Giorgia Ghione
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Giorgia Ghione has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Eros Pasero
Politecnico di Torino
5 shared publications - Vincenzo Randazzo
Politecnico di Torino
5 shared publications - Giansalvo Cirrincione
University of Picardie Jules Verne
3 shared publications - Marco Badami
Politecnico di Torino, Torino, Italy
2 shared publications - Marco di Benedetto
Department of Industrial, Electronic and Mechanical Engineering, Roma Tre University, Roma, Italy
2 shared publications - Alessandro Lidozzi
Department of Civil, Computer Science and Aeronautical Technologies Engineering, Roma Tre University, Roma, Italy
1 shared publication
Latest Publications
- Open Access
- 17 July 2026
- Published by Wiley in Advanced Intelligent Systems
- Vol. 8 (10) , e70473
Interpretable Short‐Term Electric Load Forecasting
The reliable and efficient operation of power systems hinges on accurate electrical load forecasts over various horizons and objectives, ranging from long‐term grid adequacy estimation to day‐ahead system scheduling and real‐time dispatching. In this context, deep learning (DL) models outperform conventional statistical approaches at the expense of interpretability, making the research on interpretable algorithms essential. Among these methods, the temporal fusion transformer (TFT) has proven its validity across multiple sectors, while its applicability to short‐term load forecasting (STLF) lacks sufficient corroboration. This study evaluates the TFT within a multivariate STLF task concerning a department building of an Italian university, focusing on interpretability. The performance of the model is evaluated against established benchmarks: The TFT outscores all competitors, yielding a mean absolute percentage error of 7.40% on the test subset, 25% lower than the competitors. Jointly, the interpretability of the model is studied: Variable selection weights identify the most important features, a comprehensive investigation of attention patterns highlights the most critical timesteps for generating forecasts, and the Comaniciu distance helps detect regime shifts and significant events. Overall, this study presents an effective TFT‐based model for day‐ahead STLF in terms of interpretability and forecasting performance.
- 17 July 2026
- Published by Wiley in Advanced Intelligent Systems
- Vol. 8 (10) , e70473
Open AccessInterpretable Short‐Term Electric Load Forecasting
The reliable and efficient operation of power systems hinges on accurate electrical load forecasts over various horizons and objectives, ranging from long‐term grid adequacy estimation to day‐ahead system scheduling and real‐time dispatching. In this context, deep learning (DL) models outperform conventional statistical approaches at the expense of interpretability, making the research on interpretable algorithms essential. Among these methods, the temporal fusion transformer (TFT) has proven its validity across multiple sectors, while its applicability to short‐term load forecasting (STLF) lacks sufficient corroboration. This study evaluates the TFT within a multivariate STLF task concerning a department building of an Italian university, focusing on interpretability. The performance of the model is evaluated against established benchmarks: The TFT outscores all competitors, yielding a mean absolute percentage error of 7.40% on the test subset, 25% lower than the competitors. Jointly, the interpretability of the model is studied: Variable selection weights identify the most important features, a comprehensive investigation of attention patterns highlights the most critical timesteps for generating forecasts, and the Comaniciu distance helps detect regime shifts and significant events. Overall, this study presents an effective TFT‐based model for day‐ahead STLF in terms of interpretability and forecasting performance.
- Open Access
- 18 April 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (8) , 3955
Enhanced Neural Real-Time Digital Twin for Electrical Drives
This paper presents a real-time digital twin (DT) of the power conversion system used in offshore wind applications. The proposed DT is exploited to identify key electrical parameters of both the permanent magnet synchronous generator (PMSG) and the three-phase boost rectifier and has been developed with a Condition Monitoring (CM)-oriented approach. A Gated Recurrent Unit (GRU) neural network is adopted as a real-time digital model (RTDM) to estimate online the PMSG phase resistance and synchronous inductance, as well as the DC-link capacitance at the rectifier output. The network is trained in MATLAB using data generated by a Typhoon HIL 606 emulator, covering both balanced and unbalanced operating conditions and a wide range of parameter variations. The trained GRU is then deployed on the control board and implemented in LabVIEW Real-Time for embedded execution. Experimental tests on a PMSG-based generating unit confirm the effectiveness of the proposed RTDM, achieving low root-mean-square and mean percentage errors in parameter estimation. The results demonstrate that the enhanced neural real-time DT is a promising tool for condition monitoring and predictive maintenance of power conversion systems in offshore wind applications.
- 18 April 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (8) , 3955
Open AccessEnhanced Neural Real-Time Digital Twin for Electrical Drives
This paper presents a real-time digital twin (DT) of the power conversion system used in offshore wind applications. The proposed DT is exploited to identify key electrical parameters of both the permanent magnet synchronous generator (PMSG) and the three-phase boost rectifier and has been developed with a Condition Monitoring (CM)-oriented approach. A Gated Recurrent Unit (GRU) neural network is adopted as a real-time digital model (RTDM) to estimate online the PMSG phase resistance and synchronous inductance, as well as the DC-link capacitance at the rectifier output. The network is trained in MATLAB using data generated by a Typhoon HIL 606 emulator, covering both balanced and unbalanced operating conditions and a wide range of parameter variations. The trained GRU is then deployed on the control board and implemented in LabVIEW Real-Time for embedded execution. Experimental tests on a PMSG-based generating unit confirm the effectiveness of the proposed RTDM, achieving low root-mean-square and mean percentage errors in parameter estimation. The results demonstrate that the enhanced neural real-time DT is a promising tool for condition monitoring and predictive maintenance of power conversion systems in offshore wind applications.
