Method and control system for dynamically adjusting load of air separation system based on liquid air separation tank level

Abstract

The application discloses a method and a control system for dynamically adjusting load of an air separation system based on liquid level of a liquid air separation tank, relates to the field of liquid level control of air separation tanks, and acquires multi-source data such as liquid level, three-dimensional acceleration of the tank, pressure and operation parameters of the air separation system through a sensor group, obtains accurate liquid level values through synchronous processing and dynamic compensation calibration, generates corrected liquid level values in abnormal states, generates load adjustment and valve opening control signals through an adaptive PID algorithm based on liquid level deviation and change slope, realizes closed-loop control, and the control system comprises data acquisition, processing, control decision, execution, data storage and feedback units; the application improves system operation safety and stability, reduces energy consumption and manual intervention, enhances adaptability to complex working conditions, and effectively avoids accidents such as tank overflow and air extraction.

Description

Technical Field

[0001] This invention relates to the field of air separation tank level control technology, specifically to a method and control system for dynamically adjusting the load of an air separation system based on the liquid level of the air separation tank. Background Technology

[0002] Air separation systems are core infrastructure equipment in industries such as chemical and metallurgical processing. The products they produce, such as liquid oxygen and liquid nitrogen, need to be stored in liquid air separation tanks. The control of the tank liquid level directly affects the system's operational safety and production efficiency.

[0003] In existing technologies, liquid level control in liquid air separation storage tanks mostly adopts fixed threshold alarms or manual adjustment modes, which have many drawbacks: First, the response is lagging, and manual intervention is required after the liquid level exceeds the limit. It is impossible to adjust the production load in real time, which can easily lead to safety accidents such as tank overflow and cavitation, damaging the equipment and wasting the medium. Second, the load matching is unbalanced. It relies on manual experience to adjust the load and cannot dynamically adapt the production rate according to the real-time storage needs of the tank, resulting in ineffective energy consumption and increased production costs.

[0004] Therefore, there is an urgent need to develop a method and control system that can dynamically and accurately adjust the load of the air separation system based on the real-time liquid level status of the liquid air separation storage tank and combined with multi-source operating parameters, so as to eliminate problems such as response lag, excessive energy consumption and low degree of automation. Summary of the Invention

[0005] The purpose of this invention is to provide a method and control system for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank, characterized by comprising the following steps: Step 1: Collect real-time liquid level data, motion environment data, and air separation system operating parameters of the liquid air separation storage tank using a sensor array. The motion environment data includes the three-dimensional acceleration of the storage tank. , , ) and real-time pressure data at the bottom of the storage tank and above the liquid surface ( , The operating parameters of the air separation system include fluid flow rate, temperature, pressure, and fluid density. ); Step 2: Perform time synchronization and unification processing on the real-time liquid level data, motion environment data and air separation system operating parameters to generate a standardized dataset. Based on the fluid statics and dynamic compensation model combined with historical data, dynamically adjusted compensation factors are used to compensate and calibrate the real-time liquid level data to obtain accurate liquid level values. Step 3: Extract feature parameters representing the operating status of the air separation system based on the standardized dataset, generate a status parameter set, and generate a correction liquid level value when at least one indicator in the status parameter set meets the anomaly judgment condition. Step 4: Based on the deviation between the accurate liquid level value or the corrected liquid level value and the preset target liquid level range, and combined with the liquid level change slope, dynamically calculate the load adjustment command and valve opening control signal using an adaptive PID algorithm; Step 5: Control the production load of each unit of the air separation system according to the load adjustment command, and adjust the opening of the fluid input / output valves of the storage tank according to the valve opening control signal to form a closed-loop control.

[0007] Preferably, in step 1, the specific steps for collecting real-time liquid level data include: Step 1.1: After comparing the collected raw liquid level data with two liquid levels to determine the authenticity of the liquid level, take the average value or discard false liquid levels and take the true value. Step 1.2: Collect real-time pressure data at the bottom of the storage tank and above the liquid surface ( , ), combined with the distance data between the liquid surface ( ), synchronously collect three-dimensional acceleration data from the motion environment ( , , ) and fluid density in the operating parameters of the air separation system ( The pressure data is filtered, the optical ranging data is temperature compensated, and the height difference between the real liquid surface and the static liquid surface is calculated based on the hydrostatic and dynamic compensation model to compensate the original liquid level data.

