Ammonia intelligent flow regulation system
Through real-time data acquisition and dynamic Bayesian network adjustment, the support vector machine model is optimized, and the rapid response of ammonia intelligent flow regulation system to real-time flow changes is achieved, solving the problems of existing system response lag and inaccurate control, and improving the Stability and safety of ammonia supply.
Patent Information
- Application Number
- CN202411237640.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-05
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-09-05
AI Technical Summary
The existing ammonia gas intelligent flow regulation system lacks the ability to quickly adapt to real-time data changes when processing gas flow, resulting in system response lag and inaccurate control, which may cause safety accidents and reduced efficiency.
Data acquisition and real-time monitoring modules are used to collect ammonia flow data, combined with pressure sensor feedback signal, dynamic Bayesian network adjustment module corrects Bayesian network structure and probability distribution, SVM flow prediction and regulation module optimizes support vector machine model, generates predicted flow Control instructions and real-time adjustment of valve position through control execution with feedback module to match predicted needs.
The fine management of ammonia flow is achieved, the system's response speed and accuracy to environmental changes is enhanced, the stability and reliability of ammonia supply is ensured, and safety risks and efficiency reductions are avoided due to improper control.
Smart Images

Figure CN118760244B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of flow regulation, and in particular to an ammonia intelligent flow regulation system. Background Art
[0002] The field of flow regulation technology mainly involves systems and equipment for precise control of the flow of various liquids or gases. This technical field includes the use of various control valves, sensors and control algorithms to regulate the amount of fluid flowing through pipelines or other transmission media. Flow regulation systems can be used in many industrial and civil scenarios, such as chemical production, water treatment, HVAC systems and medical equipment. The key is to maintain or adjust the fluid to a predetermined flow level to optimize performance and energy efficiency and ensure process safety. In the gas supply system, this technology is equally important to ensure that the gas is delivered to the user in a safe and efficient manner.
[0003] Among them, the ammonia intelligent flow control system refers to an intelligent system specifically used to regulate and control the flow of ammonia. This system uses sensors to monitor the flow and pressure of ammonia, and automatically adjusts the control valve or other regulating equipment through advanced control algorithms to accurately control the supply of ammonia. The main uses of the ammonia intelligent flow control system include ammonia treatment processes in industrial refrigeration, chemical production and environmental protection facilities, where precise control of ammonia flow is crucial to ensure system efficiency and safety.
[0004] Existing technologies usually use fixed algorithms and parameters when dealing with gas flow regulation, and lack the ability to quickly adapt to real-time data changes. In industrial applications, due to the rapid changes in environmental conditions and production needs, fixed control strategies cannot immediately reflect changes in actual flow, which may cause system response lags or inaccurate control. The static nature of this control system will lead to reduced efficiency in dynamically changing application scenarios, and may even cause safety accidents due to improper control. For example, in the ammonia treatment process, inaccurate control may cause the ammonia concentration to be too high or too low, affecting the efficiency and safety of the chemical reaction, or causing emission standards to fail to meet standards in environmental protection facilities, causing damage to the environment. These shortcomings highlight the urgent need for dynamic adjustment capabilities and real-time feedback mechanisms in existing technologies. Summary of the invention
[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an ammonia intelligent flow regulation system.
[0006] In order to achieve the above object, the present invention adopts the following technical solution: the ammonia intelligent flow regulation system comprises:
[0007] The data acquisition and real-time monitoring module is based on ammonia production equipment, collects real-time flow data, combines the pressure sensor feedback signal, analyzes the dynamic changes of ammonia flow, extracts key fluctuations and trend data, and generates ammonia flow characteristic signals;
[0008] The dynamic Bayesian network adjustment module uses the ammonia flow characteristic signal to check and correct the nodes and connections of the Bayesian network, recalculate the correlation and conditional dependence between multiple nodes, dynamically update the network structure, adjust the probability distribution, and obtain updated Bayesian network parameters;
[0009] The SVM flow prediction and regulation module applies the updated Bayesian network parameters to adjust and train the support vector machine model, analyzes the currently received ammonia flow data, optimizes the support vector machine model and predicts flow changes, and generates predicted flow control instructions;
[0010] The control execution and feedback module executes the predicted flow control instructions, adjusts the valve control in real time to match the predicted demand, monitors the adjustment effect in real time, and obtains the adjusted flow feedback result.
