PV Fault Simulator

An interactive Shiny for Python application for simulating photovoltaic plant performance under fault conditions, soiling, and measurement noise. Built on pvlib and the Sandia Array Performance Model (SAPM).

---

Features

Tab	Description
1 · Config & Layout	Configure equipment, site location, and array layout; visualize the field layout colored by inverter, combiner, tracker, or fault status
2 · Faults	Add, edit, and delete tracker stalls and inverter outages using a click-to-select table; apply an HSU soiling profile from air quality data
3 · Weather	Inspect the weather input file (irradiance, temperature, wind)
4 · Simulation	Run the SAPM string-level power simulation; explore time series and annual energy distributions
5 · Noise	Generate spatiotemporal measurement noise (slow/medium/fast components) and overlay it on simulation results
6 · SCADA	Aggregate string results to inverter / combiner / tracker SCADA signals; browse, plot, and download CSVs

---

Requirements

• Python 3.10+
• pvlib
• Shiny for Python
• numpy, pandas, matplotlib, scipy

Install dependencies:

pip install pvlib shiny numpy pandas matplotlib scipy

---

Running the App

shiny run app.py

Then open the URL printed in the terminal (typically http://127.0.0.1:8000).

---

File Structure

pv_fault_simulator/
├── app.py                        # Main Shiny application
├── config.py                     # Plant configuration (equipment, layout, site, faults)
├── main.py                       # Equipment loading and array construction
├── plant_design_functions.py     # SAPM simulation and SCADA signal generation
├── noise_generator.py            # Spatiotemporal noise generation
├── hsu_soiling.py                # HSU soiling model reference code
├── combined_air_quality_2021.csv # Hourly PM2.5, PM10, and rainfall data
├── weather_5m.csv                # 5-minute weather data (GHI, DNI, DHI, temp, wind)
├── SNL_logo.jpeg                 # Navbar logo
└── Baseline_SAPM.ipynb           # Original analysis notebook

---

Configuration

Edit config.py to set default equipment, layout, site coordinates, and fault scenarios — or use the 1 · Config & Layout tab and click Save to config.py.

Key settings:

WEATHER_FILE = "weather_5m.csv"
MODULE_NAME  = "Sandia module name"
INVERTER_NAME = "Sandia inverter name"

ARRAY_DICT = {
    "number_inverters": 2,
    "combiners_per_inverter": 4,
    "trackers_per_combiner": 4,
    "strings_per_tracker": 2,
    ...
}

SITE_DICT = {"latitude": 35.05, "longitude": -106.54, "altitude": 1500}

TRACKER_STALLS = [
    {"inverter_number": 1, "combiner_number": 2, "tracker_number": 3,
     "start_time": "2021-06-01 08:00:00 -07:00", "duration_hrs": 4},
]

INVERTER_OUTAGES = [
    {"inverter_number": 1, "start_time": "2021-01-15 00:00:00 -07:00", "duration_hrs": 24},
]

---

Soiling Model

The app uses the pvlib HSU soiling model. Provide a combined air quality CSV with columns for datetime, PM2.5, PM10, and rainfall:

TmStamp,PM2_5,PM10,Rainfall_mm
2021-01-01 00:00:00-07:00,12.3,28.1,0.0
...

The Edit Air Quality Data panel lets you scale PM values or insert rain events over any date range to simulate soiling scenarios without modifying the source file.

---

Noise Model

Measurement noise is generated as a superposition of three spatiotemporal components:

• Slow — long spatial and temporal correlation (soiling gradients, degradation)
• Medium — moderate correlation (cloud shadows, temperature gradients)
• Fast — uncorrelated white noise (sensor measurement error)

Each component is generated on a spatial grid via FFT-based filtering and interpolated to string locations.

---

Workflow

1. Config & Layout — set up your plant and verify the field layout
2. Faults — define tracker stalls, inverter outages, and/or a soiling profile
3. Weather — confirm the weather file looks correct
4. Simulation — run the SAPM; re-run whenever faults change (staleness warning shown)
5. Noise — generate and overlay measurement noise on simulation results
6. SCADA — generate aggregated signals and download CSVs for further analysis