Rubén Mulero | Research Scientist & PhD Candidate

A Fresh Start: Relaunching My Personal Research Website

Thu, 29 Oct 2026 00:00:00 +0000

Welcome to the new version of my personal academic website!

When I launched this site back in 2020, it was a modest space where I shared initial thoughts as I began my journey into applied research. Over the past few years, my work at Tecnalia Research & Innovation and my doctoral research at the University of the Basque Country (UPV/EHU)—co-supervised by Dr. Roberto Garay Martinez (DeustoTech, University of Deusto) and Dr. Iñigo Mendialdua (UPV/EHU)—has matured significantly.

As my research expanded into advanced control paradigms—ranging from Deep Reinforcement Learning (DRL) and Model Predictive Control (MPC) to empirical testing at Tecnalia’s KUBIK experimental building—I felt it was time for the website to evolve as well.

What’s New in this Version?

Here are the key improvements integrated into this release:

Entirely in English: To foster international collaboration, share preprints, and engage with the global building energy and machine learning research communities.
Comprehensive Publications Catalog: A dedicated, searchable, and filterable Publications page featuring all 17 peer-reviewed journal articles, conference papers, and book chapters, complete with direct DOI links and one-click BibTeX citations.
Structured Research Lines: A clear overview of my core Research Lines, detailing methodologies, simulation workbenches, and physical testbed validations.
Full-Featured RSS 2.0 Feed: You can now subscribe directly via RSS (/feed.xml) in your favourite feed reader (NetNewsWire, Feedly, Thunderbird, etc.) without relying on social media algorithms to catch new posts and paper announcements.
Verified Academic Profiles: Direct connections to my profiles on Google Scholar, ORCID, and the Tecnalia Science Portal (Pure).
Modern, Responsive Design: Built from the ground up with clean typography (Inter + Merriweather), a light and dark mode toggle, and a responsive mobile interface.

Looking Ahead

Moving forward, I will use this blog to share:

Breakdowns and plain-language summaries of newly published papers.
Practical insights on training reinforcement learning agents on thermodynamic building models.
Open-source tools, simulation workbenches, and dataset releases.
Takeaways and presentations from scientific conferences.

Feel free to browse around, check out the Publications, subscribe via RSS, or reach out via email or LinkedIn if you are interested in collaborating!

A Training Workbench Based on Transient Building Models for Creating Intelligent Energy Operators

Sun, 15 Jun 2025 00:00:00 +0000

We are excited to share insights from our latest research published in Data-Centric Engineering (Cambridge University Press): “A training workbench based on transient building models for creating intelligent energy operators” (DOI: 10.1017/dce.2025.10016).

Why Do We Need Intelligent Energy Operators?

Modern buildings are intricate thermodynamic systems with significant thermal inertia, non-linear thermal storage dynamics, and volatile external driving forces (solar radiation, outdoor temperature, internal gains, and fluctuating electricity tariffs).

Traditional Rule-Based Control (RBC) or classic PID systems are rigid: they cannot anticipate future events or adjust to dynamic tariff structures without burdensome manual re-tuning.

While Deep Reinforcement Learning (DRL) offers an appealing paradigm—learning optimal control policies through interaction—training DRL agents directly on physical buildings is fundamentally infeasible due to safety risks, equipment wear, and sample inefficiency.

The Training Workbench Architecture

To bridge this critical gap, we formulated a modular training workbench that couples:

Transient High-Fidelity Building Models: Accurately replicating multi-zone heat transfer, thermal storage, and HVAC plant equipment.
Standardized DRL Environments: An OpenAI Gym / Gymnasium compatible API allowing rapid integration with leading algorithms (PPO, SAC, DQN).
Multi-Objective Generalized Reward Functions: Formulations that jointly penalize thermal discomfort (predicted mean vote / temperature violation hours) and electrical cost.

# Conceptual reward function structure
def calculate_operator_reward(indoor_temp, comfort_band, energy_cost, alpha=0.6):
    comfort_penalty = max(0, comfort_band.min - indoor_temp) + max(0, indoor_temp - comfort_band.max)
    cost_penalty = energy_cost
    reward = - (alpha * comfort_penalty + (1.0 - alpha) * cost_penalty)
    return reward

Empirical Validation & Next Steps

Using our testbench, autonomous operators learned supervisory schedules that achieved significant energy cost reductions while preserving strict thermal comfort limits during peak heating hours.

Our ongoing work focuses on deploying these trained policies into physical test facilities, including Tecnalia’s KUBIK experimental building.

Read the full paper on Cambridge University Press.

Rubén Mulero's Website is Now Online

Fri, 01 May 2020 00:00:00 +0000

After some thought, I decided to establish this personal website to openly share my scientific journey, ongoing research experiments, and engineering projects.

My intention is to disseminate research findings, connect with fellow researchers and practitioners in machine learning and building energy systems, and foster collaborative work.

Feel free to explore my publications, examine my research lines, or subscribe via the RSS Feed.