Impact of COVID-19 Pandemic on Energy Demand in Estonia

The COVID-19 pandemic triggered a question of how to measure and evaluate adequacy of the applied restrictions. Available studies propose various methods mainly grouped to statistical and machine learning techniques. The current paper joins this line of research by introducing a simple-yet-accurate linear regression model which eliminates effects of weekly cycle, available daylight, temperature, and wind from the electricity consumption data. The model is validated using real data and enables the qualitative analysis of economical impact. © 2022 IEEE.

Short-Term Load Forecasting Model Considering the Impact of COVID-19 Lockdown in Bangladesh

The impact of COVID-19 lockdown on short-term load forecasting in Bangladesh has been investigated in this paper. Machine learning models have been proved to be the most efficient regarding such prediction. Models like Artificial Neural Network (ANN), Long Short-Term Memory (LSTM) and Random Forest (RF) have been used in this study to build robust models taking the COVID-19 lockdown situation into account. Data sets for the models were formulated by taking daily generation reports, weather indicators and holidays. This study aims to compare different machine learning models to find out the best model for load forecasting keeping into account the impact of COVID-19 lockdown. The results of these methods have been compared based on accuracy metrics. It has been observed that LSTM shows the least error among the compared models. © 2021 IEEE.

Impact of COVID-19 on Ontario’s electricity market: Load, generation, emissions

The COVID-19 outbreak not only threatened global health, it has also –affected the energy markets around the world. This paper studies the impact of the pandemic on Ontario’s electricity market assessing the demand and supply balance over three distinct periods: pre-pandemic, start of the pandemic and during the period 2020–2021. The paper also evaluates the contribution of work-from-home and other mandates in reducing GHG emission. Furthermore, the impact of such rare events is studied on load forecasting. Our analysis shows that although demand dropped by 12% during the beginning of pandemic, it started rising to levels higher than the previous years. Consequently, due to the changes in the daily load profile, primarily due to the changes in consumers’ behavior, the emissions declined significantly during the lockdown and increased afterwards. Finally, this paper provides a short-term Feed Forward Neural Network (FFNN) model to predict future demand. The model performance was evaluated during the three distinct periods and showed high accuracy even in the initial stages of the pandemic: MAPE of 3.21% pre-pandemic, 13.86% beginning of pandemic and 4.23% during pandemic.

Electricity Load Forecasting under COVID-19

The outbreak of novel coronavirus disease in 2020 has profoundly impacted all aspects of lives and posed a unique challenge in energy load forecasting. With the increase of the COVID-19 cases, governments worldwide impose strict social distancing and limit the mobility of the population, which causes a shift in load consumption magnitude and pattern. In this paper, we first identify the most influential COVID-19 features for load reduction. Then, we propose a new load forecasting model that includes the new features. The case study on the New York City data set demonstrates that our new forecasting model can efficiently provide new load prediction in the pandemic period. © 2021 IEEE.

Changes in hourly electricity consumption under COVID mandates: A glance to future hourly residential power consumption pattern with remote work in Arizona

The transition to remote work brings uncertainty to the future power consumption pattern. The COVID mandates in 2020 have accelerated the transition to remote work, generating major uncertainty regarding how residential power consumption patterns will shift. Understanding these shifts is vital for regional operators who will need to implement long-term planning strategies if companies continue to adopt remote work practices. Additionally, if new COVID variants prompt extended stay-at-home mandates, the resulting behavior shifts will decide the optimal combination of power generation in a region. Our study examines changes in hourly residential power consumption patterns resulting from COVID mandates in Arizona. We estimate how the COVID mandates and subsequent remote work practices could change the power consumption patterns using individual-consumer-level hourly power consumption data for 6,309 consumers and a machine learning framework. We also simulate how the hourly power consumption pattern will change with increasing penetration of remote work under winter and summer temperature settings. We then use our simulations to test the policy effectiveness of changing time-of-use (TOU) rates. Our results show that COVID mandates likely increase the power consumption in the afternoon by 13%, and can change the power consumption pattern in winter from a two-peaked shape to a one-peaked shape. Furthermore, we show that the residents' income, race, and house size are significantly correlated with the changes in power consumption, and the correlation is not linear. We find that, by increasing the peak hour prices and decreasing the off-peak hour prices by 10% of the TOU pricing, the peak electricity demand could be reduced by 10%. Our results show under the new remote work era: (1) the need for modifying previous energy generation combination planning due to changing peak demand hours;(2) equity concerns regarding TOU pricing and the inability of vulnerable groups to shift electricity consumption;(3) the ability of governments and utilities to lower the maximum load of power consumption by modifying the TOU rates.

Improving load forecast in energy markets during COVID-19

The abrupt outbreak of the COVID-19 pandemic was the most significant event in 2020, which had profound and lasting impacts across the world. Studies on energy markets observed a decline in energy demand and changes in energy consumption behaviors during COVID-19. However, as an essential part of system operation, how the load forecasting performs amid COVID-19 is not well understood. This paper aims to bridge the research gap by systematically evaluating models and features that can be used to improve the load forecasting performance amid COVID-19. Using real-world data from the New York Independent System Operator, our analysis employs three deep learning models and adopts both novel COVID-related features as well as classical weather-related features. We also propose simulating the stay-at-home situation with pre-stay-at-home weekend data and demonstrate its effectiveness in improving load forecasting accuracy during COVID-19. © 2021 ACM.