Event-Driven-WiFi-HIL-MPPT-Data

This dataset contains the measurement data, analysis scripts and ESP32 firmware that support the article "Event-Driven WiFi Hardware-in-the-Loop Co-Simulation for MPPT Validation: Quantifying the Effect of Network Latency and Packet Loss on a Low-Cost ESP32" (J. E. Zabala Daza, J. A. Gómez Pemberty, J. Mena Palomeque; submitted to Results in Engineering). In the platform, a Python model of a photovoltaic panel and a switched boost converter exchanges measurements and voltage references over HTTP with MPPT firmware running on an ESP32. The data quantify how the latency and packet loss of a commodity WiFi channel change the tracking efficiency measured in such a closed loop. All data were recorded on 23–24 September 2026 with a firmware build that reports, with every HTTP exchange, the on-chip algorithm execution time, the local turnaround time and the free heap. Contents Closed-loop factorial campaign (data/factorial/): 90 runs of a 2×3×3 design, crossing two MPPT algorithms (P&O and a two-phase hybrid PSO), three channel conditions (local software-in-the-loop, Wokwi ESP32 emulator, physical ESP32 over a 2.4 GHz WiFi LAN) and three 10 s irradiance profiles, with five repetitions per cell. It also includes 30 runs of a control condition with a fixed 28 ms delay and no jitter or loss. Each run provides a per-exchange log (RTT, algorithm time, turnaround, heap, inputs and returned reference), the plant trace at 0.5 ms resolution and summary metrics (tracking efficiency, settling time, ripple). Aggregates, statistical tests and the list of discarded runs, with the exclusion criterion, are included. Delay-injection sweep (data/sensitivity_rtt/): 30 closed-loop runs on the physical ESP32 with an additional delay of 50–500 ms. RTT probes (data/rtt_probe/): 1000 open-loop requests per algorithm on the emulator and on the physical board, with an identical input sequence, including a check that the firmware returns the same references as the reference Python implementation. Results (results/): RTT statistics and Monte Carlo packet-loss results (0–20 % loss, 1800 simulations). Code (scripts/, figs/, firmware/): the campaign runner and Python port of the firmware, the analysis and figure scripts, a script that checks every table of the manuscript against the data, and the ESP32 firmware (PlatformIO; WiFi credentials not included). Previous version (legacy_2026-06/): data of an earlier campaign (June 2026, record 10.5281/zenodo.20751824), superseded and kept for traceability.