[0008] Preferably, in step 1.2.1, based on the fundamental equations of fluid statics, the pressure difference corresponding to the liquid level height under static conditions is calculated: ,in The height of the still liquid level. It is the acceleration due to gravity. Real-time fluid density; Step 1.2.2: Under dynamic conditions, calculate the relationship between the actual pressure difference and the actual liquid level height based on the combined acceleration: ,in The composite acceleration of the storage tank ( ), Given the actual liquid level, we can then deduce the theoretical height difference: ; Steps 1.2.3: Calculate the dynamic compensation factor using multiple sets of historical data. Historical data includes historical pressure data, historical optical ranging data, historical motion environment data, and historical fluid density data. The calculation formula is as follows: ,in The number of historical data sets. For the first The actual height difference in the historical data set. , , , The first Fluid density, composite acceleration, dynamic pressure difference, and static pressure difference from a set of historical data; Step 1.2.4: Correct the theoretical height difference using a dynamic compensation factor to obtain the final height difference: And based on this height difference, the original liquid level data is compensated.

[0009] Preferably, in step 3, feature parameters characterizing the operating state of the spatial division system are extracted based on the standardized dataset to generate a state parameter set, including: Step 3.1: Extract the characteristic values ​​of instantaneous liquid flow rate, liquid level height, temperature, pressure, water content trend coefficient, and liquid level fluctuation coefficient from the standardized dataset; Step 3.2: Normalize the eigenvalues ​​and perform dimensionality reduction using principal component analysis to construct a low-dimensional flow state representation vector; Step 3.3: Compare the low-dimensional flow state representation vector with the preset normal operation state reference model to generate a state parameter set including gas-liquid ratio change index, state deviation index and flow stability index.

[0010] Preferably, in step 3, when at least one indicator in the set of state parameters meets the anomaly determination condition, a correction liquid level value is generated, including: Step 3.4: Set up abnormal judgment conditions. The abnormal judgment conditions include gas-liquid ratio change exceeding a preset threshold, state deviation index exceeding a preset threshold, liquid level fluctuation coefficient exceeding a set limit, and pressure parameter exceeding a preset pressure threshold. Step 3.5: After determining that the state is abnormal, extract the liquid level height and fluid temperature at the corresponding time, and construct a multi-dimensional input vector with the liquid level change data in multiple adjacent metering cycles; Step 3.6: Input the multidimensional input vector into the nonlinear liquid level compensation model trained based on historical operating data, and perform weighted correction by combining the weighting factors calculated by the temperature change rate and the liquid level change gradient to obtain the corrected liquid level value.

[0011] Preferably, step 4 involves dynamically calculating the load adjustment command and valve opening control signal using an adaptive PID algorithm, including: Step 4.1: Construct a closed-loop control structure, set the median value of the preset target liquid level range as the reference liquid level, and calculate the deviation between the precise liquid level value or the corrected liquid level value and the reference liquid level as the control error signal; Step 4.2: Based on the control error signal, call the PID control model to calculate the proportional response, integral response, and derivative prediction, and then weight them to obtain the initial control signal. Step 4.3: Based on the current liquid level change rate, gas-liquid flow pattern type, and historical error change trend, dynamically adjust the gain coefficient of the PID control model to optimize the initial control signal; Step 4.4: If the control error is detected to fluctuate within a range exceeding the preset tolerance range during the continuous metering cycle, the anti-integral saturation mechanism and variable step size adjustment strategy are invoked to dynamically correct the optimized control signal and generate load adjustment commands and valve opening control signals.

[0012] Preferably, step 5, which involves controlling the production load of each unit of the air separation system according to the load adjustment command, includes: Step 5.1: Preset multi-level liquid level control thresholds, which include lower limit thresholds and upper limit thresholds; Step 5.2: When the liquid level drops to the lower limit of the first stage threshold, instruct the air separation unit to increase production by the first preset ratio and shut down some standby equipment; Step 5.3: When the liquid level drops to the lower limit of the second stage threshold, the air separation unit is instructed to increase production by a second preset ratio, which is less than the first preset ratio. Step 5.4: When the liquid level rises back to the intermediate threshold, the system automatically restores the standard load; Step 5.5: When the liquid level approaches the upper limit of the first stage threshold, instruct the air separation unit to reduce production by the first preset ratio; Step 5.6: When the liquid level approaches the upper limit of the second stage threshold, instruct the air separation unit to reduce production by the second preset ratio, trigger an alarm and start the emergency discharge process.

