The International Energy Agency (IEA) defines firm power generation as the ability for an ensemble of generating resource to meet a given electrical load 24x365. A growing body of work undertaken under the aegis of the IEA demonstrates that PV and wind can cost-effectively meet this firm power criterion if operating with optimal amounts of battery storage (BESS) and overbuilding (aka implicit storage). The firm power cost-effectiveness of these variable renewable (VRE) resources can be further enhanced with optimal PV/wind blending and by including a small fraction of dispatchable thermal generation operated with clean (albeit expensive) GHG-free e-fuels in the generation mix. The IEA results also suggest that, almost anywhere on the planet, cost-effective firm VRE solutions could be achieved locally without requiring major transmission build-up over large distances (to both capture high VRE resource regions and mitigate their variability.) In this article, we systematically quantify this assertion for the continental US (CONUS) by analyzing how the cost of firm PV/wind generation evolves as a function of the renewable generating footprint from a single point to a subcontinental scale. Results indicate that self-contained localized 100% renewable solutions can be viable for footprints of less than 50,000 km2 (i.e., the size of a small US state).

Forecasts of PV energy generation are used by generators to offer supply to the energy market. Generators face deviation costs for over/under forecasting energy production. We derived a forecast that minimized the expected deviation cost given a probabilistic forecast of energy generation: the minimized expected cost (MEC) forecast. We evaluated the average deviation cost of the MEC forecast against five deterministic forecasts: a persistence forecast, an NWP derived forecast, the mean of the probabilistic forecast, a “rule of thumb” under-forecast of the mean, and a gradient boosted trees regression model forecast. We find the MEC forecast to have the lowest average cost when evaluated against four hypothetical deviation cost functions. The MEC forecast reduced deviation costs by between 2% and 6% compared against the two most skilled alternate forecast methods, and by between 77% and 195% compared against the persistence and NWP derived reference forecasts.

A growing body of work demonstrates that firm wind/solar power generation capable of meeting current and future electric demand 24/365 can be affordable if enabled by effective regulations. The question we pose here is whether DPV hosting capacities would be increased if DPV systems actively participated in the larger grid's firm power generation objective. We show, based on 20 years of hourly wind/solar data and two Central US case studies, that this is indeed the case with the possibility of multifold DPV hosting capacity increases.

The increasing penetration of grid-connected solar photovoltaics (PV) has challenged the existing method of planning and operating the distribution grid. Electric utilities increasingly require visibility on the Behind- The-Meter (BTM) PV installations within their territories. They are interested in knowing the physical characteristics of the installed PV systems as well as their energy production profiles. This paper provides a novel method to infer the azimuth, tilt, and size of a PV system using only net load data and PV system location. Net load (total load minus PV production) represents the combined effect of consumption and PV system production. Our method utilizes Solar Anywhere® to produce plane of array irradiance for various azimuth and tilt combinations which serves as the ground truth data. The optimal PV system specification for each individual PV system is inferred by matching characteristics derived from net load data to the ones derived from the ground truth data. The proposed method is demonstrated on a test site in California and then implemented for all residential PV systems in a utility’ service territory. All locations have both net load and PV production data. We use the net load data to infer the system specifications and then the PV production data to validate the method’ accuracy.

Overbuilding and dynamic curtailment are increasingly acknowledged as central to cost-optimally transforming intermittent PV and wind resources into firm power resources. While this strategy is not currently monetizable, firm power generation will be a prerequisite at ultra-high renewable penetration when demand will have to be met 24/365 without reliance on underlying dispatchable generation. A distinct overbuilding/curtailment strategy is increasingly implemented today: inverter-limited curtailment. This strategy can take advantage of some existing remuneration systems. In this article, we compare the effectiveness of the two strategies to deliver firm power generation at least cost. We consider the extreme case of PV meeting demand with 100% certainty using two MISO's load balancing areas (#4 and #10) as experimental support. We show that, while both strategies can achieve firm power generation at a lower cost than curtailment avoidance would, dynamic curtailment is far more cost-effective than inverter-limited curtailment. Importantly, we also show that optimally combining both strategies can further reduce firm power generation cost.

