ABSTRACT

Following the history of steady state optimization applied to OMV's distribution and transmission systems, the development of an optimization method, based on Sequential Linear Programming will be described. An iterative cooperation between a detailed simulation model and a simplified optimization model enables the application to fulfill any desired degree of accuracy and complexity. The optimization goal is the minimization of operating costs. Practical experiences made during development and application as well as general problems of steady state optimization will be pointed out. Finally conclusions and suggestions on how to overcome some of these problems are given.

Biography

Erwin Sekirnjak holds a Ph.D. in Physics from the University of Vienna. After one year of experience in designing optical systems he joined the Austrian Oil and Gas Company, OMV. For 27 years he has been working in various areas of computer applications with an emphasis on operations research methods. His responsibility covered applications like reservoir simulation, refinery optimization, optimization for long-term company investments, gas control systems, planning and optimization models for gas transportation systems.

1. Introduction

OMV's aas transnortation system: OMV operates a national distribution system as well as two international transmission systems (see Fig. 7). Gas is moved from external sources, few domestic fields, 5 storage fields, through 800 miles of pipelines, 10 compressor stations and various gas treatment plants to local distribution companies and international delivery stations. The transmission systems represent a small, but important part of the European network as shown in Fig. 8. They connect several entry / exit locations to our neighboring countries: Slovakia (the major entrypoint of Russian gas supply), Italy (TAG), Germany (WAG), Slovenia (SOL), and Hungary (HAG). The Austrian portion of the transmission systems currently consist of 600 miles of high-pressure pipelines, and 4 compressor stations, delivering 630 MMMCF/year, whereas the national delivery amounts to 220 MMMCF/year. How does Optimization support decision making? Searching for good decisions dispatchers and planners are dealing with various entities: the real world: pipelines, compressors, plants, companies, people…. current requirements: demands, technical and economic constraints (contracts, schedules, forecasts) etc. models: data from the scads system, graphic images, mental images (experience), physical laws, mathematical and computer based images etc. Decision processes typically consist of the following phases:

  1. knowing and understanding the current situation

  2. finding alternative strategies which fulfill the current requirements

  3. evaluating and comparing the alternatives (testing)

  4. selecting the most appropriate solution

  5. implementing the selected strategy in the real world, feedback on performance Phases 1 and 2 are the traditional areas supported by simulation.

A well tuned online model permanently promotes good knowledge and understanding of the current situation and predictive models are especially useful for planning of transient operations. But even if the decisions concern steady state conditions it is often wise to apply transient simulation in phases 3 or 4 to ensure technically feasible transition from the current to the new state.

This content is only available via PDF.
You can access this article if you purchase or spend a download.