Industrial equipment corrosion state real-time monitoring and corrosion inhibitor intelligent adding system
By constructing a multi-dimensional factor corrosion risk assessment model and an adaptive corrosion inhibitor dosing strategy, the problems of inaccurate corrosion status monitoring and imprecise corrosion inhibitor dosing in existing technologies have been solved. This has enabled real-time monitoring of equipment corrosion status and efficient utilization of corrosion inhibitors, thereby improving the corrosion protection effect and safety of the equipment.
Patent Information
- Application Number
- CN202511100627.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-07
- Publication Date
- 2025-11-21
AI Technical Summary
Existing systems rely on traditional manual monitoring and experience-based methods, which cannot achieve real-time, accurate monitoring and dynamic adjustment of the corrosion status of industrial equipment. They lack adaptive capabilities, resulting in inaccurate corrosion inhibitor dosage, with the risk of over- or under-dosing. Furthermore, they cannot accurately predict corrosion trends in complex environments, leading to resource waste and safety hazards.
By acquiring historical corrosion status data of equipment, a multi-dimensional factor corrosion risk assessment model is constructed. Combining LSTM long short-term memory network and attention mechanism, the future corrosion status is predicted. Based on the adaptive corrosion inhibitor dosing strategy, the dosage and frequency of corrosion inhibitor are adjusted in real time, and the dosing strategy is optimized using a proportional-integral-derivative controller.
It enables precise quantification of the corrosion status of industrial equipment and prediction of future trends, ensuring the efficient use of corrosion inhibitors, avoiding over- or under-dosing, optimizing corrosion protection, extending equipment life, and reducing resource waste and maintenance costs.
Smart Images
Figure CN120993799A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of industrial equipment monitoring technology, specifically to a real-time monitoring system for corrosion status of industrial equipment and an intelligent system for adding corrosion inhibitors. Background Technology
[0002] Most existing systems rely on traditional manual monitoring and experience-based methods, which cannot achieve real-time, accurate corrosion status monitoring and dynamic adjustment. The control of corrosion rate and corrosion inhibitor dosage often lacks adaptive capability, resulting in inaccurate corrosion inhibitor dosage and the risk of over- or under-dosing. In addition, existing technologies cannot accurately predict the corrosion trend of equipment in complex environments, lack in-depth analysis of historical data and equipment status, and are difficult to achieve optimal management of corrosion protection, resulting in resource waste and potential safety hazards to equipment. Summary of the Invention
[0003] To address the aforementioned technical problems, this paper provides a real-time monitoring solution for corrosion status of industrial equipment and an intelligent corrosion inhibitor dosing scheme. This solution solves the problems of existing systems that mostly rely on traditional manual monitoring and experience-based methods, which cannot achieve real-time, accurate corrosion status monitoring and dynamic adjustment; the control of corrosion rate and corrosion inhibitor dosage often lacks adaptive capability, leading to inaccurate corrosion inhibitor dosage and the risk of over- or under-dosing; in addition, existing technologies cannot accurately predict the corrosion trend of equipment in complex environments, lack in-depth analysis of historical data and equipment status, making it difficult to achieve optimal management of corrosion protection, resulting in resource waste and potential safety hazards to equipment.
[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows:
[0005] A method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitors for industrial equipment includes:
[0006] Acquire historical corrosion status data of equipment, analyze the trend of corrosion rate change of equipment per unit time, and predict the future corrosion status of equipment;
[0007] Based on sensor data, the current corrosion status of the equipment is obtained. Combined with the historical corrosion status data of the equipment, a dynamic multi-dimensional factor corrosion risk assessment model is constructed to quantify the current corrosion level of the equipment in real time. The multi-dimensional factors include: equipment material, operating environment and historical data.
[0008] By combining the current corrosion state of the equipment with its future corrosion state, an adaptive corrosion inhibitor dosing strategy is generated, dynamically linking the equipment corrosion rate with the corrosion inhibitor requirement.