While it is widely known that the solar resource is sufficient to meet the world's energy demand many times over, the questions of where and how much to deploy in a realistic context do not have such clear-cut answers. The objective of this paper is to address and inform these questions in a context where solar (embodied by PV) would be applied locally to firmly meet the bulk of energy demand from regional economies. Sensible answers are important in light of growing societal mandates to displace carbon-based energy resources. We aim to provide realistic and comprehensive numbers that can effectively inform planning decisions at local and regional levels. We focus on the continental United States (CONUS) and develop state-specific PV deployment requirements informed by: A full accounting of states' energy requirements from the electric sector as well as [to be] electrified transportation and building (HVAC) sectors. Positing that the bulk of this demand will be met within each state with an optimized blend of PV and wind with a small residual allowance for dispatchable generation - an optimized blend of resources estimated from our recent grid-specific investigations in diverse climatic and socioeconomic environments. A recognition that electrical demand must be met firmly, hence that intermittent renewables must be transformed into firm, effectively dispatchable resources available 24/365 to maintain a stable electrical grid. A recognition that the least-cost solution to achieve this intermittent-to-firm transformation implies overbuilding and proactively curtailing these resources we apply herein an estimated overbuild amount estimated from our recent investigations in diverse environments. Not accounting for likely energy efficiency improvements in any of the three considered demand sectors. Therefore, the numbers developed can be considered to be conservatively high. From these requirements, we explore PV deployment options using two distinct approaches: a top down approach assigning a fraction of plausible deployment potential to land use classes as defined by the US geological Survey, and a bottom-up approach starting from end-use applications prospectively amenable to PV deployment without change of function. In addition, we provide readers with an online interactive capability to modify fractional land use selections applied in this article and further investigate state-specific potentials. We conclude that a majority of the three-sector firm power requirements could be met economically and firmly by locally-deployed PV resources with ample deployment room to grow, even in the most densely populated northeastern states. This conclusion applies even before considering energy efficiency improvements or tapping other renewable resources that may be available locally (e.g., hydropower).

We introduce firm solar forecasts as a strategy to operate optimally overbuilt solar power plants in conjunction with optimally sized storage systems so as to make up for any power prediction errors, and hence entirely remove load balancing uncertainty emanating from grid-connected solar fleets. A central part of this strategy is the plant overbuilding that we term implicit storage. We show that strategy, while economically justifiable on its own account, is an effective entry step to achieving least-cost ultra-high solar penetration where firm power generation will be a prerequisite. We demonstrate that in the absence of an implicit storage strategy, ultra-high solar penetration would be vastly more expensive. Using the New York Independent System Operator (NYISO) as a case study, we determine current and future costs of firm forecasts for a comprehensive set of scenarios in each ISO electrical region, comparing centralized vs. decentralized production and assessing load flexibility’s impact. We simulate the growth of the strategy from firm forecast to firm power generation. We conclude that ultra-high solar penetration enabled by the present strategy, whereby solar would firmly supply the entire NYISO load, could be achieved locally at electricity production costs comparable to current NYISO wholesale market prices.

We present the perfect forecast concept as both an effective forecast validation metric and an operational strategy to integrate increasing amounts of variable solar power generation on power grids. The costs incurred in transforming imperfect into perfect predictions define the new metric: these include the costs of backup storage and output curtailment necessary to make-up for any over/under predictions. We illustrate the concept with the most recent version of the SUNY forecast model for hour-ahead and day-ahead forecast examples with single power plants as well as distributed PV fleets. We show that delivering perfect predictions – i.e., fully eliminating grid-operators uncertainty -is achievable at small operational cost. Most importantly, we show that a perfect forecast strategy with optimized least-cost storage and overbuild/curtailment is an effective first step of a longterm strategy to cost-optimally transform variable PV generation into firm, effectively dispatchable generation capable of displacing conventional dispatchable and baseload generation. Key-words: solar resource, irradiance, forecast, storage, high-penetration, firm power generation

The SUNY solar irradiance forecast model is implemented in the SolarAnywhere platform. In this article, we evaluate its latest version and present a fully independent validation for climatically distinct individual US locations as well as one extended region. In addition to standard performance metrics such as mean absolute error or forecast skill, we apply a new operational metric that quantifies the lowest cost of operationally achieving perfect forecasts. This cost represents the amount of solar production curtailment and backup storage necessary to correct all over/under-prediction situations. This perfect forecast metric applies a recently developed algorithm to optimally transform intermittent renewable power generation into firm power generation with the optimal - least-cost – amount of curtailment and energy storage. We discuss how perfect forecast logistics can gradually evolve and scale up into firm solar power generation logistics, with the objective of cost-optimally displacing conventional [dispatchable] power generation.

