Probabilistic Forecasting and Simulation of Electricity Markets via Online Dictionary Learning

Abstract

Abstract-The problem of probabilistic forecasting and online simulation of real-time electricity market with stochastic generation and demand is considered. By exploiting the parametric structure of the direct current optimal power flow, a new technique based on online dictionary learning (ODL) is proposed. The ODL approach incorporates real-time measurements and historical traces to produce forecasts of joint and marginal probability distributions of future locational marginal prices, power flows, and dispatch levels, conditional on the system state at the time of forecasting. Compared with standard Monte Carlo simulation techniques, the ODL approach offers several orders of magnitude improvement in computation time, making it feasible for online forecasting of market operations. Numerical simulations on large and moderate size power systems illustrate its performance and complexity features and its potential as a tool for system operators.Index Terms-Dictionary learning, electricity market, machine learning in power systems, power flow distributions, probabilistic price forecasting. NOMENCLATURE c(·)Real-time generation cost function. Shadow prices for max/min transmission constraints at time t.

A Holistic Approach to Forecasting Wholesale Energy Market Prices

“…In the former case, the full knowledge of supply bids, grid topology, and physical properties of the network allow for the explicit computation of nodal LMPs as dual variables of the corresponding OPF optimization. The relevant papers mostly study the impact of uncertainty in total grid load or renewable generation on the resulting prices, while relying on the fact that changes in LMP regimes happen at the so called critical load levels [24]- [28]. The number of the critical load levels exponentially grows with the size of a grid, making the proposed approaches intractable for use in practice.…”

Optimal Facility Allocation and Determination of Demand Response Participation Rate Considering Uncertainties in Power Systems

“…Second, we assume that the PV outputs and DR capabilities for 1 h follow a Gaussian distribution. In fact, this assumption is not perfect but we employ this because Gaussian distribution has often been used in other research related to RE and demands in order to simplify the formulation, including uncertainties [9][10][11][12]. The use of other kinds of probability distributions is reserved for future work.…”

“…1, pp. [10][11][12][13][14][15][16][17][18][19]2017 Paper Optimal Facility Allocation and Determination of Demand Response Participation Rate Considering Uncertainties in Power Systems* Ryusuke Konishi † § and Masaki Takahashi ‡ § It is necessary to install a large number of renewable energy (RE) systems such as photovoltaic systems (PVs) to solve environmental problems. However, an electricity grid with RE systems experiences problems such as power shortages and surpluses because of uncertainties in generating outputs.…”

An Online Risk Based Security Assessment Via Conditional Value-at-risk in Uncertain Environment

“…Depends on the Monte Carlo simulation technique, the generated sample sets, high-dimensional uncertainty power injections located at different buses, are concentrated as a cluster according to the character of the forecast error distributions. We provide a fast SCOPF algorithm based on dynamic critical region exploration for the online simulation and forecast [6]. Based on the very short-term load/ renewable generator output probabilistic forecast, we only focus on the particular critical regions the parameters belongs to, instead of partition the whole space.…”

“…Assume that there is no degeneracy, the critical region  and corresponding optimal value have the following relationships [6]:…”