[0009] Based on the equipment's adaptive corrosion inhibitor dosing strategy, the dosage and frequency of corrosion inhibitor addition are adjusted in real time to achieve corrosion protection of the equipment and efficient utilization of the corrosion inhibitor.
[0010] Preferably, based on the historical corrosion status data of the equipment, the data is normalized and linear interpolation is used to correct outliers;
[0011] As a further detail, the historical corrosion status data of the equipment includes: corrosion rate, temperature, humidity, and chemical composition;
[0012] Using the sliding window algorithm, a fixed window length is set, and the time series data within the current window is extracted as input features, while the data of the next adjacent time step outside the window is used as labels to construct a time series sample set of equipment corrosion status.
[0013] Using a time series sample set of equipment corrosion state as input, an LSTM long short-term memory network is trained to capture long-term dependencies in the equipment corrosion state data and output the hidden state at each time step.
[0014] Preferably, by combining an attention mechanism, the attention weight of the historical corrosion state data time step is dynamically calculated based on the hidden state of each time step, and the historical corrosion state data of the equipment is weighted and summed to obtain the weighted memory unit of the equipment at the current moment, as shown in the following formula:
[0015]
[0016] Where, α t Here, t represents the attention weights for each time step of the historical corrosion status data of the equipment, Softmax is the normalization function, V is the attention score vector, tanh is the hyperbolic tangent activation function, and W is the attention weight. h With W x h is a learnable weight matrix. t Let x be the hidden state at each time step. t Let b be the input feature vector at time step t, and m be the bias term. t α is the weighted memory unit of the device at the current moment. k h represents the attention weights at time step k. k This represents the hidden state at time step k;
[0017] By combining weighted historical corrosion status data of the equipment with the hidden states of the LSTM long short-term memory network, the future corrosion status of the equipment is predicted through regression calculation using fully connected layers.
[0018] Preferably, based on sensor data, the current corrosion status of the equipment is obtained, including the equipment material, the operating environment and historical data, and the data is preprocessed.
[0019] Training and testing sets were divided based on equipment material, usage environment, and historical data. The Bootstrap sampling method was used to randomly select multiple subsets of the training set to train decision trees. Each subset was split into multiple independent decision trees by randomly selecting some features. A random forest model was trained, and the number of decision trees, maximum depth, and minimum number of leaf node samples were set and adjusted as hyperparameters. The model was optimized by combining cross-validation method to construct a dynamic multi-dimensional factor corrosion risk assessment model for equipment.
[0020] Preferably, based on the dynamic multidimensional factor corrosion risk assessment model for equipment, the current corrosion risk of the equipment is calculated in real time, the corrosion level of the equipment is obtained, a corrosion risk threshold range is set according to industry standards, the current corrosion state of the equipment is defined as different levels, and the current corrosion level of the equipment is quantified in real time, as shown in the following formula:
[0021]
[0022] Where R(q) p Let q be the probability value of corrosion risk for the equipment at time p. p Let p be the input feature vector at time p, σ be the Sigmoid activation function, and e be the input feature vector at time p. i Let f be the dynamic weight of the i-th tree. i (q p ) represents the original output of the i-th decision tree, ε p For time-varying correction terms;
[0023] As a further detail, the corrosion levels of the equipment include: mild corrosion, moderate corrosion, and severe corrosion.
[0024] Preferably, based on the current corrosion state and future corrosion state of the equipment, the corrosion rate, corrosion risk level and environmental parameters of the equipment are defined as the state space, and the current corrosion state of the equipment and the amount of corrosion inhibitor added are defined as the action space to construct a state and action model.
[0025] Based on the state and action model, a reward function is designed with the goal of minimizing the corrosion rate, ensuring corrosion control within a safe range, and saving on the use of corrosion inhibitors.
