California Independent System Operator Research Articles

Market-Driven Energy Storage Planning for Microgrids with Renewable Energy Systems Using Stochastic Programming

Battery Energy Storage Systems (BESS) can mitigate effects of intermittent energy production from renewable energy sources and play a critical role in peak shaving and demand charge management. To optimally size the BESS from an economic perspective, the tradeoff between BESS investment costs, lifetime, and revenue from utility bill savings along with microgrid ancillary services must be taken into account. The optimal size of a BESS is solved via a stochastic optimization problem considering wholesale market pricing. A stochastic model is used to schedule arbitrage services for energy storage based on the forecasted energy market pricing while accounting for BESS cost trends, the variability of renewable energy resources, and demand prediction. The uniqueness of the approach proposed in this paper lies in the convex optimization programming framework that computes a globally optimal solution to the financial trade-off solution. The approach is illustrated by application to various realistic case studies based on pricing and demand data from the California Independent System Operator (CAISO). The case study results give insight in optimal BESS sizing from a cost perspective, based on both yearly scheduling and daily BESS operation.

Wind Power Providing Flexible Ramp Product

The deepening penetration of renewables in power systems has contributed to the increasing needs for generation scheduling flexibility. Specifically, for short-term operations, flexibility here indicates that sufficient ramp capacities should be reserved to respond to the expected changes in the load and intermittent generation, also covering a certain amount of their uncertainty. To address the growing requirements for flexible ramp capacity, markets for ramp products have been launched in practice such as the ones in California ISO and Midcontinent ISO. Some-times, to guarantee sufficient ramp capacity, expensive fast start units have to be committed in real-time. Moreover, with higher penetration of renewable generation, the flexibility provided by the conventional units might not be enough. Actually, wind power producers are physically capable of offering flexibility, which is sometimes also economically efficient to the entire system. In this paper, we aim to explore the mechanism and possibility of including wind power producers as ramp providers to increase the supply of flexibility. To conduct the anal-yses, a two-stage stochastic real-time unit commitment model considering ramp capacity adequacy is formulated. Case studies indicate that both the system and the wind power producers can benefit if the wind power is allowed to provide flexible ramp products.

Unit Commitment Under Gas-Supply Uncertainty and Gas-Price Variability

We propose a two-stage mixed-integer linear stochastic optimization model to analyze the scheduling of electricity-production units under natural gas-supply uncertainty due to pipeline congestion and natural gas-price variability. The first stage of this stochastic optimization model represents the day-ahead scheduling ( i . e ., unit commitment) stage, while the second stage represents actual real-time operations through a number of scenarios. We use this model to analyze the effect on unit commitment and dispatch of two types of natural gas-supply conditions. First, we analyze a case involving low-cost natural gas supply with natural gas-transmission issues related to potential gas-pipeline congestion. We then examine a case involving higher-cost natural gas, which is used solely to attain feasibility with fast-ramping events. The first case mimics situations in the ISO New England system, in which relatively low-cost natural gas supply is uncertain in cold-weather conditions due to natural gas-transmission bottlenecks. The second case is reminiscent of situations in the California ISO system, in which relatively expensive but flexible natural gas-fired units need to be used to handle rapid changes in net demand in the early mornings and late afternoons.

Optimal Bidding in Performance-Based Regulation Markets: An MPEC Analysis With System Dynamics

In this paper, we address the problem of optimal bidding in performance-based regulation markets for a large price-maker regulation resource. Focusing on the case of the California Independent System Operator (ISO), detailed market components are considered, such as regulation capacity payment, regulation mileage payment, performance accuracy adjustment, automatic generation control dispatch, and participation factor. Our analysis also incorporates system dynamics of the regulation resource for different resource types and technologies. In principle, our problem formulation is a mathematical program with equilibrium constraints (MPEC). However, our fundamentally new formulations introduce several new challenges in solving the MPEC problem in the context of performance-based regulation markets that are not previously addressed. In fact, global optimization techniques fail to solve the original nonlinear program due to its complexity. Therefore, we undergo several innovative steps to transform the problem into a mixed-integer linear program which is solved with accuracy, reliability, and computational efficiency. Insightful case studies are presented using data from a California ISO regulation market project.

