We're modeling a solar array for a building which will have a demand charge as part of its electricity tariff. Since demand is often calculated on a month basis, and the cloud cover / solar impact can vary dramatically during any month, we suspect that any demand savings shown by our model will be overstated - perhaps by 80-100%.

Have you considered this? Determined a method for evaluating the demand savings when a solar array is in use?

Thanks for your insight.