[Paper Review] Graph based Platform for Electricity Market Study, Education and Training

Tao Chen, Chen Yuan+2 more|arXiv (Cornell University)|Apr 3, 2018

Electric Power System Optimization9 references3 citations

TL;DR

This paper proposes a TigerGraph-based graph computing framework that accelerates security-constrained unit commitment (SCUC) and economic dispatch (SCED) for electricity market clearing using parallelized graph power flow (PGPF) and optimized LU decomposition. The system enables real-time visualization of locational marginal prices (LMP) and supports education and training for market operators through interactive, scalable simulation of large power networks.

ABSTRACT

With the further development of deregulated electricity market in many other countries around the world, a lot of challenges have been identified for market data management, network topology processing and fast market-clearance mechanism design. In this paper, a graph computing framework based on TigerGraph database is proposed to solve a security constrained unit commitment (SCUC) and security constrained economic dispatch (SCED) problem, with parallelized graph power flow (PGPF) and innovative LU decomposition techniques, for electricity market-clearance. It also provides a comprehensive visualization platform to demonstrate the market clearing results vividly, such as locational marginal price (LMP), and is able to be utilized for electricity market operators' education and training purpose.

Motivation & Objective

• To address challenges in managing large-scale market data, network topology, and fast market-clearing in deregulated electricity markets.
• To develop a scalable, high-performance simulation framework for day-ahead and real-time electricity market operations.
• To integrate AC power flow constraints into SCUC/SCED with computational efficiency using graph computing.
• To provide an interactive, visual platform for training and educating electricity market operators.

Proposed method

• The framework uses TigerGraph’s native graph database and computing engine to model power systems as graphs with buses as vertices and transmission lines as edges.
• It implements parallelized graph power flow (PGPF) using TigerGraph’s bulk-synchronous parallel (BSP) model for efficient AC power flow computation.
• Innovative LU decomposition techniques are applied to solve large linear systems arising in SCUC/SCED problems in a distributed, graph-parallel manner.
• The system employs a time-tree data structure to organize temporal market data across multiple hours, enabling dynamic bidding and simulation.
• Market data, including bids, generation limits, and network parameters, are stored as properties on graph vertices and edges for flexible querying and processing.
• A GraphStudio-based visualization layer renders LMPs, congestion patterns, and dispatch results in real time for intuitive interpretation.

Experimental results

Research questions

• RQ1 How can graph computing be leveraged to accelerate SCUC and SCED computations in electricity markets with AC power flow constraints?
• RQ2 What performance gains can be achieved using a graph-based parallel computing model compared to traditional methods in large-scale power systems?
• RQ3 Can a unified graph data model effectively represent complex market and network data for real-time market clearing simulations?
• RQ4 How effective is the platform in supporting education and training for electricity market operators through interactive visualization?
• RQ5 To what extent does the time-tree structure enable efficient temporal data management and dynamic market simulation?

Key findings

• The system achieved a market-clearing runtime of 2804.396 ms for the IEEE-14 bus system using 8 threads, demonstrating high computational efficiency.
• The graph-based approach maintained near-linear computational complexity, avoiding the exponential growth seen in traditional methods.
• Visualization of LMPs and congestion effects was successfully rendered in real time using GraphStudio, enabling intuitive understanding of market outcomes.
• The time-tree structure enabled efficient storage and retrieval of temporal market data, supporting dynamic bidding and scenario analysis.
• The platform successfully simulated a 118-bus system, confirming scalability and performance on large-scale networks.
• The integration of PGPF and optimized LU decomposition allowed accurate AC power flow solutions within a market-clearing framework, improving upon DC approximation methods.