[0026] Preferably, by analyzing the changing trend of equipment corrosion rate, a threshold condition for corrosion inhibitor addition is set. When the equipment corrosion rate exceeds the preset threshold, an intelligent corrosion inhibitor addition mechanism is triggered to dynamically adjust the amount of corrosion inhibitor added. By iteratively updating the Q-value table, the agent learns the rewards obtained by taking different actions in different states, thus obtaining the optimal adaptive corrosion inhibitor addition strategy for the equipment, as shown in the following formula:
[0027]
[0028] in, Given the current corrosion state of the equipment The optimal dosage of corrosion inhibitor. Let Q1(S) represent the corrosion state of the device at time t, and argmax be the action that maximizes the objective function. t A t ) is the Q-value function, representing the state S. t Take action A t The expected long-term returns.
[0029] Preferably, based on the equipment adaptive corrosion inhibitor dosing strategy, the equipment corrosion rate, corrosion inhibitor demand and actual corrosion rate are used as inputs, and the required amount and frequency of corrosion inhibitor dosing are calculated in real time as outputs.
[0030] The dosage and frequency of corrosion inhibitor are adjusted in real time by a proportional-integral-derivative controller. The control signal is calculated based on the deviation between the equipment corrosion rate and the corrosion inhibitor requirement, and the dosage of corrosion inhibitor is dynamically adjusted. According to historical dosage data and the changing trend of equipment corrosion status, the frequency of corrosion inhibitor addition is optimized to avoid over-addition or under-addition, thereby achieving corrosion protection of equipment and efficient utilization of corrosion inhibitor.
[0031] Furthermore, a real-time monitoring system for corrosion status of industrial equipment and an intelligent dosing system for corrosion inhibitors are provided to implement the aforementioned method for real-time monitoring of corrosion status of industrial equipment and intelligent dosing of corrosion inhibitors, comprising:
[0032] The system includes modules for future corrosion status of equipment, dynamic multi-dimensional corrosion risk assessment model, adaptive corrosion inhibitor dosing strategy, corrosion protection, and efficient utilization of corrosion inhibitors.
[0033] The device future corrosion status module is used to acquire historical corrosion status data of the device, analyze the trend of corrosion rate change of the device per unit time, and predict the future corrosion status of the device.
[0034] The dynamic multidimensional factor corrosion risk assessment model module is used to obtain the current corrosion status of the equipment based on sensor data, and combine it with the historical corrosion status data of the equipment to construct a dynamic multidimensional factor corrosion risk assessment model for the equipment, and quantify the current corrosion level of the equipment in real time. The multidimensional factors include: equipment material, operating environment and historical data.
[0035] The adaptive corrosion inhibitor dosing strategy module is electrically connected to the dynamic multidimensional factor corrosion risk assessment model module and the equipment future corrosion state module. The adaptive corrosion inhibitor dosing strategy module is used to combine the current corrosion state of the equipment with the future corrosion state of the equipment to generate an adaptive corrosion inhibitor dosing strategy for the equipment, and dynamically associate the equipment corrosion rate with the corrosion inhibitor requirement.
[0036] The corrosion protection and high-efficiency utilization module of the corrosion inhibitor is electrically connected to the adaptive corrosion inhibitor dosing strategy module. The corrosion protection and high-efficiency utilization module of the corrosion inhibitor is used to adjust the dosage and frequency of corrosion inhibitor dosing in real time according to the equipment's adaptive corrosion inhibitor dosing strategy, so as to realize the corrosion protection and high-efficiency utilization of the corrosion inhibitor of the equipment.
[0037] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0038] This invention proposes a real-time corrosion status monitoring and intelligent corrosion inhibitor dosing scheme for industrial equipment. By acquiring historical corrosion data and monitoring the equipment's status in real time, combined with a multi-dimensional factor evaluation model, this scheme can accurately quantify the corrosion level and predict future corrosion trends. The system can automatically generate an adaptive corrosion inhibitor dosing strategy, adjusting the dosage and frequency of the inhibitor in real time. This dynamic adjustment ensures efficient use of the corrosion inhibitor, avoiding over- or under-dosing, thereby optimizing the corrosion protection effect, extending the equipment's service life, reducing resource waste, and lowering maintenance costs. Attached Figure Description
[0039] Figure 1 A flowchart of a method for real-time monitoring of corrosion status and intelligent addition of corrosion inhibitors for industrial equipment;
[0040] Figure 2 This is a framework diagram of a real-time monitoring system for corrosion status of industrial equipment and an intelligent system for adding corrosion inhibitors. Detailed Implementation
[0041] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.
