a large pv installation at fairfield university

A three-year research project using a PV installation in an extensive, real-life, residential environment is now under way at Fairfield University. Design and development studies will be conducted using a grid-connected PV installation with a set of 880 solar roof shingles covering a roof area of approximately 3,000 sq.ft. in a townhouse with eight student apartments. There are two merit elements of note in this project, namely the size of the generating plant, estimated at 14.5 kW, and the fact that it is grid-connected and integrated into a living residential environment, providing power for the daily use of most electric devices in this environment. Interfacing a “new” power-generating system with daily residential activity, will allow us to learn the high points and pitfalls in the traditional components of a PV installation, and guide us to design more efficient systems. Engineering studies will be augmented by studies on the ecological and financial aspects of the project.

To maximize the opportunity for experimentation and data collection, the solar roof array, recently completed, has been segmented into three sub-arrays, each one with its own controls, and each one feeding power to a different set of student apartments.

Background and Objectives

The present PV project is in two phases.  Phase I has been funded by a $100,000 grant from the W. M. Keck Foundation awarded in January 2000, and a recent $10,000 grant from United Illuminating.  The power generator is a set of 880 solar roof shingles, each with a generating power of 17 Watts, estimated to provide a total of approximately 14.5 kW. This will probably be one of the largest grid-connected PV installations in use in a residential environment in the country. Phase I of the project includes the design of the PV system, the purchase and installation of the solar shingles in suitable arrays on the 3,000 sq.ft. roof of townhouse #10 on the Fairfield campus, and connecting the shingles to Combiner Boxes. These tasks are approaching completion.

Phase II includes (a) the acquisition and installation of the necessary electronic equipment, battery packs, and data loggers with connections to inverters and phone lines for data acquisition, management and analysis, and (b) engineering studies and development of efficient electronic controls to optimize the storage and use of PV array generated power in a real-life residential environment, augmented by economic and ecological analyses. The crucial objective is to provide credibility and value for roof PV arrays based on data and experience. Experimentation, monitoring and data gathering will be accomplished via an automatic data acquisition system on site. Statistically significant data analysis will aid in constructing quantitative models for production and efficient use of solar energy.

In addition to engineering design and development of the PV control system, studies in the economic and ecological ramifications of the installation are planned. The development team includes Engineering faculty, as well as faculty from the Department of Economics and the School of Business, representatives from the local electric power industry (United Illuminating) and the building industry (PS Designs). Data loggers in the PV installation will continuously monitor the system and accumulate data that will provide the basis for engineering development, as well as for the economic assessment of the system.

Technical issues

The solar shingles for the Fairfield PV installation are manufactured by the United Solar Systems Corp of Troy, Michigan. The core of the shingles is a triple junction amorphous silicon (a-Si) alloy solar cell, each one composed of three semiconductor junctions stacked on top of each other. Each shingle is rated at 17 Watts. The National Renewable Energy Laboratory has been evaluating these roofing modules in its Golden, Colorado, facility since 1993.

To facilitate engineering studies with the PV installation on the roof of townhouse 10 on the Fairfield campus, the solar array has been segmented into three sub-systems, each providing power to a different set of residential units, i.e., units 1 and 2, units 3,4, and 5, and units 6,7,and 8. Each subsystem has its own inverter, and each one can be easily bypassed using the inverter-bypass switch and/or DC and AC disconnect switches. 

The roof exposure in townhouse 10 is south-southwest, and it has a 270 tilt. This is about 150 less than the 420 latitude in Connecticut.  Under these circumstances, the average solar radiation for flat-plate collectors from month to month is as follows:

Average Solar Radiation for flat-plate facing South at a fixed tilt, in kWh/m2/day

Tilt

Jan

Feb

Mar

Apr

May

June

July

Aug

Sept

Oct

Nov

Dec

Av./  year

-150 from Lat.

2.9

3.7

4.4

5.1

5.5

5.8

5.8

5.5

4.9

4.0

2.8

2.4

4.4

Only the first PV sub-system, providing power to residential units 1 and 2, includes a battery pack for storing energy. This system utilizes an SW5548 Trace inverter, while the other two sub-systems include Omnion inverters. In the present configuration, the losses are estimated to be 53% in the system with the Trace inverter and batteries, and about 30% in the systems with inverters only and no battery packs. These losses will be verified when the system is operational and will be the focus of the planned engineering studies to reduce them appreciably.

The battery pack that will be used with the first sub-system consists of 24 2-Volt C&D CPV780 batteries.  Power from the solar arrays will be provided for all the appliances in the eight residential units except for the electric stove and air conditioner. Also, excess power from the array will be “net-metered”, i.e., fed into the existing grid(3).

University records show that the average monthly consumption of electric energy in each of the residential units is approximately 500 kWh.  With a year round average irradiance of 4.4 kWh/m2/day, as shown in the Table above, and taking into account the efficiency of the solar cells and initial estimated energy losses in the system, it is calculated that each residential unit will receive approximately 55-60% of the average consumed energy from the PV solar array. It is expected that the proposed engineering studies will measurably improve this number.

To facilitate data collection a CR10X Data logger from Campbell Scientific will be utilized.