CN-121995893-A - DCS automatic control system of natural gas purification station technical process

CN121995893ACN 121995893 ACN121995893 ACN 121995893ACN-121995893-A

Abstract

The invention relates to the technical field of natural gas purification and industrial automation control, in particular to a DCS automatic control system of a natural gas purification station process flow, which comprises a data acquisition step, a manifold mapping step, a trend prediction step, a steady state navigation step and a path deviation correction and execution step, wherein the data acquisition step is used for acquiring process monitoring data, the manifold mapping step is used for mapping the data into a real-time state tensor and judging steady state or abrupt evolution signals by utilizing manifold curvature, the trend prediction step is used for analyzing jacobian field distribution information to feed back future track deviation when generating abrupt signals, the steady state navigation step is used for acquiring optimal path coefficients based on geodesic distances when generating steady state signals, generating navigation correction coefficients, and the path deviation correction and execution step is used for quantitatively matching the correction coefficients by utilizing a covariant derivative to generate dynamic set values and issuing maintenance or adjustment instructions.

Inventors

DU RUIHAO
ZHAO YAXU
BI LIANG
YU HENG

Assignees

益通天然气股份有限公司

Dates

Publication Date: 20260508
Application Date: 20260410

Claims (8)

1. The DCS automatic control system for the natural gas purification station process flow is characterized by comprising a data acquisition center, a manifold mapping unit, a trend prediction unit, a steady-state navigation unit, a path deviation correcting unit and a closed-loop execution unit; the data acquisition center is used for acquiring process monitoring data of the natural gas purification station, sending the process monitoring data to the manifold mapping unit for Riemann manifold space mapping analysis to obtain a real-time state tensor, and judging and processing manifold curvature values of the real-time state tensor to obtain a steady state evolution signal or a mutation evolution signal; when the abrupt evolution signal is generated, the trend prediction unit is used for carrying out future track deviation evaluation feedback analysis on the collected jacobian field distribution information of the real-time state tensor to obtain a safe track signal or a divergent track signal; When a steady state evolution signal is generated, the steady state navigation unit is used for carrying out optimal path coefficient acquisition analysis on the acquired geodesic distance information of the real-time state tensor, processing the acquired energy consumption characteristic value and quality characteristic value to obtain a navigation correction coefficient, and the path correction unit is used for carrying out covariant derivative control quantitative matching analysis on the acquired navigation correction coefficient, and comparing and analyzing the acquired dynamic set value to obtain a holding instruction or an adjusting instruction.
2. The DCS automatic control system of a natural gas purification plant process flow of claim 1, wherein said riman manifold space mapping analysis process is as follows: the method comprises the steps of collecting the current sampling time of a natural gas purification station, setting the current sampling time of the natural gas purification station as a time anchor point, setting each sensor in the natural gas purification station as a data node, obtaining physical monitoring values of each data node in the time anchor point, wherein the physical monitoring values represent temperature, pressure, flow and component concentration, and performing tensor processing on the physical monitoring values of the data nodes; If a high-dimensional tensor of the physical monitor value of the data node is generated, a manifold projection instruction is generated, and simultaneously setting the high-dimensional tensor corresponding to the manifold projection instruction as a real-time state tensor, and if the physical monitoring value of the data node is not generated, generating an abnormal alarm instruction.
3. The DCS automatic control system of the process flow of a natural gas purification plant of claim 2, wherein the preset curvature threshold of the real-time state tensor in the time anchor point is obtained, and the module length of the rich curvature tensor in the manifold region where the real-time state tensor is located in the time anchor point is obtained, and the module length of the rich curvature tensor is set as the current curvature value; And setting a value obtained by subtracting a preset curvature threshold from the current curvature value as a curvature deviation value, judging the curvature deviation value, generating a sudden change evolution signal if the curvature deviation value is greater than zero, and generating a steady state evolution signal if the curvature deviation value is less than or equal to zero.
