Loading...

Development of An Agent-based Framework For Optimization of Water and Energy Network Structure

Babaei, Farshid | 2023

61 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 56677 (06)
  4. University: Sharif University of Technology
  5. Department: Chemical and Petroleum Engineering
  6. Advisor(s): Bozrogmehry Boozarjomehry, Ramin; Pishvaie, Mahmoud Reza
  7. Abstract:
  8. Information and Communication Technology and Artificial Intelligence (AI) can support the transition from traditional practices toward smart supply chain coordination schemes. Such a direction allows for integrating organizational levels of natural gas transportation enterprises, comprising Location, Inventory, and Routing Problem (LIRP) decisions. Specific works have investigated the LIRP of natural gas-to-product and energy networks. However, due to the absence of a systematic ICT-assisted mindset, no study has been oriented from the perspective of the dynamic inventory management and geographic data-driven features of land resources. To handle these challenges, this study partially conveys a novel framework for co-expansion planning of an integrated gas network and methanol plant with natural gas feedstock while enforcing power network constraints. The framework is novel due the fact that it integrates the non-linear complexities and challenges induced by the chemical plant and power system constraints. The work also considers spatiotemporal and quantitative distributions of water sources, water withdrawal, and consumption rates of involved power and chemical technologies. Furthermore, the framework incorporates land resource characteristics, including geological surface features and elevation changes, and devises a strategy for practical cost estimation of different gas transmission expansion stages. Results indicate that the chemical plant and power system integrations, incorporated water, land resource characteristics, and categorized construction costs considerably influence expansion alternatives of gas and chemical systems in terms of power and gas supply reliabilities and economic criteria. Moreover, this work introduces a vertically-coordinated heuristic optimization framework, in the form of an autonomous multi-agent system, for the LIRP of gas transmission infrastructures as major carrier systems to supply feedstock and energy demands. Various extendable behavior classes are programmed into the framework to support automated agent specializations regarding procedural tasks of planning, transient inventory coordination, routing, and spatial data management in the specific LIRP settings. Additionally, the study develops an ontology-driven knowledge system, formally conceptualizing the intended application-level knowledge according to a general domain semantic model. Leveraging the knowledge system, the agents determine the necessary steps and suitable ready-made tools to solve the problem collaboratively. Several industrial applications showcase the capabilities of the proposed framework. The results manifest that the devised ontology automates model construction and the entire solution process while supporting agent interoperation and scaling the platform applications to emergent domains. Additionally, transient demand fluctuations, dynamic inventory levels, and topological considerations heavily influence long-term planning. Compared to existing practices, the ontology-driven multi-agent framework expounds a more realistic view of the trade-off between strategic investments and dynamic operational welfare in the gas system. Such an automated and practical methodology, in turn, realizes AI-oriented cross-functional integrations in the enhanced product and energy sector
  9. Keywords:
  10. Methanol Production ; Ontology-Based Information Management ; Integrated Gas and Power Networks ; Water And Energy Networks Multi-Scale Planning ; Agent-Based Optimization ; Water and Land Resources

 Digital Object List

 Bookmark

No TOC