[0011] As a further solution of the present invention, the step of acquiring the ammonia flow characteristic signal is specifically as follows:
[0012] Flow data is collected from ammonia production equipment in real time, and the raw data is preprocessed using the sliding window average method to reduce the impact of noise. The formula is: ;
[0013] Calculate the moving average of traffic and obtain pre-processed traffic data ;
[0014] in, represents the size of the moving window, represents the i-th flow data point in the window;
[0015] The pre-processed flow data is combined with the feedback signal from the pressure sensor, and the flow data is adjusted according to the pressure change, using the formula: ;
[0016] Calculate and obtain adjusted flow data;
[0017] in, Represents the adjusted traffic data, represents the pressure sensitivity coefficient, Represents the current pressure, Indicates the reference pressure;
[0018] The adjusted flow data is used to extract the characteristic signal of the flow data through logarithmic difference, using the formula: ;
[0019] Extract key fluctuations and trends to obtain characteristic signals of ammonia flow;
[0020] in, Represents the characteristic signal of ammonia flow, Represents the base flow value, represents the adjustment coefficient, Represents the natural logarithm.
[0021] As a further solution of the present invention, the step of obtaining the updated Bayesian network parameters is specifically as follows:
[0022] Starting from the ammonia flow characteristic signal, identify the nodes that need to be adjusted in the Bayesian network, using the formula: ;
[0023] The initial influence of the calculated node is calculated by weight adjustment to generate a node influence evaluation;
[0024] in, It is The adjustment weight of each node indicates the influence of the node in the network. It is the node sensitivity standard, which is used to standardize the influence of traffic characteristic signals;
[0025] The node influence evaluation is used to calculate the relationship between nodes in the network, using the formula: ;
[0026] Calculate the conditional dependencies between nodes and obtain the network dependency structure;
[0027] in, is the network dependency structure, is the standardization coefficient, which assists in adjusting The nonlinear effect of
[0028] According to the network dependency structure, the probability distribution of the Bayesian network is updated using the formula: ;
[0029] Update the probability model to obtain updated Bayesian network parameters;
[0030] in, is the updated Bayesian network parameter, To fine-tune the parameters, used to adjust the impact of traffic signals on probability distribution, Is the reference flow value, used for standardization impact.
[0031] As a further solution of the present invention, the step of obtaining the predicted flow control instruction is specifically as follows:
[0032] The updated Bayesian network parameters are used to adjust the parameters of the support vector machine model using the formula: ;
[0033] Generate adjusted support vector machine parameters;
[0034] in, are the adjusted support vector machine parameters, It is used to adjust The linear influence coefficient of Indicates taking The square root of to reduce its magnitude and increase the model sensitivity, It is adjustment The nonlinear influence coefficient of represents the original parameters of the support vector machine model, Indicates taking The square of is used to enhance the model's response to changes in the original parameters;
[0035] Based on the adjusted support vector machine parameters, the currently received ammonia flow data is analyzed using the formula: ;
[0036] Estimate the fit of the current data to the model parameters and generate a fit assessment;
[0037] in, is the fitness assessment, yes for The standardized ratio uses square root to keep the data scale moderate and avoid excessive influence of extreme values;
[0038] According to the fitness evaluation, the support vector machine model is optimized and the flow change is predicted using the formula: ;
[0039] generating predicted flow control instructions;
[0040] in, is the predicted flow control instruction, is a weighting factor used to balance current forecasts with historical data, Indicates the historical prediction results The square of is used to increase the influence of historical data on the new forecast.
[0041] As a further solution of the present invention, the step of obtaining the adjusted flow feedback result is specifically:
[0042] Implement the predicted flow control instructions and adjust the valve position to match the predicted demand using the formula: ;
[0043] Calculate and generate new valve position;
[0044] in, represents the new valve position, represents the transition coefficient, Represents the current valve position, Used to maintain smoothness of valve adjustment;
[0045] Monitor the valve adjustment effect in real time and calculate the deviation between the actual flow and the target flow using the formula: ;
[0046] Generate bias assessment results ;
[0047] in, represents the deviation ratio, Represents the actual flow rate, Represents target flow, absolute value symbol Used to express the absolute value of the deviation;
[0048] Based on the deviation evaluation results, the new valve setting is adjusted to meet the flow demand, using the formula: ;
[0049] Calculate and obtain the adjusted traffic feedback result;
[0050] in, Represents the adjusted traffic feedback result, represents the regulating factor, Represents the deviation assessment result.