4. The DCS automatic control system of a natural gas purification plant process flow of claim 1, wherein said future track deviation assessment feedback analysis process is as follows: acquiring tangential vector information of a real-time state tensor in a time anchor point, wherein the tangential vector information comprises a reaction rate vector and a mass transfer driving force vector, extracting the direction of the reaction rate vector, and setting a vector formed by extracting the direction of the reaction rate vector as an evolution direction vector; And simultaneously obtaining a geodesic equation of the ideal manifold surface, setting a vector formed by extracting tangential vectors of the geodesic equation as an ideal direction vector, calculating a jacobian field deviation value between the evolution direction vector and the ideal direction vector, generating a safety track signal if the jacobian field deviation value is smaller than a preset safety deviation threshold value, and generating a divergence track signal if the jacobian field deviation value is larger than or equal to the preset safety deviation threshold value.
5. The DCS automatic control system of a natural gas clean up yard process flow of claim 4, wherein when generating divergent trajectory signals, the predicted state point at a future time is compared to the predicted critical distance threshold with the predicted critical distance threshold: Generating an adjusting instruction if the Riemann distance is smaller than a preset critical distance threshold value, and generating an emergency cut-off instruction if the Riemann distance is larger than or equal to the preset critical distance threshold value; the Riemann distance represents a product value obtained by multiplying the integral length of a path projected from the real-time state tensor to the ideal manifold surface along the geodesic line by a numerical value obtained by carrying out dimensionless normalization processing on the jacobian field divergence coefficient, and the jacobian field divergence coefficient represents the separation speed of the adjacent geodesic line in the evolution process.
6. The DCS automatic control system of a natural gas purification plant process according to claim 1, wherein said optimal path coefficient acquisition and analysis process is as follows: Acquiring geodesic distance information of a real-time state tensor in a time anchor point, wherein the geodesic distance information comprises an energy consumption characteristic value and a quality characteristic value, and comparing the energy consumption characteristic value and the quality characteristic value with a preset energy consumption threshold value and a preset quality threshold value for analysis; And simultaneously calculating the composition ratio of the lean solution circulation flow value and the reboiler steam consumption, generating an inertia compensation factor to dynamically correct the weight of the weighted sum if the composition ratio deviates from a preset balance interval, and setting the corrected weighted sum as a navigation correction coefficient.
7. The DCS automatic control system for a natural gas purification plant process as set forth in claim 6, wherein: The energy consumption characteristic value represents a product value obtained by multiplying a lean solution circulation flow value of a natural gas purification station by a numerical value obtained by carrying out data normalization treatment on the lean solution circulation flow value and a reboiler steam consumption value, wherein the lean solution circulation flow value represents the total amount of amine liquid entering an absorption tower in unit time, and the reboiler steam consumption value represents a steam enthalpy value required for maintaining the temperature of a regeneration tower; The quality characteristic value represents a product value obtained by multiplying the predicted product gas sulfur content of the real-time state tensor and a preset sulfur content standard value by a numerical value after data normalization processing, and the predicted product gas sulfur content represents the concentration of hydrogen sulfide evolving to an outlet time along the current geodesic line.
8. The DCS automatic control system of a natural gas purification plant process of claim 6, wherein said covariate derivative control quantitative match analysis process is as follows: Acquiring control constraint information of a real-time state tensor in a time anchor point, wherein the control constraint information represents a safety operation boundary of equipment, and setting a value obtained by multiplying a navigation correction coefficient by a numerical value corresponding to the safety operation boundary of the equipment as a dynamic set value; Calculating the absolute value of the difference between the dynamic set value and the current set value of the DCS system, generating a holding instruction if the absolute value of the difference is smaller than or equal to a preset adjustment dead zone value, and generating an adjustment instruction if the absolute value of the difference is larger than the preset adjustment dead zone value; The analysis process of the equipment safety operation boundary comprises the steps of obtaining the total times of the flooding factors of the absorption tower in the time anchor point, obtaining the occurrence frequency of entrainment of the absorption tower, obtaining the corrosion rate monitoring value of the regeneration tower, multiplying the corresponding values of the total times of the flooding factors of the absorption tower, the occurrence frequency of entrainment and the corrosion rate monitoring value, and setting the product value obtained by multiplying the corresponding values of the total times of the flooding factors, the occurrence frequency of entrainment and the corrosion rate monitoring value as the equipment safety operation boundary.

