual media react with air to form sulfuric acid, causing under-deposit corrosion. According to industry statistics, unplanned shutdowns due to corrosion account for an average of 28% of all coal chemical plant shutdowns annually, with direct economic losses from a single shutdown reaching tens of millions of yuan.

[0004] Current traditional assessment methods have significant limitations: single-point corrosion rate monitoring based on linear polarization resistance can only capture uniform corrosion conditions and cannot reflect the coupling effect of localized corrosion and multiple corrosion morphologies; empirical life prediction models that rely on fitting historical data are ill-suited to the fluctuating operating conditions caused by frequent load changes in coal chemical plants. For example, a coal-to-olefins project predicted a pipeline life of 8 years using traditional methods, but a perforation leak occurred after only 5 years of actual operation, exposing the failure risk of the traditional assessment system under complex operating conditions. Therefore, there is an urgent need to construct an assessment system that can dynamically quantify the interaction of multiple corrosion mechanisms to achieve accurate protection of coal chemical equipment throughout its entire life cycle. Summary of the Invention

[0005] The purpose of this invention is to provide a method and application for equipment corrosion assessment and life prediction. By integrating multi-source data and intelligent algorithms, it can achieve accurate assessment of equipment corrosion status and accurate prediction of remaining life, thereby ensuring the safe and efficient operation of coal chemical equipment.

[0006] To achieve the above objectives, the present invention provides a method for equipment corrosion assessment and life prediction, comprising the following steps:

[0007] S1, Multi-source data acquisition

[0008] Deploy a sensor network in the easily corroded areas of coal chemical equipment to collect multi-dimensional data such as corrosion current density, environmental parameters, wall thickness, stress, and surface temperature images of the equipment surface;

[0009] S2, Data Preprocessing

[0010] The multi-source data collected by S1 is transmitted to the edge computing node for outlier removal, data compression, time synchronization and spatiotemporal alignment preprocessing.

[0011] S3. Corrosion Feature Extraction and Evaluation

[0012] Image features are extracted using a convolutional neural network, and the LSTM-attention model is used to analyze the S2-processed data to establish a fuzzy comprehensive evaluation matrix to assess the corrosion level.

[0013] S4. Remaining life prediction

[0014] A physical model based on Faraday's law of electrolysis and a data-driven model based on Transformer networks were constructed, and the output was fused through a Bayesian network and combined with Monte Carlo simulation to predict the remaining lifetime.

[0015] S5, Dynamic Calibration and Method Support

[0016] Based on offline detection data, changes in corrosion level, and remaining life threshold triggering model calibration, maintenance methods are formulated according to the warning level.

[0017] Preferably, in S1, the sensor network includes a linear polarization resistance sensor, a sampling analysis sensor, a pH sensor, a temperature and humidity sensor, an ultrasonic thickness gauge, a strain gauge, and an infrared thermal imager.

[0018] Preferably, in S2, outliers are removed based on the 3σ principle, high-frequency corrosion current density data is compressed using sliding window mean filtering, time synchronization is achieved through the NTP protocol, and spatiotemporal alignment and integration are performed using a timestamp-based Kalman filter algorithm.

[0019] Preferably, in S3, a convolutional neural network is used to process the temperature images acquired by the infrared thermal imager to extract the image features of rust layer thickness and crack distribution; the corrosion rate, environmental parameters, equipment wall thickness data, strain data, and equipment surface temperature image data are analyzed by an LSTM-attention model to identify the coupling relationship between key influencing factors and corrosion rate.

[0020] Preferably, in S3, the evaluation indicators of the fuzzy comprehensive evaluation matrix include corrosion rate, corrosion depth, environmental erosion index, and stress corrosion risk, with weights of 0.4, 0.3, 0.2, and 0.1 respectively. The corrosion level is determined by Z-score normalization, construction of a membership matrix, and calculation of a comprehensive evaluation vector. The corrosion level includes five levels: intact, light corrosion, moderate corrosion, heavy corrosion, and failure risk.

[0021] In an even more preferred embodiment, in S3, the corrosion rate is the synergistic corrosion rate that establishes the synergistic effect of CO2 corrosion, H2S corrosion, and erosion corrosion.

