Improvements in the PLC Systems for Smart Grids Environments

Abstract

During recent past years there has existed an international effort to develop Smart Grid's technologies to help achieving more efficient ways of energy management. However, even though it seems clear that new smart grids should also include an information technology (IT) infrastructure, it is not so clear under what type of technologies it should be developed. One of the promising communication's technologies for smart grid's implementation is Power Line Communications (PLC). The main reason of PLC' suitability is that, since it uses the power cables as communication medium, this technology reaches all places where energy is present. However, PLC has to deal with the fact that power lines present a very harmful environment for signal transmission due to the presence of several noisy sources and severe attenuations.

The two main objectives of this PhD thesis are: first, propose modifications and techniques that increase the performance of current narrowband PLC standards; and second, develop a simulation tool that, considering all physical phenomena provides the expected system's performance in terms of end-to-end communication's latencies.

Using a simulation model developed in this thesis, a considerable weaker performance of PRIME can be observed when compared to G3-PLC in terms of number of bit errors. This difference is incremented in environments where impulsive noise is present, which is a common scenario in PLC networks. This kind of noise is very damaging in OFDM systems.

In order to decrease the effect of impulsive noise in narrowband PLC transmissions, a compressed sensing technique modification is proposed. Thanks to a reordering of the subcarriers used for data transmission in the OFDM spectrum, an enhanced reconstruction of the impulsive component is achieved. This reordering follows a difference set pattern. This proposal is tested in the PRIME transceiver's model implemented. There, an almost complete cancellation of the impulsive noise effect is observed. An additional advantage of this technique is that redundancy included in the message is not incremented, thus data bit rates are not decreased.

Following the second objective, a network simulator is used to emulate an Advanced Meter Infrastructure (AMI) system's behavior in terms of communications. Due to its good performance with respect to runtime and memory consumption, OMNeT++ network simulator is chosen for this task. The network simulation software uses physical performance from the previous Matlab analysis. This combined simulation framework creates an accurate modeling by taking into account both the whole communication's layer stack and the channel's effects. To the best knowledge of the author, this kind of approach has never been addressed in the literature before.

Main results obtained from network simulations show that, under impulsive environments, communication's latencies are too high to implement a tele-metering application using narrow band PLC over a low voltage network. By introducing the proposed modifications to PRIME in an impulsive noise environment, latencies are considerably reduced.

Finally, the author finds the use of this kind of simulation tools be very helpful to companies that are currently deploying their AMI systems. These simulations can provide the average number of meters that can be polled in a given period of time, which is an important parameter in the dimensioning of the network.