-
Electricity Price Prediction Using Multi-Kernel Gaussian Process Regression Combined with Kernel-Based Support Vector Regression
Authors:
Abhinav Das,
Stephan Schlüter,
Lorenz Schneider
Abstract:
This paper presents a new hybrid model for predicting German electricity prices. The algorithm is based on combining Gaussian Process Regression (GPR) and Support Vector Regression (SVR). While GPR is a competent model for learning the stochastic pattern within the data and interpolation, its performance for out-of-sample data is not very promising. By choosing a suitable data-dependent covariance…
▽ More
This paper presents a new hybrid model for predicting German electricity prices. The algorithm is based on combining Gaussian Process Regression (GPR) and Support Vector Regression (SVR). While GPR is a competent model for learning the stochastic pattern within the data and interpolation, its performance for out-of-sample data is not very promising. By choosing a suitable data-dependent covariance function, we can enhance the performance of GPR for the tested German hourly power prices. However, since the out-of-sample prediction depends on the training data, the prediction is vulnerable to noise and outliers. To overcome this issue, a separate prediction is made using SVR, which applies margin-based optimization, having an advantage in dealing with non-linear processes and outliers, since only certain necessary points (support vectors) in the training data are responsible for regression. Both individual predictions are later combined using the performance-based weight assignment method. A test on historic German power prices shows that this approach outperforms its chosen benchmarks such as the autoregressive exogenous model, the naive approach, as well as the long short-term memory approach of prediction.
△ Less
Submitted 14 January, 2025; v1 submitted 28 November, 2024;
originally announced December 2024.
-
Optimal Control of a Battery Storage On the Energy Market
Authors:
Stephan Schlüter,
Abhinav Das,
Mathew Davison
Abstract:
Electricity storage is crucial for a successful transition towards carbon-neutral energy production. Despite considerable research and a number of promising future alternatives such as hydrogen, battery storages currently remain the first choice. However, costs remain high and it remains to be shown whether an investment can be profitable. This article addresses this question by modelling a batter…
▽ More
Electricity storage is crucial for a successful transition towards carbon-neutral energy production. Despite considerable research and a number of promising future alternatives such as hydrogen, battery storages currently remain the first choice. However, costs remain high and it remains to be shown whether an investment can be profitable. This article addresses this question by modelling a battery storage operating in the German power market. We consider two periods with very distinct price dynamics, namely a calm year (2020) and a turbulent year (2023). It shows that even for low battery costs a 2020 style price environment does not allow for profitable battery operation, whereas current market conditions allow for positive payoffs.
△ Less
Submitted 29 July, 2024;
originally announced July 2024.
-
Interval Forecasts for Gas Prices in the Face of Structural Breaks -- Statistical Models vs. Neural Networks
Authors:
Stephan Schlüter,
Sven Pappert,
Martin Neumann
Abstract:
Reliable gas price forecasts are an essential information for gas and energy traders, for risk managers and also economists. However, ahead of the war in Ukraine Europe began to suffer from substantially increased and volatile gas prices which culminated in the aftermath of the North Stream 1 explosion. This shock changed both trend and volatility structure of the prices and has considerable effec…
▽ More
Reliable gas price forecasts are an essential information for gas and energy traders, for risk managers and also economists. However, ahead of the war in Ukraine Europe began to suffer from substantially increased and volatile gas prices which culminated in the aftermath of the North Stream 1 explosion. This shock changed both trend and volatility structure of the prices and has considerable effects on forecasting models. In this study we investigate whether modern machine learning methods such as neural networks are more resilient against such changes than statistical models such as autoregressive moving average (ARMA) models with conditional heteroskedasticity, or copula-based time series models. Thereby the focus lies on interval forecasting and applying respective evaluation measures. As data, the Front Month prices from the Dutch Title Transfer Facility, currently the predominant European exchange, are used. We see that, during the shock period, most models underestimate the variance while overestimating the variance in the after-shock period. Furthermore, we recognize that, during the shock, the simpler models, i.e. an ARMA model with conditional heteroskedasticity and the multilayer perceptron (a neural network), perform best with regards to prediction interval coverage. Interestingly, the widely-used long-short term neural network is outperformed by its competitors.
△ Less
Submitted 23 July, 2024;
originally announced July 2024.
-
A geometric state function for two-fluid flow in porous media
Authors:
James E. McClure,
Ryan T. Armstrong,
Mark A. Berrill,
Steffen Schlüter,
Steffen Berg,
William G. Gray,
Cass T. Miller
Abstract:
Models that describe two-fluid flow in porous media suffer from a widely-recognized problem that the constitutive relationships used to predict capillary pressure as a function of the fluid saturation are non-unique, thus requiring a hysteretic description. As an alternative to the traditional perspec- tive, we consider a geometrical description of the capillary pressure, which relates the average…
▽ More
Models that describe two-fluid flow in porous media suffer from a widely-recognized problem that the constitutive relationships used to predict capillary pressure as a function of the fluid saturation are non-unique, thus requiring a hysteretic description. As an alternative to the traditional perspec- tive, we consider a geometrical description of the capillary pressure, which relates the average mean curvature, the fluid saturation, the interfacial area between fluids, and the Euler characteristic. The state equation is formulated using notions from algebraic topology and cast in terms of measures of the macroscale state. Synchrotron-based X-ray micro-computed tomography (μCT) and high- resolution pore-scale simulation is applied to examine the uniqueness of the proposed relationship for six different porous media. We show that the geometric state function is able to characterize the microscopic fluid configurations that result from a wide range of simulated flow conditions in an averaged sense. The geometric state function can serve as a closure relationship within macroscale models to effectively remove hysteretic behavior attributed to the arrangement of fluids within a porous medium. This provides a critical missing component needed to enable a new generation of higher fidelity models to describe two-fluid flow in porous media.
△ Less
Submitted 24 May, 2018;
originally announced May 2018.
-
Biofilm growth in porous media: experiments, computational modeling at the porescale, and upscaling
Authors:
Malgorzata Peszynska,
Anna Trykozko,
Gabriel Iltis,
Steffen Schlueter,
Dorthe Wildenschild
Abstract:
Biofilm growth changes many physical properties of porous media such as porosity, permeability and mass transport parameters. The growth depends on various environmental conditions, and in particular, on flow rates. Modeling the evolution of such properties is difficult both at the porescale where the phase morphology can be distinguished, as well as during upscaling to the corescale effective pro…
▽ More
Biofilm growth changes many physical properties of porous media such as porosity, permeability and mass transport parameters. The growth depends on various environmental conditions, and in particular, on flow rates. Modeling the evolution of such properties is difficult both at the porescale where the phase morphology can be distinguished, as well as during upscaling to the corescale effective properties. Experimental data on biofilm growth is also limited because its collection can interfere with the growth, while imaging itself presents challenges.
In this paper we combine insight from imaging, experiments, and numerical simulations and visualization. The experimental dataset is based on glass beads domain inoculated by biomass which is subjected to various flow conditions promoting the growth of biomass and the appearance of a biofilm phase. The domain is imaged and the imaging data is used directly by a computational model for flow and transport. The results of the computational flow model are upscaled to produce conductivities which compare well with the experimentally obtained hydraulic properties of the medium. The flow model is also coupled to a newly developed biomass--nutrient growth model, and the model reproduces morphologies qualitatively similar to those observed in the experiment.
△ Less
Submitted 24 September, 2015;
originally announced October 2015.