A multi-scale optimization framework for electricity market participation

Power grids coordinate a diverse set of energy systems (generators, loads, storage devices) to ensure that supply and demands are matched at multiple timescales (from hours to milliseconds). Such coordination is achieved through hierarchical market transactions. This work presents an optimization framework to evaluate revenue opportunities provided by these multi-scale market hierarchies and to determine optimal participation strategies for individual participants. The proposed framework models day-ahead and real-time transactions of energy, ancillary services, and virtual bidding products provided by independent system operators (ISOs). We apply the framework to a combined heat and power system and a utility-scale battery to determine revenue potential from different market layers and products. Analysis using real price signals for 2015 from the California ISO reveals that 60–90% of the total revenue potential (obtained by participating in all markets) is provided by real-time markets alone (which operate at fast timescales). Our studies also indicate that providing ancillary services (in addition to day-ahead and real-time energy products) increases revenue potential by 40–100%, depending on the physical flexibility of the technology. The proposed framework can be used to identify which market layers and products offer the greatest economic potential for different energy technologies. Our results also highlight that existing techno-economic studies that focus exclusively on day-ahead energy markets (operating at slower timescales) can dramatically undervalue dynamic flexibility.

Quantification and regional comparison of water use for power generation: A California ISO case study

Abstract Analysis of water use for power generation has, in the past, focused on large geographical regions and time scales. Attempting to refine this analysis on the time and spatial scales could help to further understand the complex relationships involved in the energy–water nexus, specifically, the water required to generate power. Water factors for different types of plants and cooling systems are used from literature in combination with power generation data for different balancing authorities to model water use as a function of time based on the fuel mix and power generated for that region. This model is designed to increase public awareness of the interrelation between the energy consumed and water use that can be taken into account when making decisions about electrical energy use. These results confirm that areas with higher renewable energy penetration use less water per unit of power generated than those with little or no renewable technologies in the area, but this effect is heavily dependent on the distribution of the types of renewable and conventional generation used.

Impact of Virtual Bidding on Forward Premia in the California Independent System Operator

In the California wholesale electricity market, physical power is sold competitively in a multi-unit multi-settlement single-price auction comprised of a day-ahead forward market and a real-time spot market. Historically, significant divergences have arisen between forward and spot prices, leading to market inefficiencies. In 2011, the California Independent System Operator implemented virtual bidding as a partial solution to converge prices between markets by opening the forward market to financial speculators without any ties to physical generation or load. Using 5 years of data between April 2009 and March 2014, I estimate hourly time-series models of the system-wide forward premium, and forward premia at the two largest trading hubs. I find virtual bids and off ers to have had no effect on the forward premia during off -peak hours. During peak hours, I find virtual bidding to have caused forward and spot prices to diverge due to the large number of market participants looking to hedge against — or speculate on — the occurrence of infrequent but large spot price spikes by placing virtual demand bids. Virtuals placed at the largest trading hubs were associated with even larger price divergences, suggesting virtual demand bids also distorted the price of congestion and losses at the nodal level.

Short-Term Forward and Spot Prices in Organized Wholesale Electricity Markets

In modern organized wholesale electricity markets, power is sold competitively in a two-settlement auction comprised of a short-term forward (or day-ahead) market with elastic demand, and a spot (or real-time) imbalance market characterized by an inelastic residual demand curve - a construct that the literature on the subject tends to ignore. To understand the strategic behavior of firms looking maximize joint pro fits over both markets, I construct a two-settlement model of electricity prices in which firms compete to meet demand from consumers, first in quantities, then in prices. I show that the ensuing forward premium depends on the costliness of spot commitment relative to firms' ability to exercise market power by setting quantities in the forward market. Using hourly data from the California Independent System Operator, I find that the model helps explain price formation in the California market, and that the forward market solutions are close to the Cournot outcome due to the relative insensitivity of the spot equilibrium price to individual firm's forward commitment decisions.

A Closer Look at Demand Bids in California ISO Energy Market

One year of demand bids in the California energy market are analyzed and observations are reported on type, size, shape, and other characteristics of the bids. The implications of these observations, the underlying causes, and the potentials to improve demand bids by exploiting load flexibility are discussed.