- 1 April 2026
- Published by Springer Nature
- p. 299-311
Passive Parameters Identification of a Three-Phase AC–DC Converter via a GRU Network and Derivative Approximations
- 1 April 2026
- Published by Springer Nature
- p. 299-311
Passive Parameters Identification of a Three-Phase AC–DC Converter via a GRU Network and Derivative Approximations
- Open Access
- 17 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 128562-128581
Optimal Cogeneration Scheduling: A Comparison of Genetic and POMDP-Based Deep Reinforcement Learning Approaches
Large processing facilities require multiple types of energy, such as electrical and thermal (hot water or steam). Cogeneration, or Combined Heat and Power (CHP), can provide significant economic and energy savings. However, scheduling its operation in real-time is challenging. This work compares deep reinforcement learning (DRL) and genetic algorithm (GA) approaches to control a real CHP in a processing facility. Traditionally, the CHP economic dispatch problem is modelled as a Markov Decision Process (MDP) with the assumption of complete observability. Due to the uncertainty of future electric and thermal demands, this assumption is unrealistic in real-world scenarios. Thus, this work proposes using a partially observable MDP (POMDP) for hourly CHP dispatch scheduling to address this partial observability. The selected DRL algorithms are Deep Q Network (DQN), Deep Deterministic Policy Gradient (DDPG), and Soft Actor-Critic (SAC), along with six GA variants. Performance was evaluated using multiple economic metrics, including Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA), an environmental analysis, and a sensitivity analysis under variable electric pricing. This work shows that POMDP effectively models the hourly dispatch scheduling problem of CHPs. The insights gained from this analysis offer multiple potential avenues for future research, including the development more advanced DRL algorithms for CHP economic dispatch and the evaluation of their resilience when inaccurate measurements and anomalous conditions occur.
- 17 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 128562-128581
Open AccessOptimal Cogeneration Scheduling: A Comparison of Genetic and POMDP-Based Deep Reinforcement Learning Approaches
Large processing facilities require multiple types of energy, such as electrical and thermal (hot water or steam). Cogeneration, or Combined Heat and Power (CHP), can provide significant economic and energy savings. However, scheduling its operation in real-time is challenging. This work compares deep reinforcement learning (DRL) and genetic algorithm (GA) approaches to control a real CHP in a processing facility. Traditionally, the CHP economic dispatch problem is modelled as a Markov Decision Process (MDP) with the assumption of complete observability. Due to the uncertainty of future electric and thermal demands, this assumption is unrealistic in real-world scenarios. Thus, this work proposes using a partially observable MDP (POMDP) for hourly CHP dispatch scheduling to address this partial observability. The selected DRL algorithms are Deep Q Network (DQN), Deep Deterministic Policy Gradient (DDPG), and Soft Actor-Critic (SAC), along with six GA variants. Performance was evaluated using multiple economic metrics, including Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA), an environmental analysis, and a sensitivity analysis under variable electric pricing. This work shows that POMDP effectively models the hourly dispatch scheduling problem of CHPs. The insights gained from this analysis offer multiple potential avenues for future research, including the development more advanced DRL algorithms for CHP economic dispatch and the evaluation of their resilience when inaccurate measurements and anomalous conditions occur.
- 24 May 2025
- Published by Springer Nature
- p. 45-54
Comparison of LSTM-Based Models for the 6-Hour-Ahead and Day-Ahead Forecasting of Photovoltaic Power Generation
- 24 May 2025
- Published by Springer Nature
- p. 45-54
Comparison of LSTM-Based Models for the 6-Hour-Ahead and Day-Ahead Forecasting of Photovoltaic Power Generation
- Open Access
The geometry of BERT
Transformer neural networks, particularly Bidirectional Encoder Representations from Transformers (BERT), have shown remarkable performance across various tasks such as classification, text summarization, and question answering. However, their internal mechanisms remain mathematically obscure, highlighting the need for greater explainability and interpretability. In this direction, this paper investigates the internal mechanisms of BERT proposing a novel perspective on the attention mechanism of BERT from a theoretical perspective. The analysis encompasses both local and global network behavior. At the local level, the concept of directionality of subspace selection as well as a comprehensive study of the patterns emerging from the self-attention matrix are presented. Additionally, this work explores the semantic content of the information stream through data distribution analysis and global statistical measures including the novel concept of cone index. A case study on the classification of SARS-CoV-2 variants using RNA which resulted in a very high accuracy has been selected in order to observe these concepts in an application. The insights gained from this analysis contribute to a deeper understanding of BERT's classification process, offering potential avenues for future architectural improvements in Transformer models and further analysis in the training process.
Open AccessThe geometry of BERT
Transformer neural networks, particularly Bidirectional Encoder Representations from Transformers (BERT), have shown remarkable performance across various tasks such as classification, text summarization, and question answering. However, their internal mechanisms remain mathematically obscure, highlighting the need for greater explainability and interpretability. In this direction, this paper investigates the internal mechanisms of BERT proposing a novel perspective on the attention mechanism of BERT from a theoretical perspective. The analysis encompasses both local and global network behavior. At the local level, the concept of directionality of subspace selection as well as a comprehensive study of the patterns emerging from the self-attention matrix are presented. Additionally, this work explores the semantic content of the information stream through data distribution analysis and global statistical measures including the novel concept of cone index. A case study on the classification of SARS-CoV-2 variants using RNA which resulted in a very high accuracy has been selected in order to observe these concepts in an application. The insights gained from this analysis contribute to a deeper understanding of BERT's classification process, offering potential avenues for future architectural improvements in Transformer models and further analysis in the training process.