[0042] Reference Figure 1 As shown, a method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitors for industrial equipment includes:
[0043] S1. Obtain historical corrosion status data of the equipment, analyze the trend of corrosion rate change of the equipment per unit time, and predict the future corrosion status of the equipment.
[0044] Step S1 includes the following:
[0045] Based on the historical corrosion status data of the equipment, the data is normalized and outliers are corrected by linear interpolation.
[0046] As a further detail, the historical corrosion status data of the equipment includes: corrosion rate, temperature, humidity, and chemical composition;
[0047] Using the sliding window algorithm, a fixed window length is set, and the time series data within the current window is extracted as input features, while the data of the next adjacent time step outside the window is used as labels to construct a time series sample set of equipment corrosion status.
[0048] Using a time series sample set of equipment corrosion state as input, an LSTM long short-term memory network is trained to capture long-term dependencies in the equipment corrosion state data and output the hidden state at each time step.
[0049] Combining the attention mechanism, the attention weight of the historical corrosion state data time step is dynamically calculated based on the hidden state of each time step. The historical corrosion state data is then weighted and summed to obtain the weighted memory unit of the device at the current moment, as shown in the following formula:
[0050]
[0051] Where, α t Here, t represents the attention weights for each time step of the historical corrosion status data of the equipment, Softmax is the normalization function, V is the attention score vector, tanh is the hyperbolic tangent activation function, and W is the attention weight. h With W x h is a learnable weight matrix. t Let x be the hidden state at each time step. t Let b be the input feature vector at time step t, and m be the bias term. t α is the weighted memory unit of the device at the current moment. k h represents the attention weights at time step k. k This represents the hidden state at time step k;
[0052] By combining weighted historical corrosion status data of the equipment with the hidden states of the LSTM long short-term memory network, the future corrosion status of the equipment is predicted through regression calculation using fully connected layers.
[0053] When using it, refer to the content of step S1 above:
[0054] In existing technologies, traditional statistical methods and simple machine learning models struggle to capture the complex nonlinear relationships and long-term dependencies in the equipment corrosion process, and their ability to process multidimensional data is limited. In contrast, models based on LSTM (Long Short-Term Memory) networks and attention mechanisms can better capture the dynamic characteristics of equipment corrosion and improve prediction accuracy. LSTM networks capture long-term dependencies, and the attention mechanism dynamically weights historical data, enhancing the focus on key features, thereby providing more accurate and reliable predictions and addressing the shortcomings of traditional methods.
[0055] S2. Based on sensor data, obtain the current corrosion status of the equipment, and combine it with the historical corrosion status data of the equipment to construct a dynamic multi-dimensional factor corrosion risk assessment model for the equipment, and quantify the current corrosion level of the equipment in real time. The multi-dimensional factors include: equipment material, operating environment and historical data.
[0056] Step S2 includes the following:
[0057] Based on sensor data, the current corrosion status of the equipment is obtained, including the equipment material, operating environment and historical data, and the data is preprocessed.
[0058] Training and testing sets were divided based on equipment material, usage environment, and historical data. The Bootstrap sampling method was used to randomly select multiple subsets of the training set to train decision trees. Each subset was split into multiple independent decision trees by randomly selecting some features. A random forest model was trained, and the number of decision trees, maximum depth, and minimum number of leaf node samples were set and adjusted as hyperparameters. The model was optimized by combining cross-validation method to construct a dynamic multi-dimensional factor corrosion risk assessment model for equipment.
