Development of a DSO support tool for congestion forecast

Abstract

This paper presents a novel DSO support tool with visualisation capability for forecasting network congestion in distribution systems with a high level of renewables. To incorporate the uncertainties in the distribution systems, the probabilistic power flow framework has been utilised. An advanced photovoltaic production forecast based on sky images and a load forecast using an artificial neural network is used as the input to the tool. In addition, advanced load models and operating modes of photovoltaic inverters have been incorporated into the tool. The tool has been applied in case studies to perform congestion forecasts for two real distribution systems to validate its usability and scalability. The results from case studies demonstrated that the tool performs satisfactorily for both small and large networks and is able to visualise the cumulative probabilities of nodes voltage deviation and network components (branches and transformers) congestion for a variety of forecast horizons as desired by the DSO. The results have also shown that explicit inclusion of load-voltage dependency models would improve the accuracy of the congestion forecast. For demonstrating the applicability of the tool, it has been integrated into an existing distribution management system via the IoT platform of a DMS vendor, Atos Worldgrid.

Pathways for the Development of Future Intelligent Distribution Grids

“…The tool determines the probability of congestion to occur in a distribution network as well as at the individual elements (i.e., transformers, lines, and nodes). Currently, most of the DSOs as indicated in Subsection 4.1.1, are not equipped with similar tools, but their need is expected to grow in near future (Srivastava et al, 2019a(Srivastava et al, , 2021. This tool was demonstrated at the Gothenburg demonstration site in Sweden.…”

Predicting Renewable Curtailment in Distribution Grids Using Neural Networks

“…In literature ANN were already successfully applied in the context of congestion detection [5], RE power [15], [16], [17] and consumption prediction [18]. An overview of those ANNs is given in Table 1.…”

“…Srivastava et al proposed an approach relying on Monte Carlo simulation and probabilistic load flow to predict congestions occuring in LV grid levels. In their study generation and consumption are modeled by applying meteorological data gathered by fisheye lens cameras to predict PV power production and ANNs for load forecast [18]. Another aproach is to use ANN for congestion prediction in electricity grids.…”

“…In order to predict possible congestions occuring in the distribution grid, injections into grid also have to be predicted. The injections can be simulated by modeling RE injection and load separately [18], using meteorological information and information about consumption for phase prediction [19]. Focusing on congestion prediction in [20] they used active and reactive power of each bus as input features.…”

Single‐Site Forecasts for 130 Photovoltaic Systems at Distribution System Operator Level, Using a Hybrid‐Physical Approach, to Improve Grid‐Integration and Enable Future Smart‐Grid Operation

“…Hence, the operation of flexible grid assets or the flexibility procurement requires detailed forecasts at high spatial resolutions, preferably on day‐ahead (DA) and intraday (ID) levels. [ 4 ]…”