Predictive Demand-Side Energy Management Optimization


Demand-side management (DSM) techniques play a key role in enabling substantive energy savings, and alleviate demand peaks and carbon emissions in large urban communities such as New York City. However, these technologies are underutilized because they are often limited to electricity demand (i.e., not applied to electricity, gas and steam simultaneously).

This study introduces a novel real-time DSM framework that is not limited to electricity, and is referred to as Predictive Demand-side Energy Management Optimization (PDEMO).

PDEMO is mathematically formulated through the concept of model predictive control and optimization. The objective is to minimize the cost of building energy demand and consumption while satisfying constraints such as human comfort and operating characteristics of building equipment systems.

To illustrate the competitive benefits of PDEMO, it is applied to building startup procedures in high-rise commercial buildings equipped with cyber-physical systems. Here, the cost of energy (i.e., steam and electricity) is minimized by concurrently optimizing the set-points and sequence schedule.

Presented By

Ali Mehmani, Ph.D.
Head of Data Science and Analytics
Prescriptive Data

Ali Mehmani is a head of Data Analytics Research at Prescriptive Data (Rudin Management Co.) and an adjunct research scientist at Data Science Institute, Columbia University. He specializes in Multidisciplinary Optimization, Complex System Design, Deep Structural Networks and Cyber-Physical Systems, with focus on the following topics: heuristic optimization algorithms, machine learning, deep learning, recurrent neural networks, multi-fidelity modeling and optimization, and uncertainty analysis; and with primary applications in building energy-efficiency.

He graduated with his PhD in Mechanical Engineering. He has authored 2 book chapters, and more than 30 international journal and conference articles. He is a professional member of AEE, IEEE, ASME and a Senior Member of AIAA, and a selected member of the AIAA Multidisciplinary Design Optimization Technical Committee. He serves as a reviewer for 15 international journals in the areas of design, computing, and energy. He is also involved in organizing and chairing technical sessions on emerging topics in the ASME IDETC, and ASME PEC conferences.