[0059] Based on the dynamic multidimensional factor corrosion risk assessment model for equipment, the current corrosion risk of the equipment is calculated in real time, the corrosion level of the equipment is obtained, a corrosion risk threshold range is set according to industry standards, different levels are defined for the current corrosion state of the equipment, and the current corrosion level of the equipment is quantified in real time. The formula is as follows:
[0060]
[0061] Where R(q) p Let q be the probability value of corrosion risk for the equipment at time p. p Let p be the input feature vector at time p, σ be the Sigmoid activation function, and e be the input feature vector at time p. i Let f be the dynamic weight of the i-th tree. i (q p ) represents the original output of the i-th decision tree, ε p For time-varying correction terms;
[0062] As a further detail, the corrosion levels of the equipment include: mild corrosion, moderate corrosion, and severe corrosion.
[0063] When using it, refer to the content of step S2 above:
[0064] Existing technologies for equipment corrosion assessment often rely on single factors, lacking comprehensive consideration of multidimensional factors and dynamic changes, resulting in insufficient assessment accuracy. Traditional models fail to effectively handle time-varying data and most rely on fixed thresholds, exhibiting poor adaptability. This step, by incorporating equipment material, operating environment, and historical data, constructs a dynamic multidimensional factor corrosion risk assessment model. This model overcomes the limitations of traditional models, enabling real-time and accurate quantification of equipment corrosion levels and flexibly adapting to changes under different operating conditions, thus improving the accuracy and practicality of the assessment.
[0065] S3. Combine the current corrosion state of the equipment with the future corrosion state of the equipment to generate an adaptive corrosion inhibitor dosing strategy for the equipment, and dynamically link the equipment corrosion rate with the corrosion inhibitor requirement.
[0066] Step S3 includes the following:
[0067] Based on the current corrosion state and future corrosion state of the equipment, the corrosion rate, corrosion risk level and environmental parameters of the equipment are defined as the state space, and the current corrosion state of the equipment and the amount of corrosion inhibitor added are defined as the action space to construct a state and action model.
[0068] Based on the state and action model, a reward function is designed with the goal of minimizing the corrosion rate and ensuring that corrosion control is within a safe range while saving the use of corrosion inhibitors.
[0069] By analyzing the changing trend of equipment corrosion rate, a threshold condition for corrosion inhibitor addition is set. When the equipment corrosion rate exceeds the preset threshold, the intelligent corrosion inhibitor addition mechanism is triggered, dynamically adjusting the amount of corrosion inhibitor added. Through iterative updates to the Q-value table, the agent learns the rewards obtained by taking different actions in different states, thus obtaining the optimal adaptive corrosion inhibitor addition strategy for the equipment. The formula is as follows:
[0070]
[0071] in, Given the current corrosion state of the equipment The optimal dosage of corrosion inhibitor. Let Q1(S) represent the corrosion state of the device at time t, and argmax be the action that maximizes the objective function. t A t ) is the Q-value function, representing the state S. t Take action A t The expected long-term returns.
[0072] When using it, refer to the content of step S3 above:
[0073] Current corrosion control methods largely rely on empirical approaches, failing to dynamically adjust the amount of corrosion inhibitor added in real time based on the equipment's corrosion status and environmental changes. Existing technologies cannot accurately predict future corrosion conditions in complex environments, leading to poor corrosion control, resource waste, and safety hazards. This step generates an adaptive corrosion inhibitor addition strategy by combining the equipment's current and future corrosion status, enabling real-time adjustments. The intelligent Q-learning mechanism can minimize corrosion rates, save corrosion inhibitors, optimize equipment operational safety, extend service life, and reduce costs.
[0074] S4. Based on the equipment's adaptive corrosion inhibitor dosing strategy, adjust the dosage and frequency of corrosion inhibitor dosing in real time to achieve corrosion protection of the equipment and efficient utilization of the corrosion inhibitor.
