Ahlmahz I. Negash

Allocating the Cost of Demand Response Compensation in Wholesale Energy Markets

Currently, demand response resources can sell load reductions in wholesale energy markets. However, paying for load reductions ultimately results in an unbalanced market, where the amount of resources sold (megawatts) is less than the amount of resources bought (megawatts and “negawatts”). To resolve this imbalance, the ISO must allocate the cost of compensating demand response to those buyers who benefit from reduced LMPs. Current cost allocation methods are quite broad and based on each energy buyers share of the total load. In an uncongested network, this results in a “fair” allocation of costs, i.e., an allocation proportional to the benefits that each party accrues. However, in a congested network, this is no longer the case, as price separation occurs between nodes. In this paper, we therefore propose a cost allocation method based on LMP sensitivity that accounts for the effect of congestion on the distribution of benefits between nodes with different LMPs. Since this sensitivity-based method only takes into account the cost allocation per node, we also propose a means of allocating costs between individual load serving entities (LSEs) at a single node. Due to this refinement, LSEs are rewarded according to their individual contribution to demand response. Finally, we define a fairness index to evaluate the performance of the proposed method as compared to a load-based allocation. We find that when load reductions are small (1%-3% total load), the fairness index of the proposed method is very close to zero, indicating almost identical benefit to cost ratios for all market participants. Although the fairness index increases with increasing load reductions, results show that even for larger load reductions, the fairness index is still lower for the proposed method than for the load-based allocation method.

Multistate voltage dependent load model of a charging electric vehicle

Plug-in electric vehicles (PEVs) represent a large new load class for the distribution system. Though many intelligent charging methods have been proposed for PEVs, the models that have been previously used for simulating PEV charging are overly simplistic. In this work a multistate ZIP model of PEVs is discussed and formulated. A Nissan LEAF is also modeled in detail. Simulations show why the more detailed models are required for proper PEV charging impact analysis.

Voltage dependent load models of charging electric vehicles

Plug-in Electric Vehicle (PEV) integration on distribution grids is an important topic of research due to the potential problem involved. In order to accurately assess the impact of EVs on system voltages and power flows, accurate voltage dependent models must be used. In this work, the voltage response of four different EV types is measured and ZIP models are developed for each which can be used in EV integration studies. The measurements show that the behavior varies significantly between the different EVs and thus an accurate generic model cannot be made.

A wavelet-based method for high resolution multi-step PV generation forecasting

Forecasts play a vital role in maintaining power system stability and maximizing economic benefits of distributed energy resources. The issue with PV generation forecasting is that it relies on forecasts of solar irradiation which, due to the complex, nonlinear relationship between humidity, pressure, temperature, and cloud transients, can be quite difficult to model. Two important decisions in the forecasting process are selection of the forecasted variable (model output) and selection of explanatory variables (model inputs). This paper proposes a new method to forecast PV generation using wavelet based input selection and an output variable that directly represents clouds transients. We model this cloud effect by first determining a clear sky model (CSM) and forecasting the difference between the CSM and actual measurements of global horizontal irradiance (GHI). Potential model inputs are first decomposed using wavelet multi resolution analysis and final input selection is based on the correlation between the inputs and output at various timescales. Two separate neural network structures are designed to separately forecast sunny and cloudy days. Using the high resolution forecast of GHI (20 min increments), the next days PV generation is determined. This method improves on the persistence method by 69% on sunny days, 26% on cloudy days.

Combined optimal retail rate restructuring and value of solar tariff

Net Energy Metering (NEM) is the most common form of compensation for distributed PV and currently 46 states and the District of Columbia have NEM policies. This “easy to understand and implement” policy allows customers with distributed generation to be compensated at the retail electricity rate for any excess generation fed into the grid. However, NEM has recently come under strong criticism from consumer advocates and utilities alike for being unfair. The underlying issue is that current residential rates bundle generation, transmission and distribution costs into a single “per kilowatt-hour” charge. Thus, NEM customers provide generation but avoid generation, transmission and distribution costs. Because this can lead to unfair cost shifting, some are calling for the implementation of value of solar tariffs (VOST). Unlike NEM, a VOST credits PV based on the actual value it contributes to the utility. In this paper, we propose a methodology to co-optimize the retail rate structure with the VOST such that network, societal, and policy objectives are fully considered. We test the proposed method on a prototypical utility in Washington State. The resulting retail rate has minimal negative customer impact, while the VOST maintains the simplicity of NEM, but also ensures that PV is only credited for the benefits it provides.

Compensation of demand response in competitive wholesale markets vs. retail incentives

In 2011 the Federal Energy Regulatory Commission (FERC) issued a landmark ruling, FERC Order 745, standardizing the compensation of demand response (DR) in competitive wholesale markets. According to this order, demand response resources participating in competitive wholesale energy markets must, like generators, be paid full locational marginal price (LMP). Many economists opposed this ruling and argued that the most efficient method is to offer dynamic prices and naturally, demand reductions are rewarded with the avoided cost of the energy not used. One of the main arguments against the order is the fact that by paying LMP for demand reductions, the market collects less in revenue than it must pay out for resources, a phenomenon known as “the billing unit effect” and must therefore, allocate the shortfall. In this paper we compare wholesale DR compensation to retail level incentives. We define demand response as a short-term added cost for the load serving entity (LSE), voluntarily paid in order to save money over the long run. Based on this view of DR, we propose a benefit sharing incentive scheme at the retail level. This scheme involves the use of a publicly broadcast grid state index implemented by the California Independent System Operator (CAISO).

Allocating the cost of demand response compensation in wholesale energy markets

Summary form only given. Paying for load reductions results in a market, where the amount of resources sold is less than the amount of resources bought. To resolve this imbalance, ISOs must allocate the cost of compensating demand response to those who benefit from reduced LMPs. Current cost allocation methods are based on each energy buyers load share. In an uncongested network, this results in a “fair” allocation of costs, i.e. allocation proportional to the benefits each party accrues. However, in a congested network, this is no longer the case, as price separation occurs between nodes. In this paper, we propose a cost allocation based on LMP sensitivity that accounts for the effect of congestion. We also propose a means of allocating costs between load serving entities at a single node. Finally, we define a fairness index to evaluate the performance of the proposed method compared to load-based allocation. We find that when load reductions are small, the fairness index of the proposed method is very close to zero, indicating almost identical benefit to cost ratios for all market participants. Although the fairness index increases with increasing load reductions, even for larger load reductions, the fairness index is still lower for the proposed method.

Optimizing demand response price and quantity in wholesale markets

We propose a method to create a supply curve for demand response (DR) analogous to, but separate from, a generation supply curve. This new demand side “supply curve” is used to offer DR into the wholesale market as a demand resource, not a competitive generation source and therefore, unable to set locational marginal price (LMP). This DR supply curve is based on the value of increased gross margin due to load modifications. While this supply curve is based on optimizing a load serving entitys gross margin, the maximum level of DR allowed in the market is based on the amount of DR that brings the market balance to zero (i.e. wholesale revenue equals generator and DR expenses). The proposed method is compared to compensation according to FERC Order 745. We find that while compensation according to Order 745 always results in a negative balance (less revenue than expenses) and a need to allocate that cost, the proposed method achieves a positive balance for low levels of DR, zero balance at an optimal DR level, and a negative balance only if too much DR is purchased.