[0013] This invention also provides a control system for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank, comprising: The data acquisition unit is used to acquire real-time liquid level data and three-dimensional acceleration data of the liquid air separation storage tank through a sensor array. , , Real-time pressure data at the bottom of the storage tank and above the liquid surface ( , ), fluid flow rate, temperature, pressure, and fluid density ( The collected multi-source data is then transmitted to the data processing unit. The data processing unit is used to perform time synchronization and unification processing on the received real-time liquid level raw data, motion environment data and air separation system operating parameters, generate a standardized dataset, and perform compensation calibration on the real-time liquid level data based on the dynamic compensation factor calculated by combining fluid statics and dynamic compensation models with historical data to obtain accurate liquid level values. The control decision unit dynamically calculates the load adjustment command and valve opening control signal through an adaptive PID algorithm based on the deviation between the accurate liquid level value or the corrected liquid level value and the preset target liquid level range, combined with the liquid level change slope. The execution unit includes an air separation system load regulation module and a valve opening regulation module. The air separation system load regulation module is used to control the production load of each unit of the air separation system, and the valve opening regulation module is used to adjust the opening of the fluid input / output valve of the storage tank according to the valve opening control signal.

[0014] Preferably, it also includes a data storage and feedback unit, which is used to store historical operating condition data, standardized datasets, state parameter sets, control commands and liquid level change feedback data, providing data for dynamic compensation factor calculation, nonlinear liquid level compensation model training and PID control model optimization, and feeding back the operating status of the execution unit to the control decision unit in real time to form closed-loop control.

[0015] Preferably, the sensor group consists of a liquid level sensor, a three-dimensional acceleration sensor, a pressure sensor, a flow sensor, a temperature sensor, and a density sensor.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention improves the safety and stability of system operation through dynamic linkage and multi-source parameter fusion control of liquid level and air separation system load, effectively avoiding accidents such as tank overflow and cavitation. Its automated closed-loop control reduces manual intervention and lowers the risk of operational errors. Dynamic load matching optimizes energy distribution, reduces ineffective energy consumption, and lowers unit product cost. At the same time, the combination of dynamic compensation and anomaly correction mechanisms improves the accuracy of liquid level detection, ensures that load adjustment is accurately adapted to working conditions, and enhances the system's adaptability to complex working conditions. Attached Figure Description

[0017] Figure 1 This is a flowchart of a method for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank, according to an embodiment of the present invention.

[0018] Figure 2 This is a system block diagram of a control system for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank, according to an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 and Figure 2 The embodiments of the present invention provide a method and control system for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank, including the following steps: Step 1: Collect real-time liquid level data, motion environment data, and air separation system operating parameters of the liquid air separation storage tank using a sensor array. The motion environment data includes the three-dimensional acceleration of the storage tank. , , ) and real-time pressure data at the bottom of the storage tank and above the liquid surface ( , The operating parameters of the air separation system include fluid flow rate, temperature, pressure, and fluid density. ).

[0021] In the data acquisition and preliminary processing in step 1, the sensors consist of a liquid level sensor, a three-dimensional accelerometer, a pressure sensor, a flow sensor, a temperature sensor, and a density sensor. It is recommended that a high-precision non-contact radar level gauge be used for the liquid level sensor and installed at the center of the top of the storage tank. A MEMS triaxial accelerometer is selected as the three-dimensional accelerometer and placed close to the bottom of the outer wall of the storage tank. The pressure, flow, temperature, pressure, and density sensors are integrated on the fluid pipeline and the data is synchronously transmitted through industrial Ethernet. The raw data is preprocessed by low-pass filtering and moving average.

[0022] Furthermore, the collected raw liquid level data is compared with two liquid levels to determine the authenticity of the liquid level, and then the average value is taken, or the true value is taken after discarding false liquid levels. The real-time pressure data collected at the bottom of the storage tank and above the liquid surface are also included. , ), combined with the distance data between the liquid surface ( ), synchronously collect three-dimensional acceleration data from the motion environment ( , , ) and fluid density in the operating parameters of the air separation system ( The pressure data is filtered, the optical ranging data is temperature compensated, and the height difference between the real liquid surface and the static liquid surface is calculated based on the hydrostatic and dynamic compensation model to compensate the original liquid level data.

