Model-Based Latency Compensation for Network Controlled Modular Multilevel Converters

Abstract

The use of an internal digital communication network enhances the scalability, implementation and maintenance of Modular Multilevel Converters (MMC). However, it also introduces delays that limit the sampling frequency and the controller dynamic performance. In this paper, we propose a model-based predictor to compensate for the loop delay and overcome these limitations. Two benefits of this approach are possible: either designers can increase the sampling rate and control performance or employ a slower communication protocol/technology. In this paper, we present the mathematical description of the model-based predictor, assess the parameter sensitivity, and show matching simulation and experimental results that validate it. As constraints introduced by the use of digital communications are overcome, the results achieved encourage engineers to adopt a network into the design of Modular Multilevel Converters.

Hardware Support to Minimize the End-to-End Delay in Ethernet-Based Ring Networks

“…They are of the same order of magnitude of the MCT in small networks and they contribute to the end-to-end communication delay. As discussed in [24], the end-to-end latency adds to the loop delay and has negative influence in the closed loop system response. Moreover, in certain protocols (e.g., POWERLINK), the MAC may also be involved in the packet forwarding, causing the respective delay to impact the MCT, too.…”

Nonparametric Frequency Response Identification for Dc-Dc Converters Based on Spectral Analysis with Automatic Determination of the Perturbation Amplitude

“…The implementation of identification techniques can be done using digital control platforms such as FPGAs [ 2 , 5 , 10 , 21 , 22 , 24 ], microcontrollers [ 9 , 25 ] or Systems on Chip (SoC) [ 26 , 27 , 28 ]. FPGA implementation allows for task parallelization and a very accurate timing.…”

“…Second, spectral analysis is applied for the identification itself, allowing open and closed-loop operation. Finally, the implementation on a SoC [ 26 , 27 , 28 ] allows full online operation, integrating all processing task in the digital platform and very accurate timing control.…”