Abstract
Bayesian inference is used to calibrate a bottom-up home PLC network model with unknown loads and wires at frequencies up to 30 MHz. A network topology with over 50 parameters is calibrated using global sensitivity analysis and transitional Markov Chain Monte Carlo (TMCMC). The sensitivity-informed Bayesian inference computes Sobol indices for each network parameter and applies TMCMC to calibrate the most sensitive parameters for a given network topology. A greedy random search with TMCMC is used to refine the discrete random variables of the network. This results in a model that can accurately compute the transfer function despite noisy training data and a high dimensional parameter space. The model is able to infer some parameters of the network used to produce the training data, and accurately computes the transfer function under extrapolative scenarios.
Highlights
Bayesian Inference for Home PLCPower Line Communications (PLC)Power Line Communications (PLC) are expected to be a key component of smart-grids, allowing grid usage to be monitored without adding extensive infrastructure
PLC systems are classified into three categories based on bandwidth: ultra narrowband (UNB) systems have frequencies ranging from 0.3–3 kHz, narrowband (NB) systems have a frequency range of 3–500 kHz, and broadband (BB) systems have frequencies of
The issue with PLC systems is that because they use existing power lines that are not designed for carrying communication signals, and can have high variability in the loads and wiring, the conditions are highly dependent on the specific network topology and loads
Summary
Power Line Communications (PLC) are expected to be a key component of smart-grids, allowing grid usage to be monitored without adding extensive infrastructure. The top-down modeling approach views conductive propagation as an unknown function to be estimated based on training data, and so does not use any prior information such as the network topology or component connections, but rather implements a parametric model based on transmission line (TL) theory [11]. The frequency response of a network is calculated using TL theory to maintain a physics-driven model, but the parameter estimation results in a network model that closely matches network-specific data This allows models to be calibrated for specific networks with unknown loads or wires. This paper is organized as follows: Section 2 describes the network topology and load models, Section 3 describes the modeling framework, Section 4 shows results of the calibrated model, and Section 5 has conclusions and suggestions for future works
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
