Commercial PV System Data Monitoring, Part Two: Page 4 of 6

Conduits and circuit routing. The addition of a monitoring system can add a significant amount of conduit and circuits that must be planned and accounted for in the initial design. Verify that placing the cables in conduit does not exceed the conduit-fill allowance. This is especially true for weather sensors that have unique connectors.

When evaluating the Internet connection, keep in mind the distance limitations of the various protocols and bus drivers. TCP communication on a CAT 5 or 5e cable is limited to 300 feet. If the distance from the base station to the Internet connection exceeds this distance, consider using fiber optic cables, which have distance allowances that are counted in miles, or cellular modem connections if within range of a cellular signal.

To keep costs down in large array fields or custom structures, identify all of your conduit requirements before hardscapes like roads or parking lots are completed. This lets you get your conduit into trenches before they are backfilled and is much more cost effective than finding alternative installation solutions afterwards.


Since financials are the biggest driver in determining monitoring decisions, the following analysis attempts to demonstrate the financial effects of varying degrees of system downtime. Washington, DC, was chosen for the location due to its mature and developed solar market.

As of June 2011, the SREC value in this market with a 3-year term is averaged at $365/MWh or $0.365/kWh. In addition, as of February 2011, the average commercial electrical offset rate for this area is $0.1358/kWh, according to the US Energy Information Administration. Therefore, the total value to the investor of the PV-generated energy in this market is $0.5008/kWh.

The example system consists of 500 kW of 14% efficient modules, roof-mounted at a 10° tilt and a 180° orientation and organized into 152 strings that feed two 250 kW inverters. Using PVsyst, we can predict an average annual energy production of approximately 697,000 kWh.

Table 1 demonstrates the effects of different levels of system failures over typical durations of time. The effects are quantified in total energy production loss, percentage of annual production loss, and the total and daily financial repercussions.

These scenarios are not uncommon. As the annual production loss calculation indicates, they can have a significant impact on the performance ratio of a system. Time is money in these cases, and the integration of a data acquisition system can significantly increase system uptime.

It is not an exaggeration to estimate a 2-week time period to get an inverter back online. In fact, without data monitoring, it could be up to a month or longer before a problem is detected. Fixing the problem after it is identified typically involves time for travel, troubleshooting, shipping replacement parts, yet more travel, and replacing the component and recommissioning the subarray.

With inverter-level monitoring in place, an alarm and accompanying error code is received instantly via web portal, text message or email as soon as the inverter failure occurs. This allows for a more timely and focused response. In many cases, good inverter communication allows problems to be resolved on the first trip to a site, instead of merely being diagnosed.

Losing the input from a single string into a 250 kW inverter would not be noticeable at the system or inverter level. Without string or combiner box monitoring, it could be many months before the problem is noticed. By then, a lot of potential energy would have been lost. The loss of a single string could eventually lead to bigger problems, depending upon the causality. (See “The Bakersfield Fire,” February/March 2011 SolarPro magazine.)

Although a single string represents a small fraction of total system production, over time the value of the lost energy adds up. In the example system, using a $/kWh escalator of 4% and a discount rate of 8%, the net present value of having that string functioning properly over 25 years is around $35,000. The internal rate of return for the system also shows a drop from 17.5% to 17%. This drop alone might justify the up-front cost of providing a more sophisticated monitoring system.

For comparison purposes, a system in Oakland, California, would have a performance-based incentive value of $0.05/ kWh for 5 years and an average electricity cost offset of $0.1723/kWh, as of March 2011, for a total value of $0.2223/kWh. Based on this lower value of energy and the relatively higher insolation in Oakland compared to Washington, DC, the economic value of the production losses in Table 1 would decrease by about 50%, even though the amount of energy lost would increase by about 10%.

Article Discussion

Related Articles