ABSTRACT

With the European gas market becoming liberalized and more and more interconnected, the transport system operators (TSOs) are committed by the authorities to publish available capacities to network users and other TSOs in a transparent and unified way. According to our analysis this could be achieved using bottleneck models of local networks, i.e. minimal hydraulic models which are just sufficient to decide on the physical feasibility of a given flow pattern within predefined accuracy and predefined ranges of the boundary conditions. The article will mainly focus on a systematic and automatic procedure to derive such bottleneck models from full hydraulic models via symbolic simplification. Symbolic simplification combines computer algebra and numerical analysis to identify and extract dominant network components. The reduced model is again a hydraulic model with well interpretable parameters. The methodology has been applied successfully to a real-world pipeline network.

1 INTRODUCTION

We present results of the research project e_GASGRID, which was partially funded by the European Commission within the Fifth Framework Program and brought together partners from European universities, research institutes, software companies, consultants, and TSOs active in the gas business [1]. The European gas market becomes liberalized and more and more interconnected. This requires new communication platforms among TSOs and between TSOs and network users, who do no longer belong to the same company, now. For instance, if a shipper plans to move gas across pipelines in responsibility of several TSOs he should be supported by a software agent predicting, the amount of transport capacity available along this path. At the moment we observe a discussion between TSOs on the one hand and network users and authorities on the other hand on the question, if this kind of tools may be based on constant cross-border capacities between entry-exit zones [2]. Today, cross-border flows play a minor role and, owing to long term contracts, the flow patterns do not change a lot. Under these conditions constant cross-border capacities may be applicable. However, if shippers start to switch between operators in a liberalized market, or if there are major accidents, the flow patterns will change essentially. Then capacities depend on critical pressures and in- and outflows at other borders rather than on the cross-border pipeline itself. Therefore, we suggest to replace the concept of constant cross-border capacities by the more general concept of physical feasibility of a certain flow pattern. Typically, the requirements on such physical models needed at the interface to the business world differ considerably from those used at the operator level for monitoring or forecasting purposes. For instance, computing long run marginal costs (LRMC) of an entry-exit tariff system or screening emergency response proposals that involve severel TSOs, there are so many uncertainties that going below a certain accuracy makes no sense. On the other hand, the underlying physical models have to be evaluated quickly for a large number of modified parameter settings or linked together to a super-national model. Hence, the models we need have to be as simple as possible.

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