Weather IoT | Anticipate a Change

We’re going to pin point this one on micro grids, but the same principles apply to any outdoor activity. Concepts are simple; knowing what weather is coming changes your approach. In microgrids nearly everything that causes instability comes from the skies. The same factor applies to production, supply and demand curves. As much of the European energy market is deregulated, there is often a peak price point when weather turns “bad” (cold)  or too “good” (hot). That supply and demand curve is a known quantity.

Without local weather data you’re driving by the rear view mirror.

If you know a cloud bank is arriving, you can pre charge batteries. When wind shear increases, run the diesel slower. Expecting a heat spike at 4pm, anticipate an AC surge. You’re getting the key data points in real-time.

Grid balancing kills batteries when it’s blind. Aggressive charge and discharge, deep discharge and thermal stress damages them. Knowing the incoming weather allows you to smooth out those spikes. As a rule of thumb; good forecasting can offer 15% + extra battery life.

Microgrids often rely on diesel, gas or CHP (combined heat and power) as part of the system. Weather data allows for a more steady operating band and longer service intervals with fewer “bang” events. Blackouts are rarely due to big failures, often it is an inverter trip, protective shutdown or brownout (lowering, rather than complete stop).

Using public weather data is useful, it gives you a good overall report. Using on-site weather in real-time gives actual irradiance, real cloud edge detection, local dust and humidity levels and local wind and temperature. This moves your accuracy window much closer to the actual. If you’re running any kind of automation stack this will prove decidedly useful.

So, what kind of sensors make sense and which ones do not matter that much? A good start is GHI (Global horizontal irradiance). They measure real solar input. One pyranometer beats ten dashboards. Ambient temperature is a bit boring, but you can’t predict output or demand without it. They’re also cheap. Wind speed is another one; wind power failures usually come from variability, not from the lack of wind.

Humidity, specifically relative humidity is a handy one. It helps with fog, haze, dew, and usually outputs correlate with demand spikes. Situational sensors like barometric pressure are handy in coastal regions. In really warm regions panel heat sensors help with short term forecasting. Rain gauges tell you stuff that you may already know if you’re on-site, but if you’re remotely running a project they can add a lot of value.

As a simple rule; install only the ones you need, and whose failure you would immediately notice. If nobody cares, the numbers provide no value anyway.

You've got an IoT idea?

Our engineers have the tools.

Talk to us today