Using the lstm network to forecthe demand for hard coal

Abstract

Securing the certainty of supplies of the necessary minimum energy in each country is a basic condition for the energy security of the state and its citizens. The concept of energy security combines several aspects at the same time, as it can be considered in terms of the availability of own energy resources, it concerns technical aspects related to technical infrastructure, as well as political aspects related to the management and diversification of energy supplies. Another aspect of the issue of energy security is the environmental perspective, which is now becoming a priority in the light of the adopted objectives of the European Union’s energy policy. The restrictive requirements for reducing greenhouse gas emissions and increasing the required level of renewable energy sources in the energy balance of the Member States is becoming a challenge for economies that use fossil fuels to a large extent in the raw material structure, including Poland. Poland is the largest producer of hard coal in the European Union and hard coal is a strategic raw material as it satisfies about 50% of the country’s energy demand. In this context, the main goal of the article was to determine the future sale of hard coal by 2030 in relation to environmental regulations introduced in the energy sector. For this purpose, a mathematical model with a 95% confidence interval was developed using artificial LSTM neural networks, which belong to deep learning machine learning techniques, which reflects the key relationships between hard coal mining and the assumptions adopted in the National Energy and Climate Plan for the years 2021–2030 (NECP).