×
Community Blog Project Showcase | Effect of Weather on Energy Generation and Demand

Project Showcase | Effect of Weather on Energy Generation and Demand

This project is from Siddharth, which was awarded with the Innovation Award in the Global AI Innovation Challenge 2021 - Intelligent Weather Forecast for Better life.

This project is from Siddharth Deshpande, which was awarded with the Innovation Award in the Global AI Innovation Challenge 2021 - Intelligent Weather Forecast for Better life.

Project Introduction

The goal of the project is to analyze the effect of weather on energy generation and demand and hopefully come up with a solution that can better predict renewable energy generation and energy demand using weather forecast

1
2
3

Solution and Project Value

Project Value

  1. The end users of this project would be government energy departments at various level and electrical companies.
  2. This project will help them to understand climatic factors that can improve renewable energy generation.
  3. For example, solar energy generation is positively impacted by temperature but negatively impacted by humidity. So regions with hot and dry climate will help generate maximum solar energy. So while planning to improve country’s solar energy production, it makes sense to use regions with suitable climatic conditions to install solar panels.
  4. Furthermore, if done regionally this project can be used to determine which climatic and time based conditions are driving up electrical demand and price. This will help energy departments to predict situations where high electrical demand will be generated and plan accordingly to avoid power shortages and make people’s life easier.
  5. Many countries are recognizing contribution of climatic conditions in the energy sector and this project can help them pin point the exact climatic factors responsible for particular energy sector

Commercial Value

  1. In a warmer climate, Americans will use more electricity for air conditioning and less natural gas, oil, and wood for heating. If the nation's climate warms by 1.8°F, the demand for energy used for cooling is expected to increase by about 5-20%, while the demand for energy used for heating is expected to decrease by about 3-15%. Net expenditure in annual heating and cooling could increase by 10% ($26 billion in 1990 dollars) with a 4.5°F warming by the end of the century, and by 22% ($57 billion in 1990 dollars) with a warming of 9.0°F.
  2. Heating demand would decrease the most in the northern United States, and cooling demand would increase the most in the southern United States.
  3. Warming is likely to increase summer peak electricity demand in most regions of the United States. For example, based on a 6.3 to 9°F temperature increase, climate change could increase the need for additional electric generating capacity by roughly 10-20% by 2050.
  4. This study would help to increase the renewable energy production, by identifying important climatic factors responsible for renewable energy production which could help to handle the increased energy demand. Furthermore, it can help in better forecasting of electrical demand and electrical price which can save billions of dollars

Solution

1.  Analysis was done using a Kaggle dataset – which contains 4 years of electrical consumption, generation, pricing, and weather data for Spain.

2.  Exploratory data analysis was done on the dataset. Observations from EDA are as follows:

a) Biomass, waste and nuclear electrical generation is not affected by any climatic conditions
b) Solar electrical generation is affected by temperature, humidity, wind speed and by hours
c) Wind electrical generation is affected by wind speed and wind deg
d) Hydro electrical generation is affected by wind speed, presence of cloud and month
e) Electrical demand is correlated with solar, hydro and fossil electrical generation and hour, humidity, wind speed, weekday and temperature
f) Electrical price is correlated with fossil generation, wind generation and and wind speed, wind degree, hour, weekday and month

3.  Solar, wind and hydro energy generation prediction using climate and time parameters only. ExtraTreeRegressor algorithm showed best performance and was used to build the regression model.

4.  Energy demand prediction was done using time and energy parameters (Model 1) and time, energy and climate parameters (Model 2). Model 2 showed slightly higher accuracy than Model 1. It shows that climate parameters do not affect energy demand as significantly as energy parameters. ExtraTreeRegressor algorithm showed best performance and was used to build the regression model.

5.  Energy price prediction was done using time and energy parameters (Model 1) and time, energy and climate parameters (Model 2). Model 2 showed higher accuracy than Model 1. It shows that climate parameters affect energy price significantly. ExtraTreeRegressor algorithm showed best performance and was used to build the regression model.

Technology Highlights

  1. Dataset
  2. Additional features created include business hours, siesta hours, other hours, weekday and weekends.
  3. Exploratory Data Analysis was done using seaborn and matplotlib packages.
  4. Around 40 different regression models were used to predict energy price, energy demand, solar energy generation, wind energy generation and hydro energy generation. For entire list, please refer to solution overview.
  5. To build an application, Python Dash was used. Dash app can be later deployed to ECS

Alibaba Cloud Products Used

About the Developer

I am Interdisciplinary researcher with unique perspectives derived from translational projects combining Materials Engineering, Biochemistry, Healthcare and Data Science. I am looking for opportunities to solve problems in Climate and Healthcare. I have completed my Masters in Material science and PhD in Medicine from National University of Singapore.

0 0 0
Share on

Alibaba Cloud Project Hub

69 posts | 11 followers

You may also like

Comments