Level versus Variability Trade-offs in Wind and Solar Generation Investments: The Case of California

Hourly plant-level wind and solar generation output and real-time price data for one year from the California ISO control area is used to estimate the vector of means and the contemporaneous covariance matrix of hourly output and revenues across all wind and solar locations in the state. Annual hourly output and annual hourly revenues mean/standard deviation efficient frontiers for wind and solar resource locations are computed from this information. For both efficient frontiers, economically meaningful differences between portfolios on the efficient frontier and the actual wind and solar generation capacity mix are found. The relative difference is significantly larger for aggregate hourly output relative to aggregate hourly revenues, consistent with expected profit-maximizing unilateral entry decisions by renewable resource owners. Most of the hourly output and hourly revenue risk-reducing benefits from the optimal choice of locational generation capacities is captured by a small number of wind resource locations, with the addition of a small number of solar resource locations only slightly increasing the set of feasible portfolio mean and standard deviation combinations. Measures of non-diversifiable wind and solar energy output and revenue risk are computed using the actual market portfolio and the risk-adjusted expected hourly output or hourly revenue maximizing portfolios.

Unit Commitment With Continuous-Time Generation and Ramping Trajectory Models

There is increasing evidence of shortage of ramping resources in the real-time operation of power systems. To explain and remedy this problem systematically, in this paper we take a novel look at the way the day-ahead unit commitment (UC) problem represents the information about load, generation and ramping constraints. We specifically investigate the approximation error made in mapping of the original problem, that would decide the continuous-time generation and ramping trajectories of the committed generating units, onto the discrete-time problem that is solved in practice. We first show that current practice amounts to approximating the trajectories with linear splines. We then offer a different representation through cubic splines that provides physically feasible schedules and increases the accuracy of the continuous-time generation and ramping trajectories by capturing sub-hourly variations and ramping of load in the day-ahead power system operation. The corresponding day-ahead UC model is formulated as an instance of mixed-integer linear programming (MILP), with the same number of binary variables as the traditional UC formulation. Numerical simulation over real load data from the California ISO demonstrate that the proposed UC model reduces the total day-ahead and real-time operation cost, and the number of events of ramping scarcity in the real-time operations.

Merit-order effects of renewable energy and price divergence in California’s day-ahead and real-time electricity markets

We answer two policy questions: (1) what are the estimated merit-order effects of renewable energy in the California Independent System Operator’s (CAISO’s) day-ahead market (DAM) and real-time market (RTM)? and (2) what causes the hourly DAM and RTM prices to systematically diverge? The first question is timely and relevant because if the merit-order effect estimates are small, California’s renewable energy development is of limited help in cutting electricity consumers’ bills but also has a lesser adverse impact on the state’s investment incentive for natural-gas-fired generation. The second question is related to the efficient market hypothesis under which the hourly RTM and DAM prices tend to converge. Using a sample of about 21,000 hourly observations of CAISO market prices and their fundamental drivers during 12/12/2012–04/30/2015, we document statistically significant estimates (p-value≤0.01) for the DAM and RTM merit-order effects. This finding lends support to California’s adopted procurement process to provide sufficient investment incentives for natural-gas-fired generation. We document that the RTM-DAM price divergence partly depends on the CASIO’s day-ahead forecast errors for system loads and renewable energy. This finding suggests that improving the performance of the CAISO’s day-ahead forecasts can enhance trading efficiency in California’s DAM and RTM electricity markets.

Real-Time Markets for Flexiramp: A Stochastic Unit Commitment-Based Analysis

Markets for ramp products (called “flexiramp” in the California ISO and “ramp capability” in the Midcontinent ISO) have been proposed to address the need for generator flexibility due to increased renewable penetration. Flexiramp is committed capacity reserved to meet unexpected ramps in net load. We evaluate flexiramp market designs by comparing solutions of 1) an ISO real-time market simulation (a deterministic real-time unit commitment (RTUC) model with flexiramp requirements) with 2) a stochastic RTUC model, used as a baseline for evaluating flexiramp solutions. The flexiramp constraint improves the expected performance of ISO real-time markets, but is inefficient compared to stochastic RTUC. Flexiramp markets are inefficient in part because the deterministic model disregards the possibility of higher than anticipated energy production by capacity reserved for ramp. Thus, for instance, capacity with low commitment costs but high running costs might be overcommitted, increasing energy costs under high ramp outcomes. The amount of flexiramp acquired strongly affects costs. Too little results in under-commitment of generators and high price fluctuations and even demand curtailment; too much leads to over-commitment, inflating production costs. Hence, careful design of the ramp market and consideration of the use of full stochastic dispatch and commitment models is recommended.