[0022] CO2 corrosion is as follows:

[0023]

[0024] In the formula, is the CO2 corrosion rate, mm / a; k1 is the CO2 corrosion rate constant, mm / (a·MPa·K); CO2 partial pressure, MPa; E a The activation energy is kJ / mol, R is the ideal gas constant, and T is the temperature in °C; ηpH = 10 0.1(4-pH) , which is the pH value influence coefficient, and corrosion is significantly accelerated when pH ≤ 4.

[0025] H2S corrosion is as follows:

[0026]

[0027] In the formula, k2 is the H2 corrosion rate, mm / a; k2 is the H2 corrosion rate constant, mm / (a·√ppm); cH2S is the H2S concentration, ppm; c Cl - represents the chloride ion concentration, in ppm.

[0028] Erosion corrosion is:

[0029] υ crosion =k3·u 2 ·α·ρ s

[0030] In the formula, υ crosion kt is the erosion corrosion rate, in mm / a; k3 is the erosion corrosion coefficient, in mm / (a·(m / s)). 2 kg / m 3 ); u is the fluid velocity, m / s; α is the volume fraction of solid particles, %; ρ s Particle density, kg / m³ 3 .

[0031] The synergistic corrosion rate is:

[0032]

[0033] In the formula, υ 协同 The synergistic corrosion rate is expressed in mm / a; k4 is the synergistic effect coefficient m. 3 / (a 2 (·MPa·ppm), reflecting the interaction acceleration effect of CO2 and H2S, with a value of 0.01 to 0.05.

[0034] In an even more preferred embodiment, in S3, the image features are quantized using a convolutional neural network to output the fractal dimension D of the rust layer and the crack density δ of the corrosion region;

[0035] The fractal dimension D of the rust layer is calculated using the box-dimensional algorithm to determine the surface roughness of the rust layer.

[0036] The crack density δ is:

[0037]

[0038] In the formula, L crack A represents the total crack length in mm; ROI The area of ​​the region of interest is in mm. 2 .

[0039] In a further preferred embodiment, in S3, a model is constructed that includes image features and erosion rate υ. 协同 Environmental parameters The 10-dimensional eigenvectors of stress and strain (σ, ε):

[0040]

[0041] In the preferred embodiment, S3, the model construction of the comprehensive evaluation vector includes:

[0042] (1) Indicator standardization

[0043]

[0044] When i=1, x1 is the corrosion rate; when i=2, x2 is the corrosion depth; when i=3, x3 is the environmental corrosion index; when i=4, x4 is the stress corrosion risk.

[0045] (2) Membership function

[0046] The membership degree μ at each level is calculated using the trapezoidal distribution function. j (x′ i (j = 1~5, corresponding to 5 corrosion levels):

[0047]

[0048] In the formula, a j b j cj Threshold parameters for each level;

[0049] (3) Comprehensive evaluation vector

[0050]

[0051] In the formula, ω1, ω2, ω3, and ω4 are the weight coefficients of each evaluation index.

[0052] Preferably, in S4, the theoretical corrosion depth is calculated by combining the physical model with the material properties of the equipment. A Transformer network is used, and historical corrosion depth, environmental parameters, operating conditions and maintenance records are input to establish a data-driven model to predict the corrosion depth in the next 1-10 years.

[0053] In S4, the physical model is based on Faraday's law of electrolysis, which forms the fundamental equation for the theoretical corrosion rate.

[0054]

[0055] In the formula, υ 理论 The theoretical corrosion rate is given in mm / a; K is a constant, K = 3.27 × 10⁻⁶. -3 i corr The corrosion current density is expressed in μA / cm. 2 M is the molar mass of the metal, g / mol; n is the oxidation state number; ρ is the density of the metal, g / cm³. 3 ;

[0056] In the preferred embodiment, S4, the formula for calculating the theoretical corrosion depth is:

[0057] d 理论 =υ 理论 ·t+d0

[0058] In the formula, t represents time, in years; and d0 represents the initial corrosion depth.

[0059] In the preferred embodiment, S4, the formula for predicting the corrosion depth is:

[0060]

[0061] Preferably, in S4, the outputs of the physical model and the data-driven model are dynamically fused through a Bayesian network, and then Monte Carlo simulation is used to sample uncertain parameters and run it 100,000 times to generate a corrosion depth probability distribution. The remaining lifetime is defined as the time quantile when the corrosion depth reaches the critical value, and the 5th quantile, 50th quantile, and 95th quantile are output.

