Emission-Aware Optimization of Gas Networks: Input-Convex Neural Network Approach

TL;DR

This paper introduces an input-convex neural network approach to optimize gas network planning under emission constraints, ensuring feasible solutions and outperforming traditional solvers in accuracy and reliability.

Contribution

The paper presents a novel ICNN-based optimization method that accurately approximates gas flow physics and guarantees feasible solutions where standard methods fail.

Findings

ICNN-aided optimization outperforms non-convex and relaxation-based solvers.

Larger optimality gains are observed with stricter emission targets.

The approach provides feasible solutions when non-convex solvers fail.

Abstract

Gas network planning optimization under emission constraints prioritizes gas supply with the least CO$_2$ intensity. As this problem includes complex physical laws of gas flow, standard optimization solvers cannot guarantee convergence to a feasible solution. To address this issue, we develop an input-convex neural network (ICNN) aided optimization routine which incorporates a set of trained ICNNs approximating the gas flow equations with high precision. Numerical tests on the Belgium gas network demonstrate that the ICNN-aided optimization dominates non-convex and relaxation-based solvers, with larger optimality gains pertaining to stricter emission targets. Moreover, whenever the non-convex solver fails, the ICNN-aided optimization provides a feasible solution to network planning.