Data processing: generic control systems or specific application – Specific application – apparatus or process – Specific application of pressure responsive control system
Reexamination Certificate
2002-01-30
2004-02-24
Gandhi, Jayprakash N. (Department: 2125)
Data processing: generic control systems or specific application
Specific application, apparatus or process
Specific application of pressure responsive control system
C700S090000, C700S266000, C137S119030
Reexamination Certificate
active
06697713
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a method of controlling a gas distribution system having one or more pipeline sections that are supplied with gaseous oxygen and/or nitrogen by one or more air separation plants. More particularly, the present invention relates to such a method in which the air separation plants are controlled to supply the gaseous oxygen and/or nitrogen by a model predictive controller responsive to pipeline pressure and customer demand.
BACKGROUND OF THE INVENTION
Gaseous distribution systems include air separation plants that cryogenically rectify the air by well-known techniques to produce gaseous oxygen and gaseous nitrogen products. These products are supplied to customers through a pipeline that has one or more sections. Such gaseous distribution systems are controlled by the supplier manually to ensure that each section of a pipeline is above a certain minimum pressure set by customer contractual requirements. The pressure within a pipeline section relates to the flowrate of gaseous product within a pipeline section. Each customer that draws product from a pipeline section contracts with the supplier to draw a specific maximum flow rate of product from the pipeline. These contractual customer requirements are used to set the pipeline pressure.
The air separation plants may be located at a single geographical location or multiple locations. There can be multiple customers that draw product from each pipeline section. The individual plants are controlled by supervisory control systems in which production request changes are set to in turn control plant production. The pipeline is controlled by a pipeline operator, an individual who monitors a few strategic pipeline pressures and customer demands. If pressures begin to approach the high or low limits for a particular pipeline section, the operator calls the individual plants and verbally requests a change in production. The operator at the plant then enters the new production request level into the plant supervisory control system, at which point production is ramped to the requested level.
Where multiple air separation plants are to be controlled, it is difficult to make the best economic decision regarding which plant will supply the next increment of gas. Often, the decision is to exercise appropriate control over the plant closest to the point where the pressure change is needed. This problem is further complicated by the fact that air separation plants consume electricity and there exists electrical supply contracts that constrain the maximum electrical power a plant can draw and therefore, the amount of product a plant can produce. Moreover, plants are of varying sizes and therefore even without electrical supply contracts, they have various capabilities of supplying product.
Typically, the pipeline operator controls a pressure within a pipeline section to be substantially higher than the minimum pressure. This of course increases the overall expense in operating the pipeline system in that the air separation plants supplying the pipeline at a pressure that is higher than is necessary.
As a backup, if pipeline pressure falls below the minimum pressure, stored liquid oxygen is vaporized and then introduced into the pipeline. However, maintaining stored liquid oxygen reserves is an expensive proposition in that liquid oxygen is a more expensive product to produce than gaseous oxygen.
Simulation Studies have been conducted in which a single gas pipelines is controlled automatically, by model predictive control techniques. Examples of these studies are shown in Zhu et al., “Dynamic Modeling and Model Predictive Control of Gas Pipeline Networks” presented at the American Institute of Chemical Engineering Annual Meeting, Dallas, Texas, November 1999 by Zhu et al. and Zhu et al., and “Journal of Process Control”, Zhu et al., “Dynamic Modeling and Linear Model Predictive Control of Gas Pipeline Networks”, April, 2000. In both of these references, pre-stored models of the response of the pipeline and production of an air separation plant were used to compute an open loop response and a closed loop control action to maintain the pressure at a target value.
Automatic control has been applied to the control of multiple product compressor stations located within pipelines, such as appeared in Seiver et al., “A Pyramid Approach to Advanced Control”, Control Magazine, July, 2000. Model predictive control has also been applied to prevent overflowing of sewer systems such as disclosed in Gelormino et al., “Model-predictive Control of a Combined Sewer System”, International Journal of Control, vol. 59, 1994.
As will be discussed, the present invention applies model predictive control by way of a technique that allows gas distribution systems to be automatically controlled in an economically optimal fashion.
SUMMARY OF THE INVENTION
The present invention provides a method of controlling a gas distribution system having at least one pipeline section and at least one air separation plant. As used herein and in the claims, the term, “pipeline section” means a run of pipeline in which a pressure change along its length is solely due to frictional, fluidic losses. The at least one air separation plant is controlled by a supervisory control system responsive to production request changes to direct a production rate of at least one gaseous product to be consumed by at least one customer connected to the at least one pipeline section.
In accordance with the method, pressure is continually measured within the least one pipeline section. The pressure is dependent upon an actual flow rate of the at least one gaseous product within the at least one pipeline section. The pressure is controlled to be within a range that will ensure that the at least one customer will be able to obtain a required flow rate of the at least one gaseous product from the at least one pipeline section. The control involves measuring and storing pressure values of the pressure and production request values of production requests associated with each of the pressure values in a rolling history and inputting the pressure values as controlled variables and the production request values as manipulated variables into a model predictive controller. Within the model predictive controller, an open loop response of the pressure within the at least one pipeline section is calculated over a prediction horizon along with a set of production request changes of the at least one air separation plant, as a set of manipulated variables, required to at least in part restore the pressure to a target value within the pressure range from the open loop response.
The calculation of the target value, the open loop response, and the set of production request changes are based on at least one empirically determined step response model of pressure within the at least one pipeline section in response to a unit production change in the request value of the at least one air separation plant. The set of future production request changes are optimized to simultaneously minimize deviations of each of the production request changes and the pressure from the target value over the prediction horizon. The production request changes are inputted into the supervisory control system of the individual air separation plants to control the at least one air separation plant to produce the at least one product in accordance with the set of production request changes.
As can be seen from the above description of the present invention, unlike the prior art, model predictive control is more practically applied by the use of empirically based models. Moreover, there is no direct control of the air separation plants. Instead, control of the air separation plants remains in the supervisory control system. The model predictive control of the present invention generates production request changes based on the empirical model to in turn allow for the air separation plants to be integrated into the control scheme of the present invention with little modification if any.
Preferably, a fu
Adebekun Derin
Lennox David F.
Megan Lawrence
Scharf Paul F.
Zhu Mark
Gandhi Jayprakash N.
Praxair Technology Inc.
Rosenblum David M.
LandOfFree
Control for pipeline gas distribution system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Control for pipeline gas distribution system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Control for pipeline gas distribution system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3288540