4:15 – 4:25 pm Andy Sun
Strong SOCP Relaxations and Global Optimization for Network Constrained Nonconvex Optimization Problems.
Abstract: We will present a new global optimization scheme based on strong second-order conic programming (SOCP) relaxations for solving nonconvex optimization problems with network constraints. Extensive computation shows that the proposed scheme can be both faster and stronger than traditional SDP relaxations for solving the optimal power flow (OPF) problem on real-world sized instances.
4:25 – 4:35 pm Dileep Kalathil
Baseline Estimation and Scheduling for Demand Response
Demand Response (DR) programs serve to reduce the demand for electricity at times when the supply is scarce and expensive. Consumers with flexible consumption profiles are recruited by an aggregator who manages the DR program. These consumers are paid for reducing their energy consumption from contractually established baselines. Baselines are counter-factual consumption estimates against which load reductions are measured. Baseline consumption and the true cost of load reduction are consumer specific parameters and are unknown to the aggregator. The key components of any DR program are: (a) establishing a baseline against which demand reduction is measured, (b) designing the payment scheme for agents who reduce their consumption from this baseline, and (c) achieve these at a minimum cost to the aggregator. We propose a mechanism, self-reported baseline mechanism (SRBM), to address these problems. We show that truthful reporting of baseline and marginal utility is both incentive compatible and individually rational for every consumer under SRBM. We give a lower bound on the average cost of DR provision faced by the aggregator under any possible mechanism. We then propose a pod-sorting algorithm based DR scheduling. We show that the cost faced by the aggregator under SRBM with this pod-sorting scheduling it is nearly optimal.
4:35 – 4:45 pm Pratyush Chakraborty
Analysis of Solar Energy Aggregation under Various Billing Mechanisms
Abstract: Ongoing reductions in the cost of solar photovoltaic (PV) systems are driving their increased installations by residential households. Various incentive programs such as feed-in tariff, net metering, net purchase and sale that allow the prosumers to sell their generated electricity to the grid are also powering this trend. In this talk, we investigate sharing of PV systems among a community of households, who can also benefit further by pooling their production. Using cooperative game theory, we find conditions under which such sharing decreases their net total cost. We also develop allocation rules such that the joint net electricity consumption cost is allocated to the participants. Next, we perform a comparative analytical study on the benefit of sharing under the mechanisms favorable for sharing, namely net metering, and net purchase and sale.
4:45 – 5:00 pm Meng Wang
Data analytics by using low-dimensional models with applications in power system monitoring
Abstract: High-dimensional datasets often have intrinsic low-dimensional structures despite the ambient dimensions. For example, phasor measurement units in power systems generate synchronized phasor measurements at a rate of thirty samples per channel per second. The matrix that collects the spatial-temporal blocks of synchrophasor data can be approximated by low-rank matrices. The Hankel matrix is also low-rank due to the dynamic correlations.
In this talk, we will show that by exploiting the low-dimensional structures, one can develop computationally efficient methods with provable guarantees for missing data recovery, bad data correction, and data privacy enhancement. For instance, we have developed a first-order algorithm with a linear convergence rate to recover the original data, even when missing data and/or bad data exist across all channels consecutively. Leveraging the low-rank property, we have also developed a framework of cyber-resilient, communication-reduced, and information-preserved data collection framework.
