arXiv subjects

K. Max Zhang

Publications and source records attributed to K. Max Zhang.

4 recordsLinked to original sources

arXiv2026

The Potential Welfare Gains from Curtailment Trading Under Non-Firm Interconnection

Rapid growth of large loads, especially data centers, is straining grid capacity and increasing interest in non-firm interconnection agreements that exchange faster grid access for curtailment exposure. This shift creates opportunities for differentiated reliability, where curtailment is allocated according to the value consumers place on uninterrupted service. This value is often expressed through the value of lost load (VOLL), an estimate of the cost a consumer bears for unserved energy. Because VOLL differs by more than a hundredfold across customer classes, pro-rata allocation, which cuts every load by the same proportion, ignores variation that could be leveraged to improve grid utilization. This paper introduces the network-constrained Curtailment Credit Market (CCM), a mechanism that lets one curtailable load pay another to take on part of its curtailment obligation. In this market, a high-VOLL load can reduce its own interruption by paying a lower-VOLL load to absorb additional curtailment. Crucially, the CCM clears while enforcing transmission limits. We prove that the CCM can implement every curtailment pattern available to an idealized planner that knows each load's VOLL and assigns curtailment directly. If agents report true lost-load values, CCM clearing attains the planner's total value of served load, the highest value achievable under network constraints. We evaluate the CCM on three test networks: a 3-bus network, the IEEE 24-bus network, and a reduced New York grid spanning multiple load zones. Across these networks, the CCM raises the total value of served load by 1.41 to 1.83 times relative to pro-rata curtailment.

eess.SY↗

arXiv2021

An Open Source Representation for the NYS Electric Grid to Support Power Grid and Market Transition Studies

Under the increasing need to decarbonize energy systems, there is coupled acceleration in connection of distributed and intermittent renewable resources in power grids. To support this transition, researchers and other stakeholders are embarking on detailed studies and analyses of the evolution of this complex system, which require a validated representation of the essential characteristics of the power grid that is accurate for a specific region of interest. For example, the Climate Leadership and Community Protection Act (CLCPA) in New York State (NYS) sets ambitious targets for the transformation of the energy system, opening many interesting research and analysis questions. To provide a platform for these analyses, this paper presents an overview of the current NYS power grid and develops an open-source (https://github.com/AndersonEnergyLab-Cornell/NYgrid) baseline model using publicly available data. The proposed model is validated with real data for power flow and Locational Marginal Prices (LMPs), demonstrating the feasibility, functionality, and consistency of the model. The model is easily adjustable and customizable for various analyses of future configurations and scenarios that require spatiotemporal information about the NYS power grid with data access to all the available historical data and serves as a practical system for general methods and algorithms testing.

eess.SY↗

arXiv2021

Generalized Reinforcement Learning for Building Control using Behavioral Cloning

Advanced building control methods such as model predictive control (MPC) offer significant potential benefits to both consumers and grid operators, but the high computational requirements have acted as barriers to more widespread adoption. Local control computation requires installation of expensive computational hardware, while cloud computing introduces data security and privacy concerns. In this paper, we drastically reduce the local computational requirements of advanced building control through a reinforcement learning (RL)-based approach called Behavioral Cloning, which represents the MPC policy as a neural network that can be locally implemented and quickly computed on a low-cost programmable logic controller. While previous RL and approximate MPC methods must be specifically trained for each building, our key improvement is that our controller can generalize to many buildings, electricity rates, and thermostat setpoint schedules without additional, effort-intensive retraining. To provide this versatility, we have adapted the traditional Behavioral Cloning approach through (1) a constraint-informed parameter grouping (CIPG) method that provides a more efficient representation of the training data; (2) an MPC-Guided training data generation method using the DAgger algorithm that improves stability and constraint satisfaction; and (3) a new deep learning model-structure called reverse-time recurrent neural networks (RT-RNN) that allows future information to flow backward in time to more effectively interpret the temporal information in disturbance predictions. The result is an easy-to-deploy, generalized behavioral clone of MPC that can be implemented on a programmable logic controller and requires little building-specific controller tuning, reducing the effort and costs associated with implementing smart residential heat pump control.

eess.SY↗

arXiv2019

Convexity and monotonicity in nonlinear optimal control under uncertainty

We consider the problem of finite-horizon optimal control design under uncertainty for imperfectly observed discrete-time systems with convex costs and constraints. It is known that this problem can be cast as an infinite-dimensional convex program when the dynamics and measurements are linear, uncertainty is additive, and the risks associated with constraint violations and excessive costs are measured in expectation or in the worst case. In this paper, we extend this result to systems with convex or concave dynamics, nonlinear measurements, more general uncertainty structures and other coherent risk measures. In this setting, the optimal control problem can be cast as an infinite-dimensional convex program if (1) the costs, constraints and dynamics satisfy certain monotonicity properties, and (2) the measured outputs can be reversibly `purified' of the influence of the control inputs through Q- or Youla-parameterization. The practical value of this result is that the finite-dimensional subproblems arising in a variety of suboptimal control methods, notably including model predictive control and the Q-design procedure, are also convex for this class of nonlinear systems. Subproblems can therefore be solved to global optimality using convenient modeling software and efficient, reliable solvers. We illustrate these ideas in a numerical example.

math.OC↗