[0062] A further optimized approach involves dynamically fusing the physical model and the data-driven model using a Bayesian network, resulting in the following output:

[0063] In the formula, This is the final corrosion assessment result obtained after fusion;

[0064] d 理论 This represents the theoretical corrosion depth calculated based on a physical model.

[0065] d 预测 The corrosion depth is predicted by the data-driven model;

[0066] The dynamic weight function is determined by Bayesian optimization, with parameters k and t0, and an initial value ω = 0.5.

[0067] Preferably, in S5, when offline detection data is uploaded, or the corrosion level increases for three consecutive days and the predicted remaining lifespan is lower than the warning threshold, model calibration is triggered. The model calibration is performed using a model parameter fine-tuning algorithm based on stochastic gradient descent.

[0068] Preferably, in S5, the warning levels include yellow warning, orange warning and red warning. The maintenance method is as follows: when a yellow warning occurs, increase the sensor sampling frequency and start weekly infrared inspection; when an orange warning occurs, expand the ultrasonic thickness measurement detection range and conduct a comprehensive analysis of the corrosive components of the medium; when a red warning occurs, formulate a replacement plan and arrange shutdown for maintenance.

[0069] This invention also provides an application of the above-mentioned equipment corrosion assessment and life prediction method, which is applied to the corrosion assessment and life prediction of gasifiers, synthesis towers, separation towers, and process pipelines in coal chemical equipment, where the coal chemical equipment is in a corrosive medium environment containing high temperature, high pressure, CO2, and H2S.

[0070] The beneficial effects of this invention are:

[0071] (1) The present invention adopts the above-mentioned equipment corrosion assessment and life prediction method and application. Through multi-source data fusion acquisition, it comprehensively obtains equipment corrosion-related information, breaks through the limitations of traditional single parameter monitoring, and can accurately reflect the equipment corrosion status under the synergistic effect of multiple corrosion mechanisms.

[0072] (2) The present invention adopts the above-mentioned equipment corrosion assessment and life prediction method and application, adopts a hybrid prediction framework that combines physical model and data-driven model, and combines Bayesian network and Monte Carlo simulation. It considers both the physical and chemical nature of corrosion and utilizes the regularity of historical data, which significantly improves the accuracy and reliability of remaining life prediction.

[0073] (3) The present invention adopts the above-mentioned equipment corrosion assessment and life prediction method and application, dynamic calibration mechanism and early warning maintenance strategy, which can adjust the prediction model and maintenance plan in a timely manner according to the actual operation of the equipment, effectively reduce the risk of equipment failure caused by corrosion, and improve the safety and economy of coal chemical equipment.

[0074] The technical solution of the present invention will be further described in detail below through embodiments. Detailed Implementation

[0075] The present invention will be further described below with reference to embodiments. Unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art. The features mentioned above or in the specific examples mentioned in this invention can be combined arbitrarily, and these specific embodiments are only used to illustrate the invention and are not intended to limit the scope of the invention.

[0076] Example

[0077] This invention provides a method for equipment corrosion assessment and life prediction, comprising the following steps:

[0078] S1, Multi-source data acquisition

[0079] Deploy a sensor network in the easily corroded areas of coal chemical equipment to collect multi-dimensional data such as corrosion current density, environmental parameters, wall thickness, stress, and surface temperature images of the equipment surface.

[0080] The sensor network includes linearly polarized resistance sensors, chloride ion concentration sensors, pH sensors, temperature and humidity sensors, ultrasonic thickness gauges, strain gauges, and infrared thermal imagers. The linearly polarized resistance sensors acquire real-time corrosion current density data on the equipment surface to calculate the corrosion rate. Environmental sensors such as chloride ion concentration sensors, pH sensors, and temperature and humidity sensors obtain environmental parameters such as Cl- concentration, pH value, temperature, and humidity in the environment in which the equipment is located. The ultrasonic thickness gauge periodically measures the equipment wall thickness, and strain gauges monitor strain data in stress concentration areas. The infrared thermal imager periodically acquires surface temperature images of the equipment to identify areas of temperature anomalies caused by corrosion.

[0081] S2, Data Preprocessing

[0082] The multi-source data collected by S1 is transmitted to the edge computing node for outlier removal, data compression, time synchronization, and spatiotemporal alignment preprocessing.

[0083] Outliers in corrosion current density and wall thickness data were removed based on the 3σ principle. High-frequency corrosion current density data were compressed using a sliding window mean filter (window size 5 min). Time synchronization of data from various sensors was performed via the NTP protocol to ensure timestamp errors were <10 ms. A spatiotemporal alignment algorithm (such as timestamp-based Kalman filtering) was used to integrate heterogeneous data and construct a device corrosion state vector.