[0075] Step S4 includes the following:
[0076] Based on the equipment adaptive corrosion inhibitor dosing strategy, the equipment corrosion rate, corrosion inhibitor demand and actual corrosion rate are used as inputs, and the required amount and frequency of corrosion inhibitor dosing are calculated in real time as outputs.
[0077] The dosage and frequency of corrosion inhibitor are adjusted in real time by a proportional-integral-derivative controller. The control signal is calculated based on the deviation between the equipment corrosion rate and the corrosion inhibitor requirement, and the dosage of corrosion inhibitor is dynamically adjusted. According to historical dosage data and the changing trend of equipment corrosion status, the frequency of corrosion inhibitor addition is optimized to avoid over-addition or under-addition, thereby achieving corrosion protection of equipment and efficient utilization of corrosion inhibitor.
[0078] When using it, refer to the content of step S4 above:
[0079] Current corrosion control technologies mainly rely on empirical methods, which cannot dynamically adjust the dosage of corrosion inhibitors based on the real-time corrosion status of the equipment and environmental changes. Traditional systems cannot accurately predict the future corrosion rate of equipment in complex environments, resulting in poor corrosion control, resource waste, and potential safety hazards. This step, by monitoring the equipment's corrosion status and corrosion inhibitor requirements in real time, combines a controller to intelligently adjust the dosage and frequency of corrosion inhibitors. Based on the equipment's historical data and corrosion trends, the dosage frequency can be dynamically optimized, effectively avoiding over- or under-dosing, ensuring the equipment receives optimal corrosion protection and achieving efficient utilization of corrosion inhibitors, thereby improving equipment safety, extending service life, and reducing maintenance costs.
[0080] Reference Figure 2 As shown, a real-time monitoring system for corrosion status of industrial equipment and an intelligent system for adding corrosion inhibitors are characterized by comprising:
[0081] The system includes modules for future corrosion status of equipment, dynamic multi-dimensional corrosion risk assessment model, adaptive corrosion inhibitor dosing strategy, corrosion protection, and efficient utilization of corrosion inhibitors.
[0082] The device future corrosion status module is used to acquire historical corrosion status data of the device, analyze the trend of corrosion rate change of the device per unit time, and predict the future corrosion status of the device.
[0083] The dynamic multidimensional factor corrosion risk assessment model module is used to obtain the current corrosion status of the equipment based on sensor data, and combine it with the historical corrosion status data of the equipment to construct a dynamic multidimensional factor corrosion risk assessment model for the equipment, and quantify the current corrosion level of the equipment in real time. The multidimensional factors include: equipment material, operating environment and historical data.
[0084] The adaptive corrosion inhibitor dosing strategy module is electrically connected to the dynamic multidimensional factor corrosion risk assessment model module and the equipment future corrosion state module. The adaptive corrosion inhibitor dosing strategy module is used to combine the current corrosion state of the equipment with the future corrosion state of the equipment to generate an adaptive corrosion inhibitor dosing strategy for the equipment, and dynamically associate the equipment corrosion rate with the corrosion inhibitor requirement.
[0085] The corrosion protection and high-efficiency utilization module of the corrosion inhibitor is electrically connected to the adaptive corrosion inhibitor dosing strategy module. The corrosion protection and high-efficiency utilization module of the corrosion inhibitor is used to adjust the dosage and frequency of corrosion inhibitor dosing in real time according to the equipment's adaptive corrosion inhibitor dosing strategy, so as to realize the corrosion protection and high-efficiency utilization of the corrosion inhibitor of the equipment.
[0086] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.