This article introduces a new version of the SUNY solar forecast model, as implemented in the software SolarAnywhere®. Like the existing version, this new version is intended for direct, out-of-the-box application throughout North America without requiring training/feedback from measured data. The existing version was recently identified by EPRI as most accurate among thirteen operational models after an independent evaluation in two climatically distinct US regions. This new version shows further measurable performance improvements and operational functionality with capability of using historical satellite data for model training purposes. The advantage of incorporating historical satellite data is that it allows this forecast application to scale to all sizes of PV systems (large utility-scale down to individual rooftop) without the requirement of site-measured inputs. This new approach exhibits a 1-5% root mean square error (RMSE) improvement over the forecast horizon up to two-days in advance. Index Terms —solar resource, irradiance, solar forecasting, PV fleet forecasting.

In this article we show how a state-of-the-art satellite irradiance model - SolarAnywhere - is capable of undetected calibration issues at a trusted reference ground truth irradiance measurement station. This evidence suggests that the best satellite models have now achieved a degree of accuracy and versatility that makes them an acceptable, if not a preferred choice, for solar energy engineering applications ranging from long-term site characterization and system monitoring.

In this article, we present a state-of-the-art longtern energy prediction methodology to minimize risk in the project investment. We applied the algorithm to the SolarAnywhere data over 19 years period. he algorithm systematically creates synthetic years from the period of dataset to draw distribution that best describes the data. We conducted the experiment for projects under different climatic conditions. We compared the result with several theoretical distribution models. The results show that the method is so robust that it can predict the real production data at a site with high level of accuracy and provide confidence in the investment.

This article summarizes and analyzes recent research by the authors and others to understand, characterize and model solar resource variability. This research shows that understanding solar energy variability requires a definition of the temporal and spatial context for which variability is assessed; and describes a predictable, quantifiable variability-smoothing space-time continuum from a single point to 1000’s of km and from seconds to days. Implications for solar penetration on the power grid and variability mitigation strategies are discussed.

We approach the issue of short-term PV output intermittency from a management standpoint by determining the cost of actively mitigating it using “shock-absorbing” short-term energy buffers. Using a case study in Central California as experimental support, we determine this cost of as a function of (1) the amount of variability mitigation; (2) the considered variability time scale, (3) the PV resource's geographical footprint, and (4) the availability of accurate solar forecasts. We show that, in a plausible operational context, the cost of mitigating variability across time scales ranging from one minute to a couple of hours could be kept below 25-35 cents per installed PV kW.

Copyright © 2013 by the American Solar Energy Society Inc. All rights reserved. P hotovoltaic (PV) systems still appear to be expensive when compared without context to traditional power generation, despite immense progress over the last few years. Constituents, however, generally believe that solar energy delivers a higher value than can be monetized in a business-as-usual setting —the values that are often unaccounted for include environmental value, fuel depletion and price mitigation value, market price reduction, economic development, jobs, energy security and value linked to displacing conventional resources’ embedded incentives (e.g., see figure 1, page 20). This understanding is the reason why cities, states, provinces and countries around the world have developed financial-transfer mechanisms in an attempt to level the playing field and make up for the part of the value delivered by solar generators that is not currently monetized. These financial-transfer mechanisms are typically referred to as “incentives.” However, as we’ll discuss, the term incentive does not have to imply subsidy. Incentives/financial-transfer mechanisms have taken many forms. These include buy-down grants, solar renewable energy credits (SRECs), reverse auctions, net metering and feed-in-tariffs (FiTs), as well as income tax credits (ITC), tax abatements, tax exemptions, low-cost financing and so on (e.g., see DSIRE, 2012), which can either be tax-financed and/or utility ratepayerfinanced. In the United States, the ratepayerbased transfers of value are generally driven by renewable portfolio standards (RPS), whereby a renewable deployment goal is specified by the law and implemented by forcing utilities and grid operators to purchase renewable energy credits from renewable energy producers. We propose a modified FiT, or Smart FiT. This modification aims to take the best elements of the most effective solar energy compensation system to date, but removes some its weaknesses. In particular, the proposed Smart FiT links the tariff to value produced and includes long-term market controls leading to very high penetration.

We approach the issue of short-term PV output intermittency from a management standpoint by determining the cost of actively mitigating it using “shock-absorbing” short-term energy buffers. Using three case studies in California, Hawaii and the southern US as experimental support, we determine this cost as a function of (1) the desired amount of variability mitigation; (2) the considered variability time scale, (3) the PV resource’s geographical footprint, and (4) the availability of accurate solar forecasts. We show that, in a plausible operational context, the cost of mitigating variability across time scales ranging from one minute to a couple of hours could be kept below 25-35 cents per installed PV kW.