[0084] S3. Corrosion Feature Extraction and Evaluation

[0085] A convolutional neural network was used to process temperature images acquired by an infrared thermal imager to extract image features such as rust layer thickness and crack distribution.

[0086] Image features are quantized using a convolutional neural network to output the fractal dimension D of the rust layer and the crack density δ of the corroded area.

[0087] The fractal dimension D of the rust layer is calculated using the box-dimensional algorithm to determine the surface roughness of the rust layer.

[0088] The crack density δ is:

[0089]

[0090] In the formula, L crack A represents the total crack length in mm; ROI The area of ​​the region of interest is in mm. 2 .

[0091] In a further preferred embodiment, in S3, a model is constructed that includes image features and erosion rate υ. 协同 Environmental parameters The 10-dimensional eigenvectors of stress and strain (σ, ε):

[0092] By analyzing the corrosion rate, environmental parameters, equipment wall thickness data, strain data, and equipment surface temperature image data after S2 processing using the LSTM-attention model, the coupling relationship between key influencing factors and corrosion rate was identified.

[0093] The corrosion rate is the synergistic corrosion rate that establishes the combined effects of CO2 corrosion, H2S corrosion, and erosion corrosion.

[0094] CO2 corrosion is as follows:

[0095]

[0096] In the formula, is the CO2 corrosion rate, mm / a; k1 is the CO2 corrosion rate constant, mm / (a·MPa·K); CO2 partial pressure, MPa; E aThe activation energy is kJ / mol, R is the ideal gas constant, and T is the temperature in °C; ηpH = 10 0.1(4-pH) , which is the pH value influence coefficient, and corrosion is significantly accelerated when pH ≤ 4.

[0097] H2S corrosion is as follows:

[0098]

[0099] In the formula, k2 is the H2 corrosion rate, mm / a; k2 is the H2 corrosion rate constant, mm / (a·√ppm); cH2S is the H2S concentration, ppm; c Cl - represents the chloride ion concentration, in ppm.

[0100] Erosion corrosion is:

[0101] υ crosion =k3·u 2 ·α·ρ s

[0102] In the formula, υ crosion kt is the erosion corrosion rate, in mm / a; k3 is the erosion corrosion coefficient, in mm / (a·(m / s)). 2 kg / m 3 ); u is the fluid velocity, m / s; α is the volume fraction of solid particles, %; ρ s Particle density, kg / m³ 3 .

[0103] The synergistic corrosion rate is:

[0104]

[0105] In the formula, υ 协同 The synergistic corrosion rate is the corrosion rate, expressed in mm / a; k4 is the synergistic effect coefficient m. 3 / (a 2 (·MPa·ppm), reflecting the interaction acceleration effect of CO2 and H2S, with a value of 0.01 to 0.05.

[0106] A fuzzy comprehensive evaluation matrix was established to assess the corrosion level. Evaluation indicators included corrosion rate, corrosion depth, environmental erosion index, and stress corrosion risk, with weights of 0.4, 0.3, 0.2, and 0.1 respectively. The corrosion level was determined through Z-score normalization, construction of a membership matrix, and calculation of a comprehensive evaluation vector. The corrosion levels were categorized into five levels: intact, lightly corroded, moderately corroded, heavily corroded, and failure risk. The specific steps are as follows:

[0107] (1) Standardization of indicators: Corrosion rate, corrosion depth, environmental corrosion index, and stress corrosion risk are normalized using Z-score, with the following formula:

[0108]

[0109] When i=1, x1 is the corrosion rate; when i=2, x2 is the corrosion depth; when i=3, x3 is the environmental corrosion index; when i=4, x4 is the stress corrosion risk.

[0110] (2) Membership function

[0111] The membership degree μ of each index to the five corrosion levels was calculated using the trapezoidal distribution function. j (x′ i (j = 1~5, corresponding to 5 corrosion levels):

[0112]

[0113] In the formula, a j b j c j Threshold parameters for each level;

[0114] For example, the membership function for corrosion rate is:

[0115]

[0116] (3) Comprehensive evaluation vector

[0117]

[0118] In the formula, ω1, ω2, ω3, and ω4 are the weight coefficients of each evaluation index, and W = 0.4, 0.3, 0.2, and 0.1, respectively.

