Event-triggered distributed STW-RBFNN control of PV-battery DC microgrids under uncertainty and delays
Voltage regulation in PV-battery DC microgrids is challenging because of nonlinear converter dynamics, uncertain operating conditions, changes in PV generation and load demand, and communication delays. This paper presents a distributed smooth super-twisting radial basis function neural network (STW-RBFNN) controller for DC-link voltage regulation. The RBFNN estimates uncertain nonlinear dynamics online, while the super-twisting term deals with estimation errors and bounded disturbances. Directed pinned coordination is used for voltage regulation, and a two-stage event-triggered scheme reduces communication between units. The controller is tested on a six-node PV-battery DC microgrid under nominal conditions, 40 stochastic Monte Carlo runs, and communication delays. Under nominal conditions, the RMSE is 0.0089 V, the MAE is 0.0074 V, and the settling time is 0.601 s. The controller requires 3228 network-wide transmissions, 91.03% fewer than 1 ms periodic communication. Across the Monte Carlo runs, the mean RMSE is 0.00704 V with a standard deviation of 0.00027 V. Tests with delays from 0.5 to 10 ms show small changes in voltage error and communication. The analysis establishes conditional practical boundedness for the reduced continuous-time model. Simulations evaluate the complete digital implementation with sampling, saturation, and delayed communication; no analytical delay margin is established.
Authors
- Mazda Moattari (ORCID: https://orcid.org/0000-0001-8361-5569)
- Heydar Izadneshan (ORCID: https://orcid.org/0000-0001-6958-5359)
- Hooman Fetanat Fard Haghighi
Institutions
- Islamic Azad University, Marvdasht (IR)
Publication Details
- Journal
- Scientific Reports
- Published
- 2026-11-17
- DOI
- https://doi.org/10.1038/s41598-026-78044-w
- Primary Topic
- Microgrid Control and Optimization
- Type
- article
- Field-Weighted Citation Impact
- 0.00