Picked from the software, topic, and level of “Class LSTM Electricity Load Forecasting #1: Hourly Electricity Load Data: Daily, Weekly, and Seasonal Patterns”.
At most one email a week. Unfollow anytime with one click.
Transcript
The full narration of this video. The current sentence is highlighted while the video plays; click any sentence to jump there.
Show transcript
Introduction
Here is the final result, 5 years of hourly electricity load data showing daily patterns, differences between weekdays and weekends, seasonal cycles, and the influence of temperature. These patterns will be studied by the LSTM to forecast the next 24 hours of load. Let's get started. Welcome to the tutorial series on research, short-term electricity load forecasting with LSTM and evolutionary optimization. Power system operators must know the load for each hour tomorrow to schedule generators at the lowest cost and maintain reliability. An error of just 1% can mean billions of rupiah. In this 6-part series, we will build everything. Data, features, baseline ARIMA and regression, LSTM, optimization with genetic algorithms, up to evaluation. In the first episode, we get to know the data. We open the terminal through the terminal menu, new terminal, and then install the library series.
Step 1: Installing the library & dataset
Torque, pandas, scikit-learn, stats models for ARIMA, holidays for the list of public holidays, and matplotlib. The data is hourly electricity load for the country of Belgium from ENSOI via Open Power System Data, combined with hourly temperature data for the city from Open Meteo, 5 years from 2015 to 2019. The file is in the customer's premium project file. Create a file named data_underscore_load_dot_py.
Step 2: Loading & splitting the data by time
The function load reads a CSV file with the time column as the index. The function split divides the data based on time. The first 3 years for training, 2018 for validation, and 2019 for testing. In time series data, the data should not be shuffled, because we cannot use the future to forecast the past. Create a file named exploration_underscore_load_dot_py and run it.
Step 3: Data summary
There are 43,824 hours of data. The average load is around 9,900 MW, with a minimum of around 6,500 and a maximum of 13,600 MW. Training data is 26,000 hours Validate and test each of 8,760 hours. Negative load temperature correlation, minus 0.29. The colder it gets, the higher the load due to heating.
Step 4: Daily, weekly, seasonal & temperature patterns
Let's look at the pattern. The top-left graph shows the seasonal cycle every year. High in winter, low in summer. Top right graph showing hourly load profile. Load spikes sharply at 6 AM, peaks around 10 AM and 6 PM, and Saturday as well as weekends are much lower than workdays. Bottom left graph: monthly average, and the graph in the lower right showing daily load against temperature. The red dots on the weekend form a separate group with lower values.
Done!
Done. We have already loaded the hourly electricity load data and temperature, split the data by time, and recognize daily patterns, weekly, seasonal, as well as the effect of temperature. In the next episode, we will turn these patterns into features and set up the time series window for LSTM. Working on a thesis, thesis, Or are you doing research in the field of load forecasting and time series? Cerdas Inovasi Digital helps implement deep learning, from data load exploration, feature engineering, ARIMA and regression, LSTM with PyTor, up to hyperparameter optimization with genetic algorithms, as well as opening deep learning courses with PyTor. Want to master the entire series, from hourly load data to forecast evaluation? All six episodes, complete with step-by-step PDF guides and source code, are available to premium tutorial subscribers. Subscribe starting at Rp69.000 per month, save 28% with the 12-month package, or one-time payment of Rp999.000 for lifetime access. Want a single class only? Buy the single class with lifetime access. Subscribe now at the premium page cerdasinovasidigital.com Contact our WhatsApp, or visit cerdasinovasidigital.com Thank you for watching, see you in the next episode.
Ask the class material
Type a question or keyword; we find the part of this episode that explains it and jump the video there.
My notes
Notes are saved with the current video time and are only visible to you.
Questions & answers
Stuck on a step? Ask here; the instructor and other learners can answer.
No questions yet.
Your instructor
Cerdas Inovasi Digital Team
Software developers, BIM & 3D engineers, and AI practitioners
53 video classes1,290 episodes
We build software, BIM models, games and AI automations for clients every day, and turn that hands-on work into step-by-step classes: every step is shown on screen and comes with a PDF guide you can keep.