[0023] Specifically, based on the fundamental equations of fluid statics, calculate the pressure difference corresponding to the height of the liquid surface in a static state: ,in The height of the still liquid level. It is the acceleration due to gravity. Real-time fluid density; In a dynamic environment, the relationship between the actual pressure difference and the true liquid level is calculated based on the combined acceleration: ,in The composite acceleration of the storage tank ( ), Given the actual liquid level, we can then deduce the theoretical height difference: ; Dynamic compensation factor is calculated using multiple sets of historical data. Historical data includes historical pressure data, historical optical ranging data, historical motion environment data, and historical fluid density data. The calculation formula is as follows: ,in The number of historical data sets. For the first The actual height difference in the historical data set. , , , The first Fluid density, composite acceleration, dynamic pressure difference, and static pressure difference from a set of historical data; The final height difference is obtained by correcting the theoretical height difference using a dynamic compensation factor: And based on this height difference, the original liquid level data is compensated.

[0024] Step 2: Perform time synchronization and unification processing on the real-time liquid level data, motion environment data and air separation system operating parameters to generate a standardized dataset. Based on the fluid statics and dynamic compensation model combined with historical data, dynamically adjusted compensation factors are used to compensate and calibrate the real-time liquid level data to obtain accurate liquid level values. Step 2 involves data standardization and liquid level compensation. The real-time liquid level data, motion environment data, and air separation system operating parameters are synchronized in time and converted to a unified unit. The data is then normalized using the Z-score method to construct a standardized dataset. Based on the hydrostatic and dynamic compensation model, and combined with real-time three-dimensional acceleration, pressure difference, and fluid density, the liquid level compensation amount under dynamic conditions is calculated.

[0025] Step 3: Extract feature parameters representing the operating status of the air separation system based on the standardized dataset, generate a status parameter set, and generate a correction liquid level value when at least one indicator in the status parameter set meets the anomaly judgment condition. In step 3, feature parameters characterizing the operating state of the spatial division system are extracted based on the standardized dataset to generate a state parameter set, including: Extract the characteristic values ​​of instantaneous liquid flow rate, liquid level height, temperature, pressure, water content trend coefficient, and liquid level fluctuation coefficient from the standardized dataset; The eigenvalues ​​are normalized and dimensionality is reduced using principal component analysis to construct a low-dimensional flow state representation vector. The low-dimensional flow state representation vector is compared with a preset normal operation state reference model to generate a set of state parameters including gas-liquid ratio change index, state deviation index, and flow stability index.

[0026] Furthermore, in step 3, when at least one indicator in the set of state parameters meets the anomaly determination condition, a correction liquid level value is generated, including: Step 3.4: Set up abnormal judgment conditions. The abnormal judgment conditions include gas-liquid ratio change exceeding a preset threshold, state deviation index exceeding a preset threshold, liquid level fluctuation coefficient exceeding a set limit, and pressure parameter exceeding a preset pressure threshold. Step 3.5: After determining an abnormal state, extract the liquid level height and fluid temperature at the corresponding time, and construct a multi-dimensional input vector with the liquid level change data in the three adjacent metering cycles (each metering cycle is 10 seconds). ,in Current liquid level height Given the current fluid temperature, For the first The liquid level change over a historical period; Step 3.6: Input the multidimensional input vector into the nonlinear liquid level compensation model trained based on historical operating data. The nonlinear liquid level compensation model is an LSTM neural network with 5 input layer nodes, 16 hidden layer nodes, and 1 output layer node. The model is trained 1000 times, and the loss function is the mean squared error (MSE), combined with the temperature change rate. With liquid level change gradient Calculate the weighting factor ( After weight normalization , ), through formula A weighted correction is performed to obtain the corrected liquid level value.

[0027] Step 3 extracts feature values ​​such as instantaneous liquid flow rate, liquid level, temperature, pressure, water content trend coefficient, and liquid level fluctuation coefficient from a standardized dataset. After normalization, principal component analysis is used to reduce the dimensionality and generate a low-dimensional flow state representation vector. This vector is compared with a preset normal operation reference model to generate a set of state parameters such as gas-liquid ratio change, state deviation, and flow stability. If any parameter exceeds the anomaly judgment threshold, a nonlinear liquid level compensation model is triggered. The corrected liquid level value is calculated by combining the multidimensional input vector and weighting factors to improve the measurement robustness under abnormal operating conditions.