[0051] Compared with the prior art, the advantages and positive effects of the present invention are:
[0052] In the present invention, by real-time collection and monitoring of ammonia flow data, combined with the analysis of dynamic changes, it is possible to effectively capture key fluctuations and trends in the flow, and enhance the system's response speed and accuracy to environmental changes. The Bayesian network structure and probability distribution that are corrected in real time are used to improve the accuracy of the prediction and the adaptability of the system. This optimized prediction mechanism achieves real-time adjustment of valve control through precise flow control instructions, ensuring the fine management of ammonia supply. The feedback module monitors the adjustment effect in real time, and can correct deviations in real time to ensure stability and reliability in long-term operation. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Figure 1 is a system flow chart of the present invention;
[0054] Figure 2 Flow chart of the steps for obtaining the characteristic signal of ammonia flow rate of the present invention;
[0055] Figure 3 A flowchart of the steps for obtaining updated Bayesian network parameters of the present invention;
[0056] Figure 4 A flow chart of the steps for obtaining the flow control instructions predicted by the present invention;
[0057] Figure 5 This is a flow chart of the steps for obtaining the adjusted flow feedback result of the present invention. DETAILED DESCRIPTION
[0058] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0059] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.
[0060] Embodiment 1
[0061] See also Figure 1 , the ammonia intelligent flow regulation system includes:
[0062] The data acquisition and real-time monitoring module is based on ammonia production equipment, collects real-time flow data, combines the pressure sensor feedback signal, analyzes the dynamic changes of ammonia flow, extracts key fluctuations and trend data, and generates ammonia flow characteristic signals;
[0063] The dynamic Bayesian network adjustment module uses the characteristic signal of ammonia flow to check and correct the nodes and connections of the Bayesian network, recalculate the correlation and conditional dependence between multiple nodes, dynamically update the network structure, adjust the probability distribution, and obtain the updated Bayesian network parameters;
[0064] The SVM flow prediction and regulation module applies the updated Bayesian network parameters to adjust and train the support vector machine model, analyzes the currently received ammonia flow data, optimizes the support vector machine model and predicts flow changes, and generates predicted flow control instructions;
[0065] The control execution and feedback module executes the predicted flow control instructions, adjusts the valve control in real time to match the predicted demand, monitors the adjustment effect in real time, and obtains the adjusted flow feedback results.
[0066] The ammonia flow characteristic signal specifically includes flow data, pressure feedback signal, dynamic change analysis, key fluctuation and trend data. The updated Bayesian network parameters specifically refer to node weights, connection strength, and probability distribution. The predicted flow control instructions specifically include valve position adjustment, flow set value, and prediction accuracy. The adjusted flow feedback results specifically include valve control adjustment, predicted demand matching, and effect monitoring data.
[0067] See also Figure 2 , the specific steps for obtaining the characteristic signal of ammonia flow are as follows:
[0068] Flow data is collected from ammonia production equipment in real time, and the raw data is preprocessed using the sliding window average method to reduce the impact of noise. The formula is: ;
[0069] Calculate the moving average of traffic and obtain pre-processed traffic data ;
[0070] in, represents the size of the moving window, represents the i-th flow data point in the window;
[0071] Combine the preprocessed flow data with the feedback signal from the pressure sensor and adjust the flow data according to the pressure change using the formula: ;
[0072] Calculate and obtain adjusted flow data;
[0073] in, Represents the adjusted traffic data, represents the pressure sensitivity coefficient, Represents the current pressure, Indicates the reference pressure;
[0074] Using the adjusted flow data, the characteristic signal of the flow data is extracted by logarithmic difference, using the formula: ;
[0075] Extract key fluctuations and trends to obtain characteristic signals of ammonia flow;
[0076] in, Represents the characteristic signal of ammonia flow, Represents the base flow value, represents the adjustment coefficient, Represents the natural logarithm.
[0077] formula: ;
[0078] in:
[0079] : Preprocessed traffic data, indicating the traffic value after averaging through the sliding window.
[0080] : The size of the moving window determines the number of data points referenced in the average calculation.
[0081] : The i-th flow data point in the window.
[0082] Calculation process example:
[0083] Set a sliding window size , and the traffic data points monitored within a certain period of time for The unit is cubic meter / hour.
[0084] calculate : ;
[0085] The results showed that the average flow rate was 103 cubic meters per hour.
[0086] formula: ;
[0087] in:
[0088] : Adjusted flow data, taking into account the impact of pressure changes.
[0089] : Pressure sensitivity coefficient, which affects the adjustment intensity of flow rate due to pressure changes.
[0090] : Current pressure.
[0091] : Reference pressure.
