Skip to content
Images of particles made from a promising battery cathode material called NMC

Planning Reliable Grid with Variable Generation and Storage

Bits & Watts Initiative

PI: Sanjay Lall, Electrical Engineering; Co-PI: Dimitry Gorinevsky, Electrical Engineering;  Weixuan Gao (Ph.D. student)

Critical Need: 

Managing risk and cost of the grid with increasing variable renewable generation and storage at scales beyond anything seen before requires new probabilistic analysis tools.

Project Innovation: 

This project develops probabilistic tools that greatly improve the speed and accuracy of the reliability analysis compared to traditionally used Monte Carlo methods. The analysis approach is based on discrete convolutions of conditional probability distributions for variables that might be interdependent. The approach to storage dispatch and energy adequacy analysis is related to Information Theory.

FY19 Accomplishments and Results

The project has developed ML algorithms for analysis of grid reliability with variable distributed energy resources (DERs) and a software tool called RAPIER (Risk Analysis and Prediction for Integrated Energy from Renewables) is currently under development. At present RAPIER exists as Matlab tool that is being verified and improved in increasing number of examples and use cases. 

Potential Impact:

These new probabilistic tools model variable demand, solar, and wind generation as interdependent random variables. They allow the adequacy of the resource mix and grid reliability to be assessed more accurately and with less effort than existing approaches to planning. The tools allow risk management for resource portfolios with high percentages of variable renewable sources, and possibly, with Demand Response (DR) and battery storage. Reducing use of dispatchable generation to a minimum contributes to reduced carbon emissions. The tools allow operators to analyze and minimize the risk of the unbalance between demand and supply in the presence of large fractions of renewable energy, storage, DR, and EV load in the portfolio, thus, increasing grid reliability and resilience. The tools can be also deployed on-line to balance the costs of peak renewable generation and demand probabilistically, with high accuracy, minimizing the risk of overpayment. The outcome would be pre-defined risk of spot market exposure with minimum procurement cost.


1.  IEEE Transactions on Power System (Submitted), Probabilistic Modeling for Optimization of Resource Mix with Variable Generation and Storage, Full Paper, Weixuan Gao and Dimitry Gorinevsky,

2.  IEEE PES General Meeting 2019, Atlanta, GA, Probabilistic Planning of Minigrid with Renewables and Storage in Western Australia, Paper, Weixuan Gao, Dimitry Gorinevsky, and Dev Tayal 

3.  Reliability, Risk and Probability Application, (RRPA) Meeting, 2019, IEEE Analytical Methods for Power Systems (AMPS) Committee, LOLE Analysis with Storage and Deep Renewables, Invited Talk, Dimitry Gorinevsky, August 8, 2019, Atlanta, GA