[0119] S4. Remaining life prediction

[0120] A physical model based on Faraday's law of electrolysis and a data-driven model based on Transformer networks are constructed. The output is fused through a Bayesian network and combined with Monte Carlo simulation to predict the remaining lifetime.

[0121] The theoretical corrosion depth is calculated by combining a physical model with the material properties of the equipment. The physical model is based on the fundamental equations for the theoretical corrosion rate constructed according to Faraday's law of electrolysis.

[0122]

[0123] In the formula, υ 理论 The theoretical corrosion rate is given in mm / a; K is a constant, K = 3.27 × 10⁻⁶. -3 i corr The corrosion current density is expressed in μA / cm. 2M is the molar mass of the metal, g / mol; n is the oxidation state number; ρ is the density of the metal, g / cm³. 3 .

[0124] Using carbon steel (M = 55.85 g / mol, n = 2, ρ = 7.87 g / cm³) 3 For example, if the measured corrosion current density i corr =15μA / cm 2 Then the theoretical corrosion rate is:

[0125]

[0126] The formula for calculating the theoretical corrosion depth is:

[0127] d 理论 =0.018·t+d0

[0128] In the formula, t represents time, in years; and d0 represents the initial corrosion depth.

[0129] Using a Transformer network, and inputting historical corrosion depth, environmental parameters, operating conditions, and maintenance records, a data-driven model is established to predict the corrosion depth for the next 1-10 years. The formula for calculating the predicted corrosion depth is:

[0130]

[0131] For example, if the corrosion depth sequence of a pipeline over the past 5 years is [0.1, 0.3, 0.6, 1.0, 1.5] mm, the model predicts that the corrosion depth in the 6th year will be 1.9 mm.

[0132] The outputs of the physical model and the data-driven model are dynamically fused using a Bayesian network:

[0133]

[0134] In the formula, This is the final corrosion assessment result obtained after fusion;

[0135] d 理论 This represents the theoretical corrosion depth calculated based on a physical model.

[0136] d 预测 The corrosion depth is predicted by the data-driven model;

[0137] The dynamic weight function is determined by Bayesian optimization, with parameters k and t0, and an initial value ω = 0.5.

[0138] When the current weight is 0.6,

[0139] Then, Monte Carlo simulation was used. By sampling uncertain parameters such as corrosion rate fluctuation range (±20%) and Cl- concentration change (±15%), the simulation was run 100,000 times to generate a corrosion depth probability distribution. The remaining lifetime was defined as the time quantile when the corrosion depth reaches the critical value, and the 5th percentile (P5), 50th percentile (P50), and 95th percentile (P95) were output.

[0140] S5, Dynamic Calibration and Method Support

[0141] Model calibration is triggered based on offline detection data, changes in corrosion level, and remaining lifetime threshold: θ represents the model parameters, α = 0.01 represents the learning rate, and L represents the mean squared error loss function.

[0142] The warning levels and maintenance strategies are as follows:

[0143] When offline detection data is uploaded, or the corrosion level rises for three consecutive days and the predicted remaining lifespan falls below the warning threshold, model calibration is triggered. Model calibration uses a stochastic gradient descent-based algorithm for fine-tuning model parameters, automatically running a global calibration monthly and performing local fine-tuning weekly. Warning levels include: Yellow warning (remaining lifespan ≤ 3 years), Orange warning (remaining lifespan ≤ 2 years), and Red warning (remaining lifespan ≤ 1 year). Maintenance methods include: for a yellow warning, increasing the sensor sampling frequency to once per minute and initiating weekly infrared inspections; for an orange warning, expanding the ultrasonic thickness measurement range by 50% to cover all suspected corrosion areas and performing a comprehensive analysis of the corrosive components of the medium (e.g., fluctuations in CO2, H2S, and Cl- concentrations), and adjusting process parameters to suppress corrosion; and for a red warning, developing a replacement plan and arranging a shutdown for maintenance.

[0144] Application examples

[0145] This invention also provides an application of the equipment corrosion assessment and life prediction method described in the above embodiments. In 2016, it was applied to the corrosion assessment of the top pipeline of the carbon dioxide separation tower in a coal chemical gasification unit. The pipeline transports syngas containing CO2 / H2S, and the operating parameters are as follows:

[0146] cH2S = 600ppm, c Cl - = 180ppm, T = 200℃, conveying rate u = 6m / s;

[0147] The material is 304L, the initial wall thickness is 8mm, and the critical corrosion depth is 2.4mm (30% of the wall thickness).