[0028] Step 4: Based on the deviation between the accurate liquid level value or the corrected liquid level value and the preset target liquid level range, and combined with the liquid level change slope, dynamically calculate the load adjustment command and valve opening control signal using an adaptive PID algorithm; Step 4 involves dynamically calculating the load adjustment command and valve opening control signal using an adaptive PID algorithm, including: A closed-loop control structure is constructed, the median value of the preset target liquid level range is set as the reference liquid level, and the deviation between the precise liquid level value or the corrected liquid level value and the reference liquid level is calculated as the control error signal. Based on the control error signal, the PID control model is invoked to calculate the proportional response, integral response, and derivative prediction, and the weighted synthesis is used to obtain the initial control signal. Based on the current rate of liquid level change, the type of gas-liquid flow pattern, and the historical error trend, the gain coefficient of the PID control model is dynamically adjusted to optimize the initial control signal. If the control error is detected to fluctuate beyond the preset tolerance range within a continuous metering cycle, the anti-integral saturation mechanism and variable step size adjustment strategy are invoked to dynamically correct the optimized control signal and generate load adjustment commands and valve opening control signals. In the adaptive PID control strategy, the system uses the median value of the target liquid level range as a reference to calculate the deviation of the precise liquid level value or the corrected liquid level value, and combines the liquid level change slope as the control error signal. Through the adaptive PID algorithm, the gain coefficient of the PID control model is dynamically adjusted according to the current liquid level change rate, gas-liquid flow pattern type and historical error trend to optimize the initial control signal. Finally, a smooth and responsive load adjustment command and valve opening control signal are generated to achieve precise closed-loop regulation.

[0029] Step 5: Control the production load of each unit of the air separation system according to the load adjustment command, and adjust the opening of the fluid input / output valves of the storage tank according to the valve opening control signal to form a closed-loop control. Among them, a multi-level liquid level control threshold is preset, and the multi-level liquid level control threshold includes a lower limit tier threshold and an upper limit tier threshold. When the liquid level drops to the lower limit of the first stage threshold, the air separation unit is instructed to increase production by the first preset ratio and shut down some standby equipment. When the liquid level drops to the lower limit of the second stage threshold, the air separation unit is instructed to increase production by a second preset ratio, which is less than the first preset ratio. When the liquid level rises back to the intermediate threshold, the system automatically resumes the standard load; When the liquid level approaches the upper limit of the first grade threshold, the air separation unit is instructed to reduce production by the first preset ratio. When the liquid level approaches the upper limit of the second stage threshold, the air separation unit is instructed to reduce production by the second preset ratio, triggering an alarm and initiating an emergency discharge process.

[0030] In this embodiment, the lower limit first tier threshold is set to 30%, the lower limit second tier threshold is set to 40%, the middle threshold is set to 50%, the upper limit first tier threshold is set to 70%, and the upper limit second tier threshold is set to 80% in the multi-level liquid level control threshold system. After setting the multi-level liquid level control thresholds, the specific control process for the multi-level liquid level control thresholds is shown in the following example: When the liquid level drops to 30%, the controller instructs the air separation unit to increase production by 20% while shutting down some standby equipment. When the liquid level drops to 40%, the controller instructs the air separation unit to increase production by 10%. Once the liquid level rises to 50%, the system automatically resumes standard load. When the liquid level approaches 70%, the controller instructs the air separation unit to reduce production by 10%. When the liquid level approaches 80%, the controller instructs the air separation unit to reduce production by 20%, while simultaneously triggering an alarm and initiating an emergency discharge process.

[0031] Results Verification: After implementation, overflow accidents in storage tanks decreased by 90%, and energy consumption per unit product decreased by 5%.

[0032] In summary, by dynamically linking the liquid level and air separation system load and integrating multi-source parameters for control, the system's operational safety and stability are improved, effectively preventing accidents such as tank overflow and cavitation. Its automated closed-loop control reduces manual intervention and the risk of operational errors, while dynamic load matching optimizes energy distribution, reduces ineffective energy consumption, and lowers unit product costs. At the same time, the combination of dynamic compensation and anomaly correction mechanisms improves the accuracy of liquid level detection, ensuring precise load adjustment to adapt to operating conditions and enhancing the system's adaptability to complex operating conditions.