[0092] Calculation process example:
[0093] set up Cubic meters / hour, pressure , reference pressure , pressure sensitivity coefficient .
[0094] calculate : ;
[0095] The results show that after adjustment of the reference pressure, the flow rate is 104.63 cubic meters per hour.
[0096] formula: ;
[0097] in:
[0098] : Ammonia flow characteristic signal.
[0099] : Adjusted traffic data.
[0100] : Baseline flow value.
[0101] : Adjustment coefficient, used to adjust the sensitivity of logarithmic difference.
[0102] Calculation process example:
[0103] set up Cubic meters / hour, base flow Cubic meters / hour, adjustment factor .
[0104] calculate : ;
[0105] The results show that the flow characteristic signal is 0.2, reflecting the key fluctuations and trends of the flow.
[0106] See also Figure 3 , the specific steps for obtaining the updated Bayesian network parameters are:
[0107] Starting from the characteristic signal of ammonia flow, identify the nodes that need to be adjusted in the Bayesian network, using the formula: ;
[0108] The initial influence of the calculated node is calculated by weight adjustment to generate a node influence evaluation;
[0109] in, It is The adjustment weight of each node indicates the influence of the node in the network. It is the node sensitivity standard, which is used to standardize the influence of traffic characteristic signals;
[0110] The node influence is used to evaluate the relationship between nodes in the computing network, using the formula: ;
[0111] Calculate the conditional dependencies between nodes and obtain the network dependency structure;
[0112] in, is the network dependency structure, is the standardization coefficient, which assists in adjusting The nonlinear effect of
[0113] According to the network dependency structure, the probability distribution of the Bayesian network is updated using the formula: ;
[0114] Update the probability model to obtain updated Bayesian network parameters;
[0115] in, is the updated Bayesian network parameter, To fine-tune the parameters, used to adjust the impact of traffic signals on probability distribution, Is the reference flow value, used for standardization impact.
[0116] formula: ;
[0117] Parameter details:
[0118] : The adjusted weights of each node, which are determined based on the node's historical importance in the network or expert evaluation.
[0119] : Ammonia flow characteristic signal, set by the sensor in the actual factory environment measurement.
[0120] : Node sensitivity criteria, based on previous experimental data or model analysis.
[0121] Calculation example:
[0122] The network is set up with 3 nodes, where:
[0123] ;
[0124] (Set based on past data)
[0125] calculate :
[0126] ;
[0127] formula:
[0128] ;
[0129] Parameter details:
[0130] : Standardized coefficient, determined based on statistical analysis of the data.
[0131] Node influence evaluation.
[0132] Calculation example:
[0133] if ,
[0134] ;
[0135] formula:
[0136] ;
[0137] Parameter details:
[0138] : Network dependency structure.
[0139] : Fine-tune parameters based on recent network performance monitoring.
[0140] : Baseline flow value, set based on the long-term average flow value.
[0141] Calculation example:
[0142] set up ,
[0143] ;
[0144] See also Figure 4 ,The steps for obtaining the predicted flow control instructions are as follows:
[0145] Using the updated Bayesian network parameters, adjust the parameters of the support vector machine model using the formula: ;
[0146] Generate adjusted support vector machine parameters;
[0147] in, are the adjusted support vector machine parameters, It is used to adjust The linear influence coefficient of Indicates taking The square root of to reduce its magnitude and increase the model sensitivity, It is adjustment The nonlinear influence coefficient of represents the original parameters of the support vector machine model, Indicates taking The square of is used to enhance the model's response to changes in the original parameters;
[0148] Based on the adjusted support vector machine parameters, the currently received ammonia flow data is analyzed using the formula: ;
[0149] Estimate the fit of the current data to the model parameters and generate a fit assessment;
[0150] in, is the fitness assessment, yes for The standardized ratio uses square root to keep the data scale moderate and avoid excessive influence of extreme values;
[0151] According to the fitness evaluation, the support vector machine model is optimized and the traffic changes are predicted using the formula: ;
[0152] generating predicted flow control instructions;
[0153] in, is the predicted flow control instruction, is a weighting factor used to balance current forecasts with historical data, Indicates the historical prediction results The square of is used to increase the influence of historical data on the new forecast.
[0154] formula ;
[0155] Parameter details and calculation process
[0156] (Updated Bayesian network parameters):
[0157] How to get it: Set the value obtained through the previous step analysis and machine learning model update. .
[0158] The value represents a certain probability or weight, which has been trained by the network.
