Charlotte Verhaeghe

Mrs. Charlotte Verhaeghe

orcid

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 June 2020
Verified
Conference Paper
Open access
PDF file
CVCharlotte Verhaeghe
Charlotte Verhaeghe
AAAudenaert Amaryllis
Audenaert Amaryllis
SVStijn Verbeke
Stijn Verbeke
The rising interest in low energy building has led to an inflation in related terminology: (nearly) zero energy buildings, or (n)ZEBs, passive houses, positive energy buildings and districts, off-grid buildings, energy autarkic buildings, etc. Each of these terms involves (sometimes subtle) differences in interpretations, system boundaries, included energy end uses, etc. This paper maps the differences and overlaps in applications of various cases of residential High Energy Performance Buildings (HEPBs), aiming to contribute in the development of a novel taxonomy to evaluate the extent to which a building can be considered energy or carbon neutral. Three dimensions are suggested for specification in novel taxonomy for HEPBs: (i) the spatial dimension (energy use, locally renewable energy production and sometimes energy storage), (ii) the time dimension (during which period is the building and its systems balanced, e.g. yearly or momentary) and (iii) the end-use dimension (these are the end-uses that are included or excluded for the calculation of the total energy needs of the buildings).

Verified
Journal Article
Open access
PDF file
CVCharlotte Verhaeghe
Charlotte Verhaeghe
JMJasmine Meysman
Jasmine Meysman
LTLucas Zenichi Terada
Lucas Zenichi Terada
AVArthur Vissers
Arthur Vissers
Amaryllis AudenaertAmaryllis Audenaert
Amaryllis Audenaert
SVStijn Verbeke
Stijn Verbeke
Dynamic energy tariffs increasingly mirror renewable generation and market conditions, yet most techno-economic assessments of residential retrofit and flexibility measures still rely on simplified, static tariffs, risking-biassed cost estimates. This study systematically maps energy tariff structures across three classification tables, showing that dynamic and weather-responsive tariffs are underrepresented in ex-ante building energy modelling. Building on this gap, four ex-ante tariff-modelling methods are developed and validated, namely a fixed tariff, a time-of-use (ToU) tariff, a weather-dependent real-time-pricing (RTP) mechanism grounded in a residual-load proxy, and a capacity-based network tariff, illustrated for the Flemish context. The RTP mechanism is calibrated against two years of Belgian day-ahead prices (ENTSO-E) and Elia demand, as well as PV- and wind-generation data, reaching a monthly correlation of ρ = 0.78 (hourly ρ = 0.43, excluding 2022) and ρ = 0.71 for the final quantile-mapped tariffs, using only weather data (.epw) and three calibratable parameters. An illustrative building-level application shows that, for an identical unoptimised demand profile, a weather-driven RTP tariff changes the projected annual electricity cost by roughly 33–44%, against only 1–3% for a static ToU tariff, confirming that tariff structure must be an explicit ex-ante modelling choice.

3 September 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
LTLucas Zenichi Terada
Lucas Zenichi Terada
CVCharlotte Verhaeghe
Charlotte Verhaeghe
SVStijn Verbeke
Stijn Verbeke
Ricardo BessaRicardo Bessa
Ricardo Bessa
MRMarcos J. Rider
Marcos J. Rider
Residential controllers coordinating photovoltaic (PV) generation, a stationary battery energy storage system (BESS), and an electric vehicle (EV) often use deployment-time action projection to enforce operating limits, which can conceal how strongly the controller stack depends on corrective intervention. This study compares safety-projected operation with an independently executed unprojected closed-loop diagnostic. A residential energy-management environment with state-dependent storage limits, PV curtailment, EV availability, and five tariff structures is used to compare penalty-based and Lagrangian soft actor-critic (SAC) controllers across three sequence encoders. Device-level BESS and EV constraints are handled through feasibility-aware action mapping, while a separate Lagrangian channel represents EV departure risk. Demonstration-free stabilizers, including potential-based BESS shaping, critic dropout, and randomized critic ensembles, are compared with a behavior-cloning-initialized fine-tuning reference derived from an offline MILP teacher. Testing uses 24 independently initialized full-month rollouts from two held-out datasets, and additive quantities are reported as mean monthly totals. Most demonstration-free Lagrangian SAC variants exhibit low unprojected closed-loop grid-limit violation while attaining mean monthly operating costs comparable to the fine-tuning reference. The nominal EV-risk budget is not satisfied at the selected scale and the multiplier reaches its cap; therefore, formal CMDP-budget feasibility is not claimed. The results show that cost-only evaluation can conceal projection dependence and storage-control differences.