Comprehensive Solutions for Integration of Solar Resources into Grid Operations

-- This project is inactive --

This project primarily looks at the benefits from more cost-effective unit commitment and dispatch, and reduction in balancing reserves due to reducing uncertainty in solar forecasting. This project will improve the Pacific Northwest National Laboratory’s ramp and uncertainty prediction tool by incorporating accurate forecasting of solar generation, and then integrate the tool with the Siemens market applications software currently used by the California Independent System Operator (CAISO) to perform unit commitment and dispatch. This project will utilize probabilistic forecast algorithms for solar energy production of large-scale PV plants and rooftop PV installations, to enable CAISO to incorporate solar generation forecasts directly into their tools that perform power system operations, thus reducing the uncertainty and hence the costs of system integration of solar generation into the bulk power system.

This project addresses the uncertainty problem comprehensively by including all types of uncertainties (such as load, variable generation, etc.) and all aspects of uncertainty including the ramping requirements. The main objective is to provide rapid (every 5 min) look-ahead (up to 5 to 8 hours ahead) assessment of the resulting uncertainty ranges for the balancing effort in terms of the required capacity, ramping capability, and ramp duration. The uncertainty range is called a “performance envelope” in this work. A methodology for self-validation of the predicted performance envelope has been developed.
This project will capture the uncertainty and variability present in power forecasting, load, and other variables that challenge the flexibility of grid balancing:
- Forecasting algorithms aiming to reduce the uncertainty range rather than just the standard deviation or root mean square value of the forecast error. The difference between the old and the new objectives becomes evident when dealing with nonparametric distributions and non-stationary distributions. The size of uncertainty interval corresponding to a certain level of confidence directly influences the balancing requirements.
- Forecast geographically distributed PV generation mixed with local loads and other local generation resources.
- Provide for concurrent consideration of all sources of uncertainty and variability (solar generation, system load, uninstructed deviations of conventional units, and forced outages).The concurrent consideration will help to reduce the overall uncertainty and, consequently, the resulting balancing effort required from the grid operators.
- Apply geographically and temporally distributed forecast models help to quantify the collective impacts of all sources of renewable generation uncertainty in their interaction on the transmission system. Proactive minimization of uncertainty intervals by separating more predictable, slower quasi-deterministic components from less predictable, faster components in the forecast errors.
- Quantify uncertainty and variability on the transmission system based on the risks of transmission problems (overloads, voltage problems) in various contingencies.
- First-ever industrial implementation of close loop real-time uncertainty-based unit commitment and dispatch procedures helping to exercise preventive (rather than corrective) control and avoid accidents, when the system balancing capacity and ramping capability is not sufficient to address random deviations of the resulting system load.
- Quantify of the system balancing requirements based on the new NERC control performance standard (BAAL) and new scheduling requirements recently introduced into the grid control practices.
- Incorporate ramping information provided by forecasts in the unit commitment and dispatch processes. Develop multi-dimensional uncertainty quantification procedures including a concurrent consideration of the capacity, ramp rate, and ramp duration requirements to the balancing generators.