[0159] (Statistics of the original parameters of the support vector machine model):
[0160] How to get it: This is calculated from the model training data, setting the average parameter value .
[0161] and :
[0162] Set according to the model adjustment requirements, set The coefficients are set based on historical data to adjust and optimize the model performance.
[0163] Actual operation example
[0164] ;
[0165] Calculated The values are used to iteratively tune the parameters of the support vector machine model.
[0166] formula
[0167] ;
[0168] Parameter details and calculation process
[0169] (Current ammonia flow data):
[0170] Set the current flow rate to unit.
[0171] (Base flow value):
[0172] Set the base flow rate to Unit, which is a standard or desired flow rate.
[0173] Actual operation example
[0174] ;
[0175] The calculated Ω value is used to measure the degree of fit between the current data and the model parameters.
[0176] formula
[0177] ;
[0178] Parameter details and calculation process
[0179] (Previous prediction results):
[0180] Set the last prediction result to .
[0181] and :
[0182] According to system stability and response requirements, set .
[0183] Actual operation example
[0184] ;
[0185] value Represents the flow control forecast instructions adjusted based on current and historical data.
[0186] See also Figure 5 , the specific steps for obtaining the adjusted traffic feedback results are:
[0187] Implement the predicted flow control command and adjust the valve position to match the predicted demand using the formula:
[0188] ;
[0189] Calculate and generate new valve position;
[0190] in, represents the new valve position, represents the transition coefficient, Represents the current valve position, Used to maintain smoothness of valve adjustment;
[0191] Monitor the valve adjustment effect in real time and calculate the deviation between the actual flow and the target flow using the formula:
[0192] ;
[0193] Generate bias assessment results ;
[0194] in, represents the deviation ratio, Represents the actual flow rate, Represents target flow, absolute value symbol Used to express the absolute value of the deviation;
[0195] Based on the deviation evaluation results, adjust the new valve setting to meet the flow demand using the formula:
[0196] ;
[0197] Calculate and obtain the adjusted traffic feedback result;
[0198] in, Represents the adjusted traffic feedback result, represents the regulating factor, Represents the deviation assessment result.
[0199] formula:
[0200] ;
[0201] Parameter explanation and examples
[0202] : Flow control instruction, set by the control system according to the demand calculation, the value is 5.0.
[0203] : Transition coefficient is set to 0.7, which is set according to the response time and control sensitivity of the system.
[0204] : Current valve position, set the current valve position to 3.0. This value is obtained through real-time monitoring by the sensor.
[0205] Calculation process:
[0206] calculate ,
[0207] calculate ,
[0208] Weighted sum: ,
[0209] Normalized denominator: ,
[0210] calculate ;
[0211] Result description:
[0212] New valve position It is about 0.863, which means the valve needs to be adjusted to 86.3% of the maximum opening to match the flow control instruction.
[0213] formula:
[0214] ;
[0215] Parameter explanation and examples
[0216] : Actual flow, set to , this value is obtained through real-time monitoring of the flow meter.
[0217] : Target flow, set to , set according to system requirements.
[0218] Calculation process:
[0219] Calculate the deviation numerator: ,
[0220] Calculate the deviation ratio: ,
[0221] Take the absolute value: ,
[0222] Result description:
[0223] deviation ratio It is 0.167, which means that the actual flow rate is about 16.7% lower than the target flow rate, indicating that iterative adjustment is needed.
[0224] formula:
[0225] ;
[0226] Parameter explanation and examples
[0227] : The new valve position, obtained from the above results, is 0.863.
[0228] : Adjustment factor, set to 0.5, used to adjust the valve position to compensate for deviation.
[0229] Calculation process:
[0230] Calculate the adjustment factor: ,
[0231] Calculate the feedback flow: ,
[0232] Result description:
[0233] Adjusted traffic feedback results It is about 0.796, which means that the valve position has been adjusted according to the deviation to get as close to the target flow as possible.
[0234] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.