Optimal Bidding, Scheduling, and Deployment of Battery Systems in California Day-Ahead Energy Market

An optimal supply and demand bidding, scheduling, and deployment design framework is proposed for battery systems. It takes into account various design factors such as the day-ahead and real-time market prices and their statistical dependency, as well as the location, size, efficiency, lifetime, and charge and discharge rates of the batteries. Utilizing second-life/used batteries is also considered. Without loss of generality, our focus is on the California Independent System Operator (ISO) energy market and its two available bidding options, namely self-schedule bidding and economic bidding. While the formulated stochastic optimization problems are originally nonlinear and difficult to solve, we propose a methodology to decompose them into inner and outer subproblems. Accordingly, we find the global optimal solutions within a short amount of computational time. All case studies in this paper are based on real market data.

Sequential Contracts for Uncertain Electricity Resources

The U.S. department of energy recently listed “development of rules for market evolution that enable system flexibility” among the key strategic areas of intervention for a successful integration of renewable energy into the grid [2]. Renewable energy generation from wind and solar is intermittent and its prediction accuracy is precise only within a short time horizon (e.g. 5-15 minutes [1]). Therefore, new market mechanism solutions such as intra-day markets are proposed that allow for flexible generation of renewable energy to exploit the improved forecast accuracy of renewable resources over time [14]. In this paper we study sequential contract design problems that incorporate the arrival of new information about renewable generation and allow for flexible production. Today, the renewable energy generation receives extramarket treatment such as feed-in tariffs, guaranteed grid access, and lenient penalty rate [3, 7]. For example, the Participating Intermittent Resource Program (PIRP) mandates the California independent system operator to accept all the wind generation in real-time and treat them as negative loads. The subsequent increased cost of the required reserve generation capacity is then socialized among the load serving entities (LSE). However, such approaches cannot be sustained for high levels of renewable generation as the imposed reserve generation cost on LSEs becomes excessively high and results in social welfare loss. In the long-run, renewable energy generation needs to participate in electricity markets and be exposed to market mechanisms. An alternative approach, implemented in the U.K., requires the wind generators to bid in conventional electricity markets and pay penalty for ex-post deviation from their ex-ante contracted schedule. Such a firm contracting approach is the subject of many studies in the literature. The works of [4, 15] study the problem of optimal bidding in a two-settlement market structure with an exogenous price and penalty rate. The problem of mechanism design for wind aggregation among many wind producers that jointly participate in a two-settlement market structure with exogenous price and penalty rate is investigated in [8, 10]. The authors in [12] study the problem of auction design for such a two-settlement market structure. In this paper, we propose a simple two-step model to capture the dynamic variable nature of renewable generation and provide a general formulation for flexible forward con-

Envelope modeling of renewable resource variability and capacity

To unify the analysis of both renewable and conventional fossil-fuel generating resources in electricity systems, we develop an envelope-based modeling method. Built on Network Calculus theory (NetCal) for deterministic queuing systems from the field of telecommunications engineering, this method characterizes the variability of electricity supply and demand by upper and lower envelopes and their respective Legendre conjugates. Differing from all other modeling methods, this method not only quantifies variability across different time scales, but also captures the intrinsic tradeoff between capacity and the corresponding Quality-of-Service (QoS) performance. In particular, the QoS measures represent matching/mismatching patterns between power supply and demand and provide an intuitive interpretation of the role of storage resources. The concept of QoS leads to two QoS-based capacity metrics – guaranteed capacity and best-effort capacity – whose conceptual and numerical properties are analyzed and compared against existing capacity metrics for validation purpose. As illustration, the proposed methods are applied to data from the California Independent System Operator (CAISO), which allows us to explicitly quantify the capacity contribution (via the notion of best-effort capacity) of wind during peak hours and its negative system impact at night, and demonstrates the positive capacity contribution of storage resources even though they are net energy consumer.