Claims
1. A method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitors for industrial equipment, characterized in that, include: S1. Obtain historical corrosion status data of the equipment, analyze the trend of corrosion rate change of the equipment per unit time, and predict the future corrosion status of the equipment. S2. Based on sensor data, obtain the current corrosion status of the equipment, and combine it with the historical corrosion status data of the equipment to construct a dynamic multi-dimensional factor corrosion risk assessment model for the equipment, and quantify the current corrosion level of the equipment in real time. The multi-dimensional factors include: equipment material, operating environment and historical data. S3. Combine the current corrosion state of the equipment with the future corrosion state of the equipment to generate an adaptive corrosion inhibitor dosing strategy for the equipment, and dynamically link the equipment corrosion rate with the corrosion inhibitor requirement. S4. Based on the equipment's adaptive corrosion inhibitor dosing strategy, adjust the dosage and frequency of corrosion inhibitor dosing in real time to achieve corrosion protection of the equipment and efficient utilization of the corrosion inhibitor.
2. The method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitors for industrial equipment according to claim 1, characterized in that, S1 includes: Based on the historical corrosion status data of the equipment, the data is normalized and outliers are corrected by linear interpolation. As a further detail, the historical corrosion status data of the equipment includes: corrosion rate, temperature, humidity, and chemical composition; Using the sliding window algorithm, a fixed window length is set, and the time series data within the current window is extracted as input features, while the data of the next adjacent time step outside the window is used as labels to construct a time series sample set of equipment corrosion status. Using a time series sample set of equipment corrosion state as input, an LSTM long short-term memory network is trained to capture long-term dependencies in the equipment corrosion state data and output the hidden state at each time step.
3. The method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitors for industrial equipment according to claim 2, characterized in that, S1 further includes: Combining the attention mechanism, based on the hidden state of each time step, the attention weight of the historical corrosion state data of the device at each time step is dynamically calculated. The historical corrosion state data of the device is then weighted and summed to obtain the weighted memory unit of the device at the current moment, as shown in the following formula: Where, α t Here, t represents the attention weights for each time step of the historical corrosion status data of the equipment, Softmax is the normalization function, V is the attention score vector, tanh is the hyperbolic tangent activation function, and W is the attention weight. h With W x h is a learnable weight matrix. t Let x be the hidden state at each time step. t Let b be the input feature vector at time step t, and m be the bias term. t α is the weighted memory unit of the device at the current moment. k h represents the attention weights at time step k. k For time step k, the hidden state; By combining weighted historical corrosion status data of the equipment with the hidden states of the LSTM long short-term memory network, the future corrosion status of the equipment is predicted through regression calculation using fully connected layers.
4. The method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitors for industrial equipment according to claim 1, characterized in that, S2 includes: Based on sensor data, the current corrosion status of the equipment is obtained, including the equipment material, operating environment and historical data, and the data is preprocessed. Training and testing sets were divided based on equipment material, usage environment, and historical data. The Bootstrap sampling method was used to randomly select multiple subsets of the training set to train decision trees. Each subset was split into multiple independent decision trees by randomly selecting some features. A random forest model was trained, and the number of decision trees, maximum depth, and minimum number of leaf node samples were set and adjusted as hyperparameters. The model was optimized by combining cross-validation method to construct a dynamic multi-dimensional factor corrosion risk assessment model for equipment.
5. The method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitors for industrial equipment according to claim 4, characterized in that, S2 further includes: Based on the dynamic multidimensional factor corrosion risk assessment model for equipment, the current corrosion risk of the equipment is calculated in real time, the corrosion level of the equipment is obtained, a corrosion risk threshold range is set according to industry standards, different levels are defined for the current corrosion state of the equipment, and the current corrosion level of the equipment is quantified in real time. The formula is as follows: Where R(q) p Let q be the probability value of corrosion risk for the equipment at time p. p Let p be the input feature vector at time p, σ be the Sigmoid activation function, and e be the input feature vector at time p. i f is the dynamic weight of the i-th tree. i (q p ) represents the original output of the i-th decision tree, ε p For time-varying correction terms; As a further detail, the corrosion levels of the equipment include: mild corrosion, moderate corrosion, and severe corrosion.