[0033] All parts not described in this invention are the same as or can be implemented using existing technology. Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank, characterized in that, Includes the following steps: Step 1: Collect real-time liquid level data, motion environment data, and air separation system operating parameters of the liquid air separation storage tank using a sensor array. The motion environment data includes the three-dimensional acceleration of the storage tank. , , ) and real-time pressure data at the bottom of the storage tank and above the liquid surface ( , The operating parameters of the air separation system include fluid flow rate, temperature, pressure, and fluid density. ); Step 2: Perform time synchronization and unification processing on the real-time liquid level data, motion environment data and air separation system operating parameters to generate a standardized dataset. Based on the fluid statics and dynamic compensation model combined with historical data, dynamically adjusted compensation factors are used to compensate and calibrate the real-time liquid level data to obtain accurate liquid level values. Step 3: Extract feature parameters representing the operating status of the air separation system based on the standardized dataset, generate a status parameter set, and generate a correction liquid level value when at least one indicator in the status parameter set meets the anomaly judgment condition. Step 4: Based on the deviation between the accurate liquid level value or the corrected liquid level value and the preset target liquid level range, and combined with the liquid level change slope, dynamically calculate the load adjustment command and valve opening control signal using an adaptive PID algorithm; Step 5: Control the production load of each unit of the air separation system according to the load adjustment command, and adjust the opening of the fluid input / output valves of the storage tank according to the valve opening control signal to form a closed-loop control.

2. The method for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank according to claim 1, characterized in that: In step 1, the specific steps for collecting real-time liquid level data include: Step 1.1: After comparing the collected raw liquid level data with two liquid levels to determine the authenticity of the liquid level, take the average value or discard false liquid levels and take the true value. Step 1.2: Collect real-time pressure data at the bottom of the storage tank and above the liquid surface ( , ), combined with the distance data between the liquid surface ( ), synchronously collect three-dimensional acceleration data from the motion environment ( , , ) and fluid density in the operating parameters of the air separation system ( The pressure data is filtered, the optical ranging data is temperature compensated, and the height difference between the real liquid surface and the static liquid surface is calculated based on the hydrostatic and dynamic compensation model to compensate the original liquid level data.

3. The method for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank according to claim 1, characterized in that: Step 1.2.1: Based on the fundamental equations of fluid statics, calculate the pressure difference corresponding to the liquid level height in a static state: ,in The height of the still liquid level. It is the acceleration due to gravity. Real-time fluid density; Step 1.2.2: Under dynamic conditions, calculate the relationship between the actual pressure difference and the actual liquid level height based on the combined acceleration: ,in The composite acceleration of the storage tank ( ), Given the actual liquid level, we can then deduce the theoretical height difference: ; Steps 1.2.3: Calculate the dynamic compensation factor using multiple sets of historical data. Historical data includes historical pressure data, historical optical ranging data, historical motion environment data, and historical fluid density data. The calculation formula is as follows: ,in The number of historical data sets. For the first The actual height difference in the historical data set. , , , The first Fluid density, composite acceleration, dynamic pressure difference, and static pressure difference from a set of historical data; Step 1.2.4: Correct the theoretical height difference using a dynamic compensation factor to obtain the final height difference: And based on this height difference, the original liquid level data is compensated.

4. The method for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank according to claim 1, characterized in that: In step 3, feature parameters characterizing the operating state of the spatial division system are extracted based on the standardized dataset to generate a state parameter set, including: Step 3.1: Extract the characteristic values ​​of instantaneous liquid flow rate, liquid level height, temperature, pressure, water content trend coefficient, and liquid level fluctuation coefficient from the standardized dataset; Step 3.2: Normalize the eigenvalues ​​and perform dimensionality reduction using principal component analysis to construct a low-dimensional flow state representation vector; Step 3.3: Compare the low-dimensional flow state representation vector with the preset normal operation state reference model to generate a state parameter set including gas-liquid ratio change index, state deviation index and flow stability index.

