Correction to: Implementation of RTO in a large hydrogen network considering uncertainty

Anibal Galán, César de Prada, Gloria Gutiérrez, Daniel Sarabia, Ignacio E. Grossmann, Rafael GonzálezOptimization 2019 · DOI

Abstract

This article discusses the implementation of a robust Real-Time Optimization (RTO) framework for managing large-scale industrial hydrogen networks under uncertainty. Originally published in Optimization and Engineering, this work utilizes two-stage stochastic programming and risk measures like Conditional Value at Risk (CVaR) to achieve a measurable economic advantage over traditional deterministic methods. A formal correction to the original paper clarifies unit notation (k€ for kilo-Euro), ensuring technical precision in reporting the significant cost savings identified.

Introduction

In large industrial complexes such as refineries, hydrogen serves as a critical utility. However, its production and distribution are subject to significant uncertainty, including fluctuating demands from hydrotreating units, variable purity levels, and shifting energy costs. Traditional RTO often relies on deterministic models that assume future variables are known with certainty. This approach can lead to sub-optimal performance or operational failures when real-world conditions deviate from forecasts. This research addresses these challenges by integrating uncertainty directly into the mathematical optimization model.

Methodology

The authors employ advanced mathematical programming techniques to ensure system efficiency and resilience:

  • Two-Stage Stochastic (TSS) Programming: Decisions are split into "first-stage" (immediate) and "second-stage" (recourse) actions taken after uncertainty is revealed.
  • Risk Management (VaR and CVaR): Value at Risk (VaR) and Conditional Value at Risk (CVaR) are incorporated to hedge against "tail-end" probability distributions, prioritizing reliability alongside cost-savings.
  • Performance Metrics: The study utilizes the Expected Value of Perfect Information (EVPI) and the Value of Stochastic Solution (VSS) to quantify the benefits of the stochastic approach.

Key Results and Discussion

The analysis of the hydrogen network reveals substantial economic and operational benefits from adopting a stochastic framework.

Economic Performance

The stochastic model, represented by the Recourse Problem (RP), yielded a value of 735.936 k€/h. A critical finding is the Value of Stochastic Solution (VSS), which measures the gain of using a stochastic model over a deterministic one. The results show a VSS of 10.923 k€/h, representing a 1.51% improvement. In high-volume industrial settings, saving nearly 11,000 Euros per hour through improved mathematical modeling offers a massive return on investment.

Economic performance metrics including the Value of Stochastic Solution (VSS).

Risk Sensitivity and Computational Viability

The model's performance remained stable across different confidence levels. At both 95% and 99% confidence, the optimized cost (CVaR) was 735.88 k€/h, indicating a robust operating point. Furthermore, the "H4" hydrogen plant's capacity utilization increased from 84.24% in a risk-neutral scenario to 86.10% under higher confidence requirements, demonstrating how the model buffers production against volatility.

Crucially, the computational solve time for the 95% confidence risk constraints was approximately 71.46 seconds. This efficiency confirms that stochastic RTO is viable for real-time industrial applications where quick decision-making is essential.

Risk sensitivity analysis and computational solve times across different confidence levels.

Conclusion

The research by Galan et al. demonstrates that integrating uncertainty into RTO frameworks for hydrogen networks provides significant economic value and operational stability. By quantifying the "uncertainty tax" and proving computational feasibility, the study paves the way for more resilient industrial utility management. The recent correction further ensures that these findings are communicated with the necessary technical precision regarding unit notation.

Summary updated

Abstract

Correction to: Optimization and Engineering https://doi.org/10.1007/s11081-019-09444-3
BibTeX
@inproceedings{galan2019correction,
  author = {Anibal Galán and César de Prada and Gloria Gutiérrez and Daniel Sarabia and Ignacio E. Grossmann and Rafael González},
  title = {{Correction to: Implementation of RTO in a large hydrogen network considering uncertainty}},
  booktitle = {Optimization},
  year = {2019},
  doi = {10.1007/s11081-019-09446-1},
}
Figure