6. The method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitors for industrial equipment according to claim 5, characterized in that, S3 includes: Based on the current corrosion state and future corrosion state of the equipment, the corrosion rate, corrosion risk level and environmental parameters of the equipment are defined as the state space, and the current corrosion state of the equipment and the amount of corrosion inhibitor added are defined as the action space to construct a state and action model. Based on the state and action model, a reward function is designed with the goal of minimizing the corrosion rate, ensuring corrosion control within a safe range, and saving on the use of corrosion inhibitors.
7. The method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitors for industrial equipment according to claim 6, characterized in that, S3 further includes: By analyzing the changing trend of equipment corrosion rate, a threshold condition for corrosion inhibitor addition is set. When the equipment corrosion rate exceeds the preset threshold, the intelligent corrosion inhibitor addition mechanism is triggered, dynamically adjusting the amount of corrosion inhibitor added. Through iterative updates to the Q-value table, the agent learns the rewards obtained by taking different actions in different states, thus obtaining the optimal adaptive corrosion inhibitor addition strategy for the equipment. The formula is as follows: in, Given the current corrosion state of the equipment The optimal dosage of corrosion inhibitor. Let Q1(S) represent the corrosion state of the device at time t, and argmax be the action that maximizes the objective function. t A t ) is the Q-value function, representing the state S. t Take action A t The expected long-term returns.
8. The method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitor for industrial equipment according to claim 7, characterized in that, S4 includes: Based on the equipment adaptive corrosion inhibitor dosing strategy, the equipment corrosion rate, corrosion inhibitor demand and actual corrosion rate are used as inputs, and the required amount and frequency of corrosion inhibitor dosing are calculated in real time as outputs. The dosage and frequency of corrosion inhibitor are adjusted in real time by a proportional-integral-derivative controller. The control signal is calculated based on the deviation between the equipment corrosion rate and the corrosion inhibitor requirement, and the dosage of corrosion inhibitor is dynamically adjusted. According to historical dosage data and the changing trend of equipment corrosion status, the frequency of corrosion inhibitor addition is optimized to avoid over-addition or under-addition, thereby achieving corrosion protection of equipment and efficient utilization of corrosion inhibitor.
9. A real-time monitoring system for corrosion status of industrial equipment and an intelligent system for adding corrosion inhibitors, characterized in that, To implement any one of claims 1-8, the method for real-time monitoring of corrosion status and intelligent dosing of corrosion inhibitors for industrial equipment includes: The system includes modules for future corrosion status of equipment, dynamic multi-dimensional corrosion risk assessment model, adaptive corrosion inhibitor dosing strategy, corrosion protection, and efficient utilization of corrosion inhibitors. The device future corrosion status module is used to acquire historical corrosion status data of the device, analyze the trend of corrosion rate change of the device per unit time, and predict the future corrosion status of the device. The dynamic multidimensional factor corrosion risk assessment model module is used to obtain the current corrosion status of the equipment based on sensor data, and combine it with the historical corrosion status data of the equipment to construct a dynamic multidimensional factor corrosion risk assessment model for the equipment, and quantify the current corrosion level of the equipment in real time. The multidimensional factors include: equipment material, operating environment and historical data. The adaptive corrosion inhibitor dosing strategy module is electrically connected to the dynamic multidimensional factor corrosion risk assessment model module and the equipment future corrosion state module. The adaptive corrosion inhibitor dosing strategy module is used to combine the current corrosion state of the equipment with the future corrosion state of the equipment to generate an adaptive corrosion inhibitor dosing strategy for the equipment, and dynamically associate the equipment corrosion rate with the corrosion inhibitor requirement. The corrosion protection and high-efficiency utilization module of the corrosion inhibitor is electrically connected to the adaptive corrosion inhibitor dosing strategy module. The corrosion protection and high-efficiency utilization module of the corrosion inhibitor is used to adjust the dosage and frequency of corrosion inhibitor dosing in real time according to the equipment's adaptive corrosion inhibitor dosing strategy, so as to realize the corrosion protection and high-efficiency utilization of the corrosion inhibitor of the equipment.