Claims
1. Ammonia intelligent flow regulation system, characterized in that: The system comprises: The data acquisition and real-time monitoring module is based on ammonia production equipment, collects real-time flow data, combines the pressure sensor feedback signal, analyzes the dynamic changes of ammonia flow, extracts key fluctuations and trend data, and generates ammonia flow characteristic signals; The dynamic Bayesian network adjustment module uses the ammonia flow characteristic signal to check and correct the nodes and connections of the Bayesian network, recalculate the correlation and conditional dependence between multiple nodes, dynamically update the network structure, adjust the probability distribution, and obtain updated Bayesian network parameters; The steps for obtaining the updated Bayesian network parameters are specifically as follows: Starting from the ammonia flow characteristic signal, identify the nodes that need to be adjusted in the Bayesian network, using the formula: ; The initial influence of the calculated node is calculated by weight adjustment to generate a node influence evaluation; in, It is The adjustment weight of each node indicates the influence of the node in the network. It is the node sensitivity standard, which is used to standardize the influence of traffic characteristic signals; The node influence evaluation is used to calculate the relationship between nodes in the network, using the formula: ; Calculate the conditional dependencies between nodes and obtain the network dependency structure; in, is the network dependency structure, is the standardization coefficient, which assists in adjusting The nonlinear effect of According to the network dependency structure, the probability distribution of the Bayesian network is updated using the formula: ; Update the probability model to obtain updated Bayesian network parameters; in, is the updated Bayesian network parameter, To fine-tune the parameters, used to adjust the impact of traffic signals on probability distribution, Is the reference flow value, used for standardization The impact of The SVM flow prediction and regulation module applies the updated Bayesian network parameters to adjust and train the support vector machine model, analyzes the currently received ammonia flow data, optimizes the support vector machine model and predicts flow changes, and generates predicted flow control instructions; The steps of obtaining the predicted flow control instruction are specifically as follows: The updated Bayesian network parameters are used to adjust the parameters of the support vector machine model using the formula: ; Generate adjusted support vector machine parameters; in, are the adjusted support vector machine parameters, It is used to adjust The linear influence coefficient of Indicates taking The square root of to reduce its magnitude and increase the model sensitivity, It is adjustment The nonlinear influence coefficient of represents the original parameters of the support vector machine model, Indicates taking The square of is used to enhance the model's response to changes in the original parameters; Based on the adjusted support vector machine parameters, the currently received ammonia flow data is analyzed using the formula: ; Estimate the fit of the current data to the model parameters and generate a fit assessment; in, is the fitness assessment, yes for The standardized ratio uses square root to keep the data scale moderate and avoid excessive influence of extreme values; According to the fitness evaluation, the support vector machine model is optimized and the flow change is predicted using the formula: ; generating predicted flow control instructions; in, is the predicted flow control instruction, is a weighting factor used to balance current forecasts with historical data, Indicates the historical prediction results The square of is used to increase the impact of historical data on new forecasts; The control execution and feedback module executes the predicted flow control instructions, adjusts the valve control in real time to match the predicted demand, monitors the adjustment effect in real time, and obtains the adjusted flow feedback result.
2. The ammonia intelligent flow regulation system according to claim 1, characterized in that: The steps for obtaining the ammonia flow characteristic signal are specifically as follows: Flow data is collected from ammonia production equipment in real time, and the raw data is preprocessed using the sliding window average method to reduce the impact of noise. The formula is: ; Calculate the moving average of traffic and obtain pre-processed traffic data ; in, represents the size of the moving window, Represents the first flow data points; The pre-processed flow data is combined with the feedback signal from the pressure sensor, and the flow data is adjusted according to the pressure change, using the formula: ; Calculate and obtain adjusted flow data; in, Represents the adjusted traffic data, represents the pressure sensitivity coefficient, Represents the current pressure, Indicates the reference pressure; The adjusted flow data is used to extract the characteristic signal of the flow data through logarithmic difference, using the formula: ; Extract key fluctuations and trends to obtain characteristic signals of ammonia flow; in, Represents the characteristic signal of ammonia flow, Represents the base flow value, represents the adjustment coefficient, Represents the natural logarithm.
3. The ammonia intelligent flow regulation system according to claim 1, characterized in that: The steps for obtaining the adjusted flow feedback result are specifically as follows: Implement the predicted flow control instructions and adjust the valve position to match the predicted demand using the formula: ; Calculate and generate new valve position; in, represents the new valve position, represents the transition coefficient, Represents the current valve position, Used to maintain smoothness of valve adjustment; Monitor the valve adjustment effect in real time and calculate the deviation between the actual flow and the target flow using the formula: ; Generate bias assessment results ; in, represents the deviation ratio, Represents the actual flow rate, Represents target flow, absolute value symbol Used to express the absolute value of the deviation; Based on the deviation evaluation results, the new valve setting is adjusted to meet the flow demand, using the formula: ; Calculate and obtain the adjusted traffic feedback result; in, Represents the adjusted traffic feedback result, represents the regulating factor, Represents the deviation assessment result.