Description

DCS automatic control system of natural gas purification station technical process Technical Field The invention relates to the technical field of natural gas purification and industrial automation control, in particular to a DCS automatic control system for a natural gas purification station process flow. Background In the field of automatic control of the process flow of a natural gas purification station, a distributed control system gradually becomes a core center for guaranteeing production safety and product quality due to high reliability and flexibility; at present, a distributed control system generally supports a feedback regulation strategy based on a classical control theory, if the system is required to deal with complex chemical reactions in a natural gas purification process, a proportional-integral-differential control loop or a linear model-based predictive control method is generally adopted, steady state regulation of the purification process is realized by collecting physical quantities such as temperature, pressure and flow, but when the process flow is controlled in the mode, chemical process variables with multidimensional strong coupling characteristics are often regarded as independent data points in Euclidean space, so that an inherent manifold topological structure of chemical reaction dynamics is ignored, abrupt evolution of working conditions cannot be perceived in advance through curvature change, and an optimal path considering energy consumption and quality cannot be planned based on geodesic distance when a nonlinear system has large hysteresis, and the problems of delayed control response and higher energy consumption exist. Disclosure of Invention In order to solve the technical problems, the invention provides a DCS automatic control system for a natural gas purification station process flow, and specifically, the technical scheme of the invention comprises the following steps: the system comprises a data acquisition center, a manifold mapping unit, a trend prediction unit, a steady-state navigation unit, a path deviation correcting unit and a closed-loop execution unit; the data acquisition center is used for acquiring process monitoring data of the natural gas purification station, sending the process monitoring data to the manifold mapping unit for Riemann manifold space mapping analysis to obtain a real-time state tensor, and judging and processing manifold curvature values of the real-time state tensor to obtain a steady state evolution signal or a sudden change evolution signal; when the abrupt evolution signal is generated, the trend prediction unit is used for carrying out future track deviation evaluation feedback analysis on the collected jacobian field distribution information of the real-time state tensor to obtain a safe track signal or a divergent track signal; When a steady state evolution signal is generated, the steady state navigation unit is used for carrying out optimal path coefficient acquisition analysis on the acquired geodesic distance information of the real-time state tensor, processing the acquired energy consumption characteristic value and quality characteristic value to obtain a navigation correction coefficient, and the path correction unit is used for carrying out covariant derivative control quantitative matching analysis on the acquired navigation correction coefficient, and comparing and analyzing the acquired dynamic set value to obtain a holding instruction or an adjusting instruction. Preferably, the Riemann manifold space mapping analysis process comprises the steps of collecting the current sampling time of a natural gas purification station, setting the current sampling time of the natural gas purification station as a time anchor point, setting each sensor in the natural gas purification station as a data node, acquiring physical monitoring values of each data node in the time anchor point, performing tensor processing on the physical monitoring values of the data node, generating a manifold projection instruction if a high-dimensional tensor of the physical monitoring values of the data node is generated, setting the manifold projection instruction corresponding to the high-dimensional tensor as a real-time state tensor, and generating an abnormal alarm instruction if the physical monitoring values of the data node are not generated. Preferably, a preset curvature threshold value of the real-time state tensor in the time anchor point is obtained, meanwhile, a Legend curvature tensor module length of a manifold area where the real-time state tensor is located in the time anchor point is obtained, the Legend curvature tensor module length is set to be a current curvature value, a value obtained by subtracting the preset curvature threshold value from the current curvature value is set to be a curvature deviation value, the curvature deviation value is judged, if the curvature deviation value is larger than zero, a mutation evoluti