← 論文一覧に戻る

柔軟な地域熱供給ネットワークの最適設計

Optimaal ontwerp van flexibele warmtenetten (原題)

Martin Sollich

Liriasジャーナル2026-11-17#エネルギー転換Origin: EU経営インパクト: コスト削減対象セクター: power
原典: https://lirias.kuleuven.be/handle/20.500.12942/796737

🤖 gxceed AI 要約

日本語

地域熱供給網(DHN)の設計を自動化する物理ベースの非線形最適化フレームワークを開発した学位論文。熱源・消費者サブステーション・配管・蓄熱を統合し、離散設計を連続緩和とペナルティ法で扱うことでスケーラブルな最適化を実現。ベルギー実網を含む200超の消費者規模のケースで、コスト・CO2削減と炭素価格影響の分析可能性を示した。

English

This thesis develops a physics-based continuous nonlinear optimization framework for automated design of district heating networks, jointly optimizing heat producers, consumer substations, pipe topology, and storage. Discrete design choices are relaxed and steered via penalization for scalability. Case studies up to 200+ consumers show cost and CO2 savings, carbon-pricing sensitivity, and superior storage sizing versus simplified approaches.

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

📝 gxceed 編集解説 — Why this matters

日本のGX文脈において

日本では地域熱供給の脱炭素化と熱分野のトランジションが課題であり、炭素価格を織り込んだ熱源・配管設計の定量手法は、自治体・エネルギー事業者の長期投資計画やGX推進の参考になる。

In the global GX context

District heating is central to EU building decarbonization and the EU-27's ~€11.7bn annual investment need; the framework's carbon-pricing sensitivity links network design to transition-finance and carbon-cost planning relevant to CSRD-era capex disclosure.

👥 読者別の含意

🔬研究者:物理制約を保ちつつスケールする連続NLP最適化でDHN設計を扱う手法として、熱システム最適化研究の参照点になる。

🏢実務担当者:熱供給事業者・自治体は、熱源・配管・蓄熱の統合最適設計と炭素価格感応度を投資検討に活用できる。

🏛政策担当者:熱分野脱炭素の投資効率を示す証拠として、地域熱供給への支援策や炭素価格設計の根拠に使える。

📄 Abstract(原文)

Space and water heating accounted for 77.1% of household energy consumption in the EU in 2024, yet the share of renewables in the heating sector reached only 26.7%. Modern district heating networks (DHNs) are widely recognized as a cost-effective pathway to decarbonize the heating sector, connecting a wide range of renewable and low-carbon heat sources - including solar thermal, geothermal, residual, and aquathermal heat - to consumers across districts and cities. Realizing this potential requires both the development of new networks and the modernization of existing ones, with individual DHN projects requiring investments in the range of tens to hundreds of millions of euros. The required annual investment across the EU-27 is estimated at 11.7 billion EUR. Designing a DHN requires interdependent decisions on heat producers, pipe topology and diameters, consumer substations, heat storages, and operational regimes. Historically, this has been addressed through iterative, manual processes based on design guidelines and simulation of predefined scenarios - an approach that is time-consuming, does not guarantee an economic optimum, and does not scale to large or complex networks. Optimization-based approaches have been proposed to address this limitation. However, the majority of existing methods rely on simplified linear formulations that cannot ensure physical feasibility or accurately capture the interdependencies between network physics and costs. Nonlinear approaches are physically more accurate but are typically not scalable when formulated as mixed-integer nonlinear programs solved by combinatorial methods. To address this gap, this thesis develops a physics-based framework for the automated design optimization of modern district heating networks. The framework belongs to the category of continuous nonlinear program (NLP) optimization, achieving scalability by initially relaxing discrete design decisions and progressively steering the solution towards discrete outcomes through penalization methods, inspired by the field of topology optimization. Four main developments are presented, each extending the framework with a new design feature. First, a heat producer optimization is developed for the optimal selection, sizing, and operation of central heat producers, including the producer supply temperature. Applied to a DHN retrofit study with a multi-objective optimization of costs and CO2 emissions, the approach enables analysis of future carbon pricing impacts on network design and operation, while reliably evaluating heat and pressure losses and temperature-dependent producer efficiencies. Second, a consumer substation optimization is developed, integrating booster heat pump (HP), electric heater (EH), and enlarged heat exchanger (HX) substation configurations into the optimization framework, jointly with the producer supply temperature. Applied to DHN retrofit case studies with more than 200 consumers - substantially exceeding the scale of existing substation retrofit approaches - the methodology demonstrates significant potential for temperature and cost reduction in existing DHNs, and represents the first automated, physics-based substation retrofit demonstrated at this scale. Third, the producer, substation, and pipe optimizations are unified into a first-of-akind joint greenfield design optimization framework. Applied to a synthetic DHN with more than 200 consumers, the framework identifies a configuration combining a central HP with decentralized HX-EH substations as the most costeffective design, with cost savings arising from two complementary mechanisms: a lower central producer supply temperature reducing operational costs, and reduced peak pipe heat loads enabling substantially smaller pipe diameters and lower pipe investment costs. Lastly, a heat storage optimization is developed for the optimal integration and sizing of short-term central storage in DHNs. Applied to an existing combined heat and power DHN in Belgium, the methodology demonstrates that physicsbased storage optimization, which accounts explicitly for network heat losses, outperforms simplified single-node approaches and determines a larger, more cost-effective storage size. Across all four developments, each design study is completed in under 11 hours on a laptop, demonstrating the computational efficiency and scalability of the framework. The developed framework integrates central heat producers, heat storages, consumer booster substations, and pipe design into a single, scalable, physics-based optimization, enabling their joint and cost-optimal design for both DHN retrofit and greenfield studies. The embedded nonlinear DHN simulation, validated against measurement data, ensures the feasibility of the proposed designs and enables accurate cost assessments, supporting the planning of energy- and cost-efficient networks that contribute to the decarbonization of the heating sector.

🔗 Provenance — このレコードを発見したソース

🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録(無料・3キーワードまで)。

gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。