ResiWater deliverable report D4.1: Stable and reliable pressure dependent modeling
2017
This report describes different methods for modelling of pressurized water supply networks with consideration of pressure dependent demands (PDM: Pressure Driven Modelling). The capability of a hydraulic network simulation software to correctly deal with pressure dependent demands is essential for the development in ResiWater. In case of system failures that might be caused by pipe bursts, natural disasters or malevolent attacks by humans, the pressure is reduced due to decreased capacity of the system. In such case, the common DDM solvers (DDM: Demand Driven Modelling) deliver unrealistic results with negative pressures while maintaining the defined outflows at the side of the customers. PDM modelling fixes this problem by automatically reducing the demand when the pressure drops under the minimum service pressure. The deliverable starts with an overview of existing approaches to PDM. Problems with convergence are well documented in literature. Two different methods for solution are proposed in the main part. The first includes a step size control based on Goldstein criterion. It has been published already in a journal that's why it is not presented in detail in the deliverable. The second approach contains the so-called Content Model formulation of the PDM problem that delivers also statements about existence and uniqueness of the solution. For numerical calculation, an active set method is proposed. The methods have been implemented in Matlab and successfully tested for several test networks with up to 20.000 nodes under strong PDM conditions (almost all nodes are in PDM mode). Small examples exist also in Excel. The approach will be also implemented in Porteau and SIR 3S as part of the reliable hydraulic system solver (task 4.4). Tests were made in SIR 3S for the BWB pilot networks. The deliverable ends with Conclusion.
Keywords:
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
0
References
0
Citations
NaN
KQI