5. The method for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank according to claim 1, characterized in that: In step 3, when at least one indicator in the set of state parameters meets the anomaly determination condition, a correction liquid level value is generated, including: Step 3.4: Set up abnormal judgment conditions. The abnormal judgment conditions include gas-liquid ratio change exceeding a preset threshold, state deviation index exceeding a preset threshold, liquid level fluctuation coefficient exceeding a set limit, and pressure parameter exceeding a preset pressure threshold. Step 3.5: After determining that the state is abnormal, extract the liquid level height and fluid temperature at the corresponding time, and construct a multi-dimensional input vector with the liquid level change data in multiple adjacent metering cycles; Step 3.6: Input the multidimensional input vector into the nonlinear liquid level compensation model trained based on historical operating data, and perform weighted correction by combining the weighting factors calculated by the temperature change rate and the liquid level change gradient to obtain the corrected liquid level value.

6. The method for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank according to claim 1, characterized in that: Step 4 involves dynamically calculating the load adjustment command and valve opening control signal using an adaptive PID algorithm, including: Step 4.1: Construct a closed-loop control structure, set the median value of the preset target liquid level range as the reference liquid level, and calculate the deviation between the precise liquid level value or the corrected liquid level value and the reference liquid level as the control error signal; Step 4.2: Based on the control error signal, call the PID control model to calculate the proportional response, integral response, and derivative prediction, and then weight them to obtain the initial control signal. Step 4.3: Based on the current liquid level change rate, gas-liquid flow pattern type, and historical error change trend, dynamically adjust the gain coefficient of the PID control model to optimize the initial control signal; Step 4.4: If the control error is detected to fluctuate within a range exceeding the preset tolerance range during the continuous metering cycle, the anti-integral saturation mechanism and variable step size adjustment strategy are invoked to dynamically correct the optimized control signal and generate load adjustment commands and valve opening control signals.

7. The method for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank according to claim 1, characterized in that: Step 5, which involves controlling the production load of each unit in the air separation system according to the load adjustment command, includes: Step 5.1: Preset multi-level liquid level control thresholds, which include lower limit thresholds and upper limit thresholds; Step 5.2: When the liquid level drops to the lower limit of the first stage threshold, instruct the air separation unit to increase production by the first preset ratio and shut down some standby equipment; Step 5.3: When the liquid level drops to the lower limit of the second stage threshold, the air separation unit is instructed to increase production by a second preset ratio, which is less than the first preset ratio. Step 5.4: When the liquid level rises back to the intermediate threshold, the system automatically restores the standard load; Step 5.5: When the liquid level approaches the upper limit of the first stage threshold, instruct the air separation unit to reduce production by the first preset ratio; Step 5.6: When the liquid level approaches the upper limit of the second stage threshold, instruct the air separation unit to reduce production by the second preset ratio, trigger an alarm and start the emergency discharge process.

8. A control system for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank, characterized in that, include: The data acquisition unit is used to acquire real-time liquid level data and three-dimensional acceleration data of the liquid air separation storage tank through a sensor array. , , Real-time pressure data at the bottom of the storage tank and above the liquid surface ( , ), fluid flow rate, temperature, pressure, and fluid density ( The collected multi-source data is then transmitted to the data processing unit. The data processing unit is used to perform time synchronization and unification processing on the received real-time liquid level raw data, motion environment data and air separation system operating parameters, generate a standardized dataset, and perform compensation calibration on the real-time liquid level data based on the dynamic compensation factor calculated by combining fluid statics and dynamic compensation models with historical data to obtain accurate liquid level values. The control decision unit dynamically calculates the load adjustment command and valve opening control signal through an adaptive PID algorithm based on the deviation between the accurate liquid level value or the corrected liquid level value and the preset target liquid level range, combined with the liquid level change slope. The execution unit includes an air separation system load regulation module and a valve opening regulation module. The air separation system load regulation module is used to control the production load of each unit of the air separation system, and the valve opening regulation module is used to adjust the opening of the fluid input / output valve of the storage tank according to the valve opening control signal.

9. The control system for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank according to claim 1, characterized in that: It also includes a data storage and feedback unit, which stores historical operating data, standardized datasets, state parameter sets, control commands, and liquid level change feedback data. This provides data for dynamic compensation factor calculation, nonlinear liquid level compensation model training, and PID control model optimization. It also feeds back the operating status of the execution unit to the control decision unit in real time, forming a closed-loop control.

10. The control system for dynamically adjusting the load of an air separation system based on the liquid level of a liquid air separation storage tank according to claim 1, characterized in that: The sensor group consists of a liquid level sensor, a three-dimensional acceleration sensor, a pressure sensor, a flow sensor, a temperature sensor, and a density sensor.

Images