August, 2013

Public Service Company of New Mexico and its partners are co-locating a 2.8 MWh advanced lead acid battery with a separately installed 500kW solar photovoltaic (PV) plant to create a dispatchable distributed generation resource. This hybrid resource provides simultaneous voltage smoothing and peak shifting through advanced control algorithms and and has the capacity to switch between two configurations, end-of-feeder and beginning-of- feeder. Data collection and analysis will produce information for a wide range of applications including grid upgrade deferral. The project also yields modeling tools used to optimize battery system control algorithms and further the understanding of feeders with storage and distributed generation. The site is located in southeast Albuquerque and is easily accessible for viewing by the public.

  • DOE: What does your team consider its biggest successes with the project so far?

    We consider the design and construction of the battery storage system to be one of our biggest successes. In order for this project to work, we needed a battery storage system that is flexible and can address multiple value streams using the same physical asset. This project is different from other energy storage developments because it is comprised of mostly off the shelf components using both internal and partner expertise. As a result, our system has proven capabilities of both smoothing and peak shifting, and can perform these functions simultaneously such as firming PV during cloudy periods. This system has consistently demonstrated simultaneous PV smoothing and shifting for months.

  • DOE: What does your team consider the biggest challenges?

    One of the ongoing challenges is how to derive the economic benefit of smoothing the solar farm output with the battery. Defining the optimal algorithm, optimal gains, and associated settings requires a large amount of statistical analysis. It also requires analysis to optimize battery life and the amount of smoothing delivered, which is still a vague concept. Further, the optimized storage system must be compared to existing solutions for high penetration PV installations such as capacitor banks, load tap changers (LTC’s) on transformers, and voltage regulators. The introduction of two way flow of energy as a result of PV on the grid is new for systems that were designed for one way power flow and there is not precedent for this analysis.  Because the use of storage is relatively new in electric grid applications, the industry is still grappling with how to quantify payments for this type of power.

  • DOE: What surprised your team the most about this project?

    We were surprised at the granularity of the data we received from our control system. We are operating a very sophisticated and sensitive data acquisition system that collects data in 1 second intervals. In fact, we can see the shadows of contrails from jets flying overhead on the PV system output! Furthermore, when we pushed on the data acquisition system capabilities we were able to measure control signal speeds, from one device to another, within a mili-second timeframe. When tuning a PV smoothing battery, it turns out that you really need this high resolution capability.

  • DOE: What advice would your team give to someone starting a similar project?

    We would recommend a thorough risk analysis on the front end. The risk analysis should identify specific risks in technological, resource, and operational categories; align these risks to types of impacts (schedule, cost, objectives); assign level of impact and probability of impact; and allow for mitigation strategies. The risk analysis should be continually refreshed. In addition, a thorough analysis of the requirements allowed us to field a sophisticated, integrated control system. Without a thorough understanding of all the systems and “actors” involved, we probably would still be trying to get the controls working.

  • DOE: Was there an “ah-ha!” moment during the project where you learned something unexpected?

    There were many and most of them came while brainstorming together--writing and diagraming on a white board and some major concepts clicked. For example, early in the design process it became obvious how we could send shifting control signals; we could send a simple +/- signal via a single PI tag from a back office control system that made decisions from a variety of inputs rather than sending the inputs to the field and making shifting control decisions there.

  • DOE: Is there anything else DOE should know about your project?

    This project would make a good platform for continued demonstration of forecasting and further refinements in storage control development. It has a proven and operable system, sophisticated data acquisition capabilities, and the ability to port high volumes of 1 second data securely to project partners. Our findings from this project could also help with integrating concentrated solar where ramp rates are harsher and present more rapid changes in output than PV. We made sure that the concepts we were using are scalable and transferable across other types of variable generation, and that the algorithms we are developing would translate across different types of applications for storage.