Efficiency impact of convergence bidding in the california electricity market

The California Independent System Operator (CAISO) has implemented Convergence Bidding (CB) on February 1, 2011 under Federal Energy Regulatory Commission’s September 21, 2006 Market Redesign and Technology Upgrade Order. CB is a financial mechanism that allows market participants, including electricity suppliers, consumers and virtual traders, to arbitrage price differences between the day-ahead (DA) market and the real-time (RT) market without physically consuming or producing energy. In this paper, market efficiency is defined in terms of trading profitability, where a zero-profit competitive equilibrium implies market efficiency (Jensen in, J Financial Econ 6(2):95–101, 1978). We analyze market data in the CAISO electric power markets, and empirically test for market efficiency by assessing the performance of trading strategies from the perspective of virtual traders. By viewing DA–RT spreads as payoffs from a basket of correlated assets, we can formulate a chance constrained portfolio selection problem, where the chance constraint takes two different forms as a value-at-risk constraint and a conditional value-at-risk constraint, to find the optimal trading strategy. A hidden Markov model (HMM) is further proposed to capture the presence of the time-varying forward premium. This is meant to be a contribution to the modeling of regime shifts in the electricity forward premium with unobservable states. Our backtesting results cast doubt on the efficiency of the CAISO electric power markets, as the trading strategy generates consistent profits after the introduction of CB, even in the presence of transaction costs. Nevertheless, by comparing with the performance before the introduction of CB, we find that the profitability decreases significantly, which enables us to identify the efficiency gain brought about by CB. Convincing evidence for the improvement of market efficiency in the presence of CB is further provided by the test for the Bessembinder and Lemmon (J Finance 57(3):1347–1382, 2002) model.

Economics: Latest FERC Outlook Showed Adequate Capacity, Enhanced Transmission, Higher Prices

AbstractElectric and natural gas market conditions this summer will reflect exceptional gas burn and storage withdrawals this winter, and the anticipation of a warmer‐than‐normal summer for much of the country. The possibility of a summer El Niño may moderate temperatures and associated market impacts. California Independent System Operator (CAISO) markets face a deep drought, while market expansions in the Southwest Power Pool (SPP) and Midcontinent Independent System Operator (MISO) enter their first summer period of operation.

Do Day-Ahead Electricity Prices Reflect Economic Fundamentals? Evidence from the California ISO

This paper hypothesizes that if day-ahead markets for electricity are efficient, then the day-ahead prices will reflect the processed information and expectations of all market participants regarding the next day’s electricity load and thus the prices may be useful in actually predicting the next day’s load. We test this hypothesis using data for the PG&E aggregation area in the California ISO. The results provide evidence of a positive and significant relationship between the hourly day-ahead electricity price (relative to the natural gas price) and the subsequent actual hourly load. The reported relationship is sufficiently robust to produce a forecast based on the day-ahead hourly price relative to the price of natural gas, some binary variables, and a number of estimated ARMA disturbances that is considerably more accurate than the ISO’s day-ahead forecast.

Electricity-market price and nuclear power plant shutdown: Evidence from California

Japan׳s Fukushima nuclear disaster, triggered by the March 11, 2011 earthquake, has led to calls for shutting down existing nuclear plants. To maintain resource adequacy for a grid׳s reliable operation, one option is to expand conventional generation, whose marginal unit is typically fueled by natural-gas. Two timely and relevant questions thus arise for a deregulated wholesale electricity market: (1) what is the likely price increase due to a nuclear plant shutdown? and (2) what can be done to mitigate the price increase? To answer these questions, we perform a regression analysis of a large sample of hourly real-time electricity-market price data from the California Independent System Operator (CAISO) for the 33-month sample period of April 2010–December 2012. Our analysis indicates that the 2013 shutdown of the state׳s San Onofre plant raised the CAISO real-time hourly market prices by $6/MWH to $9/MWH, and that the price increases could have been offset by a combination of demand reduction, increasing solar generation, and increasing wind generation.