[0148] Specifically, the following steps are included:

[0149] S1. Deploy a sensor network to collect multi-source data.

[0150] Three linear polarized resistance sensors, two sampling and analysis sensors, and two temperature and humidity sensors were deployed in easily corroded areas such as welds and corners of the pipeline; two strain gauges were installed at welds where stress was concentrated in the pipeline; the wall thickness of the pipeline was measured at six points monthly using an ultrasonic thickness gauge; and the pipeline was comprehensively imaged weekly using an infrared thermal imager.

[0151] S2, Data Processing

[0152] The multi-source data collected by S1 is transmitted to the edge computing node via the LoRaWAN wireless protocol. The collected corrosion current density data is processed by outlier removal and sliding window mean filtering. Environmental parameters and strain data are synchronized in time and integrated in spatiotemporal alignment to form a pipeline corrosion state vector and stored in the InfluxDB time series database.

[0153] S3. Corrosion Feature Extraction and Evaluation

[0154] Images acquired by an infrared thermal imager were input into a convolutional neural network to identify a corrosion area on the inner wall of the pipe with a rust layer thickness of approximately 0.3 mm. Time-series data were analyzed using an LSTM-attention model to determine that CO2 corrosion and H2S corrosion are the key factors affecting the current corrosion rate.

[0155]

[0156] υ crosion =k3·6 2 ·0.05·2500=0.045mm / a

[0157] The calculated synergistic corrosion rate is υ 协同 =0.101+0.15+0.03×0.101×0..15=0.30mm / a.

[0158] Based on the fuzzy comprehensive evaluation matrix and the comprehensive evaluation vector calculation, the current corrosion level of the chemical pipeline is determined.

[0159] The fuzzy comprehensive evaluation yielded B = [0.1, 0.2, 0.3, 0.4, 0.0], which was determined to be level 4 (severe corrosion).

[0160] S4. Remaining life prediction

[0161] The theoretical corrosion depth was calculated using a physical model, and combined with parameters such as the current corrosion current density, the theoretical corrosion rate was found to be 0.075 mm / a.

[0162] The Transformer network of the data-driven model predicts an increase in corrosion depth of 0.12 mm over the next year based on historical data from the past 5 years. A Bayesian network is used to fuse the results of the physical and data models to obtain a fused corrosion depth prediction. After 100,000 Monte Carlo simulations, the fused model prediction is output. In 2024 (the 8th year of actual use), the corrosion depth was measured at 2.3 mm with an error of <5%, verifying the effectiveness of the method.

[0163] S5, Dynamic Calibration and Method Support

[0164] Six months after sensor deployment, data was collected in real time via a sensor network. After edge computing preprocessing, the data was input into an intelligent model to achieve dynamic assessment of corrosion levels and prediction of remaining life. After one year of system operation, the system successfully provided early warnings for three moderate corrosion risks, avoiding unplanned shutdowns and reducing equipment maintenance costs by 25%, thus verifying the engineering applicability of the method.

[0165] Therefore, the above-mentioned equipment corrosion assessment and life prediction method and application of the present invention can effectively realize the corrosion assessment and life prediction of coal chemical equipment under the synergistic effect of multiple corrosion mechanisms, and has good application effect and promotion value.

