Факторы в предиктивных сценарных моделях для энергетики: мета-анализ

Energy models have the goal of supporting decision makers and stakeholders in making well- founded judgments by providing comprehensive forecasts. Yet, due to the rising number of energy simulations by the scientific community and international organizations, it can be increasingly challenging to keep up to date, identify potentially misleading models and transparently compare baseline assumptions. Especially due to rapidly developing technologies, balanced and exhaustive models are hard to find and demanding to recognize. Renewable energy for instance has been celebrated as well as ridiculed for many years. Some praise the wide array of clean energy sources as the saviour of all humanity, whereas others do not acknowledge renewables will have any significant share in the primary energy mix in the foreseeable future. Which position is more compelling and how do the underlying assumptions compare?
Besides speculations, it has become common practice to not just build one model, but rather create many different scenarios with varying baseline assumptions that yield different results. For instance, the International Energy Agency (IEA) offers three scenarios: the Stated Policy scenario – where planned regulations and near-term policies are considered; the Sustainable Development scenario – assuming a hard shift in public opinion, leading to a more sustainable approach towards energy; and Current Policies scenario – assuming no change in behaviour by humanity and the progression of the current path.
Even though this presents a greater variety of possible directions and is set out to increase the scope of included possible outcomes, it makes it almost impossible to compare models with each other across different resources, as granular differences in scenarios make them unparalleled. As a result, incomparable energy models can hinder the understanding of the whole energy industry and the identification of trends. To tackle this information gap and help improve future energy scenario models it is crucial to find a common baseline of comparison between energy scenario models and build a collection of deterministic variables that influence the outcome of the respective model.
For high-level managers and decision-makers in the energy industry it is imperative to improve forecasting capabilities in the energy field as long-term planning periods are common and can lead to huge sunk costs if developments in the energy market are misjudged. This is especially relevant in times of strong attention to greenhouse gas emissions and rapid shifts in policy and strategy. The viewing angle in the context of this thesis therefore has a macro orientation with
7
Master Thesis Alexander Hoffer, BSc.
a focus on the energy industry and related stakeholders. It builds on the concepts of Operations Research (OR), which involves the incorporation of mathematical methods to analyse problems and gain insights into a management problem (Merriam-Webster, 2020).
This thesis sets out to deliver a complete collection of used variables in energy scenario models and a statistical analysis that sheds light on usage patterns. Furthermore, a methodical analysis and description provide possibilities to recommend improvements for the future of energy modelling. It is argued that an expansion of the overall knowledge base in the context of energy systems and its deterministic variables ultimately leads to a better understanding and the ability to better forecast its future development.
In order to provide a contextual setting, a literature review of key energy technologies is carried out, discussing major advantages and shortcomings. It is essential to map the industry as well as its great variety of systems and technologies prior to diving into scenario models because several mechanisms or technologies might classify as novel or unintuitive. Ultimately, this thesis focuses not on the results and predictions of each scenario, but rather takes a close look at the considered factors and variables.
After all, this thesis presents 91 unique energy scenario variables, which can either be used as an inspiration to include in future models or to check for biases with regard to content coverage. This opportunity most certainly applies to managers in the energy industry, but also to developers of such energy models. Furthermore, the thesis offers an attempt to categorize unstandardized variables in this context and builds a base structure for potential further exploration of the topic.
Moreover, the statistical analysis carried out on top of the exploratory research provides an interesting insight into the inner works and apparent weightings of variable types in energy scenarios.
The thesis consists of the collection and analysis of papers and reports which present one or multiple long-term energy scenarios (to the year 2030 or further) and consider the complete energy mix of either regional, national or global scope. The methodology of the thesis is a meta- analysis.