A study on carbon emission prediction of multi-energy complementary power system based on multiple linear regression...

Area of Science:

• Energy Systems Engineering
• Environmental Science
• Data Science

Background:

• Multi-energy complementary power systems generate vast operational data, posing challenges for accurate carbon emission prediction.
• Existing methods struggle to efficiently process large-scale, distributed data for reliable forecasting.

Purpose of the Study:

• To develop a high-precision carbon emission prediction method for multi-energy complementary power systems.
• To identify key factors influencing carbon emissions in these systems.
• To enhance computational scalability and prediction efficiency.

Main Methods:

• Analysis of multi-energy complementary power system structure and carbon emission intensity.
• Selection of preliminary carbon emission influencing factors.
• Construction and optimization of a multiple linear regression model using significance tests.
• Integration of the model with the MapReduce parallel framework.

Main Results:

• The optimized multiple linear regression model demonstrated a high goodness-of-fit.
• Prediction errors were minimal, ranging from 0.00516% to 0.00818%, indicating high accuracy.
• The model successfully processed large-scale distributed data efficiently.

Conclusions:

• The developed method accurately predicts carbon emissions in multi-energy complementary power systems.
• The model's findings provide a basis for effective carbon emission reduction policies.
• Annual carbon emissions are projected to increase until 2031 and then decline until 2034.