[0166] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for assessing equipment corrosion and predicting its lifespan, characterized in that, Includes the following steps: S1. Multi-source data acquisition: Deploy a sensor network in the easily corroded areas of coal chemical equipment to collect multi-dimensional data such as corrosion current density, environmental parameters, wall thickness, stress, and surface temperature images of the equipment surface; S2, Data Preprocessing: The multi-source data collected by S1 is transmitted to the edge computing node for outlier removal, data compression, time synchronization and spatiotemporal alignment preprocessing. S3. Corrosion Feature Extraction and Evaluation: Image features are extracted using a convolutional neural network, and the LSTM-attention model is used to analyze the S2-processed data to establish a fuzzy comprehensive evaluation matrix to assess the corrosion level. S4. Remaining life prediction: A physical model based on Faraday's law of electrolysis and a data-driven model based on Transformer networks were constructed, and the output was fused through a Bayesian network and combined with Monte Carlo simulation to predict the remaining lifetime. The theoretical corrosion depth is calculated by combining a physical model with the material properties of the equipment. Using a Transformer network, historical corrosion depth, environmental parameters, operating conditions and maintenance records are input to establish a data-driven model to predict the corrosion depth for the next 1-10 years. The physical model is based on Faraday's law of electrolysis, which forms the fundamental equation for the theoretical corrosion rate. In the formula, The theoretical corrosion rate is expressed in mm / a. K is a constant, K=3.27 10 -3 ; The corrosion current density is expressed in μA / cm. 2 ; The molar mass of the metal is expressed in g / mol. The oxidation state number; The density of the metal is expressed in g / cm³. 3 ; The formula for calculating the theoretical corrosion depth is: In the formula, For time, in years; This represents the initial corrosion depth. The formula for predicting corrosion depth is: ； By dynamically fusing the outputs of the physical model and the data-driven model through a Bayesian network, and then using Monte Carlo simulation to sample uncertain parameters and run it 100,000 times, a corrosion depth probability distribution is generated. The remaining lifetime is defined as the time quantile when the corrosion depth reaches the critical value, and the 5th quantile, 50th quantile, and 95th quantile are output. The output of dynamically fusing the physical model and the data-driven model through a Bayesian network is: In the formula, This is the final corrosion assessment result obtained after fusion; This represents the theoretical corrosion depth calculated based on a physical model. The corrosion depth is predicted by the data-driven model; , is a dynamic weight function, whose parameters are determined through Bayesian optimization. k,t 0 initial value ; S5. Dynamic calibration and method support: Based on offline detection data, changes in corrosion level, and remaining life threshold triggering model calibration, maintenance methods are formulated according to the warning level.

2. The method for equipment corrosion assessment and life prediction according to claim 1, characterized in that: In S1, the sensor network includes a linear polarization resistance sensor, a sampling analysis sensor, a pH sensor, a temperature and humidity sensor, an ultrasonic thickness gauge, a strain gauge, and an infrared thermal imager.

3. The method for equipment corrosion assessment and life prediction according to claim 1, characterized in that: In S2, outliers are removed based on the 3σ principle, high-frequency corrosion current density data is compressed using sliding window mean filtering, time synchronization is achieved through the NTP protocol, and spatiotemporal alignment and integration are performed using a timestamp-based Kalman filter algorithm.

4. The method for equipment corrosion assessment and life prediction according to claim 1, characterized in that: In S3, a convolutional neural network is used to process the temperature images acquired by the infrared thermal imager to extract the image features of rust layer thickness and crack distribution. The corrosion rate, environmental parameters, equipment wall thickness data, strain data, and equipment surface temperature image data are analyzed by an LSTM-attention model to identify the coupling relationship between key influencing factors and corrosion rate.

5. The method for equipment corrosion assessment and life prediction according to claim 1, characterized in that: In S3, the evaluation indicators of the fuzzy comprehensive evaluation matrix include corrosion rate, corrosion depth, environmental erosion index, and stress corrosion risk, with weights of 0.4, 0.3, 0.2, and 0.1, respectively. The corrosion level is determined by Z-score normalization, construction of membership matrix, and calculation of comprehensive evaluation vector. The corrosion level includes five levels: intact, light corrosion, moderate corrosion, heavy corrosion, and failure risk.

6. The method for equipment corrosion assessment and life prediction according to claim 1, characterized in that: In S5, when offline detection data is uploaded, or when the corrosion level rises for three consecutive days and the predicted remaining lifespan is lower than the warning threshold, model calibration is triggered. The model calibration is performed using a model parameter fine-tuning algorithm based on stochastic gradient descent.

7. The method for equipment corrosion assessment and life prediction according to claim 1, characterized in that: In S5, the warning levels include yellow, orange and red. The maintenance methods are as follows: when a yellow warning occurs, increase the sensor sampling frequency and start weekly infrared inspection; when an orange warning occurs, expand the ultrasonic thickness measurement range and conduct a comprehensive analysis of the corrosive components of the medium; when a red warning occurs, formulate a replacement plan and arrange for shutdown and maintenance.

8. An application of the equipment corrosion assessment and life prediction method as described in any one of claims 1-7, characterized in that: Corrosion assessment and life prediction are applied to gasifiers, synthesis towers, separation towers, and process pipelines in coal chemical equipment, which are exposed to high temperature, high pressure, and corrosive media containing CO2 and H2S.

Images