Solar Array Telemetry: Analog Smoothing, Moving Averages & Data Logging

FIG 1.1 // REMOTE SOLAR PV MICROGRID TELEMETRYSYSTEM: PV-STRING // 24VDC - 32VDC BUS SIGNAL CONDITIONING

The 4,000 False Alarms & The Weekend Database Crash

Imagine driving down a bumpy, rocky dirt road. If your car speedometer displayed the raw instant speed of each bounce every millisecond, the needle would violently blur between 20 MPH and 50 MPH! You wouldn't be able to read your speed, and your cruise control would violently jerk the throttle.

Modern cars solve this with digital smoothing—averaging recent readings to display a smooth, steady number.

In solar energy automation, sensors face the exact same problem from nature: passing clouds:

Understanding Analog Noise vs. Real Trends

When monitoring physical sensors (like solar voltage, wind speed, or temperature), readings contain two components:

• The True Engineering Trend: The actual, meaningful state of the process (e.g., the sun is shining brightly at midday).
• High-Frequency Noise: Temporary fluctuations caused by electromagnetic interference, cable induction, or passing cloud shadows.

If an automation system reacts to every millisecond of noise: Mechanical relays and contactors chatter violently, burning out contacts; databases get overwhelmed with useless junk; and false alarms wake up on-call technicians.

Stage 1: The Raw Unfiltered Comparator (Contactor Chattering)

Below is Stage 1 of our interactive lab. The raw solar voltage is wired directly into an ANALOG_COMPARATOR with a 24V threshold.

Try adjusting the slider in the simulation below:

Stage 2: Analog Smoothing with a Moving Average Filter

To solve signal chatter, we pass the raw analog signal through an AVERAGE block.

How a Moving Average Works: The block maintains a small FIFO memory buffer of recent measurements (e.g., 5 samples):

When a brief shadow or voltage spike occurs, its impact is diluted across the buffer. The resulting output is a clean, stable engineering value that reflects the true trend without jitter!

Stage 3: Full Microgrid Telemetry & Battery Charge Manager

In a production solar microgrid, the control logic must perform two essential tasks:

• Safe Window Charging: Batteries must only charge when voltage is within safe limits (e.g., between 24V and 32V). Below 24V, sunlight is insufficient; above 32V, over-voltage threatens to boil battery electrolyte!
• Periodic Telemetry Logging (DATA_LOG): Instead of logging continuous streams of data, the DATA_LOG block samples the clean, smoothed voltage at a steady, configured interval (e.g., every 500ms). This produces compact, orderly datasets ready for cloud transmission via MQTT or cellular IoT.

Commissioning & Field Calibration Checklist

When deploying analog telemetry or data loggers on a remote solar installation, follow this 5-point field checkout:

1. Verify Instrument Cable Shielding: Ensure the shielded drain wire from the voltage transducer (VT-101) is grounded at one end only (at the RTU cabinet) to avoid creating ground loops.
2. Tune the Averaging Window: Sizing the sample buffer is a balance: too few samples (e.g., 2) leaves jitter; too many samples (e.g., 50) introduces noticeable phase lag during genuine daylight transitions. 5 to 10 samples is ideal for solar PV strings.
3. Verify Logging Frequency: Confirm that data logging intervals match your network budget. On satellite or cellular IoT skids, sampling every 500ms to 5 seconds provides high fidelity without exceeding monthly data caps.
4. Calibrate Analog Span: Measure the physical string voltage with a calibrated Fluke multimeter. Ensure the scaled analog reading on the PLC display matches within ±0.1VDC.
5. Test Over-Voltage Cutoff: Inject a simulated test voltage of 33VDC into the analog channel and verify the battery charge contactor drops out immediately.

Key Engineering Takeaways