A G 3 PLC Network Simulator with Enhanced

A G 3 -PLC Network Simulator with Enhanced Link Level Modeling A. Sanz 1, D. Sancho 1, C. Guemes 1, P. Estopiñan 1 and J. A. Cortés 2 1 Microchip Technology Inc. 2 Dpt. Ingeniería de Comunicaciones, University of Málaga (Spain)

Contents l l Introduction G 3 -PLC Network Simulator Architecture l Events description l Channel model l Frame error estimation l l l Validation and network performance analysis Conclusion 2

Introduction l l Smart Metering network is the network that allows the remote management of elements of energy grid. The G 3 -PLC the OFDM specification suitable for this purpose. Network simulators are very useful tools for developing and debugging the communication stack. This work presents a G 3 -PLC network simulator enhanced to achieve a more accurate modeling while allowing faster than real-time simulation of complex networks. 3

G 3 -PLC Network Simulator: Architecture l Each G 3 -PLC node is simulated by an independent process that implements the full stack, except most parts of the physical layer, and an event machine. l Layers are implemented employing the same code used in actual G 3 -PLC devices by Microchip. l A Control module commands the simulation and ensures its coherence. l Frame transmission events are managed by the Network process, which implements the physical layer and the shared power line communications (PLC) channel. Taken from [2] A. Sanz, P. J. Piñero, S. Miguel, and J. I. Garcíaa-Nicolás as, “Distributed Event-Driven Simulation Environment for Smart Metering Protocols Evaluation, ” in Proceedings of the IEEE International Conference on Smart Grid Communications, 2012, pp. 151– 156. 4

G 3 -PLC Network Simulator: Events description l Example: l l l Events associated to the transmission of two frames in a simplified scenario with three nodes plus the network coordinator. Frames transmitted by node 1 and node 3 reach the coordinator and node 2 with a signal level above the receiver sensitivity. Direct communication between node 1 and node 3 is not possible. 5

G 3 -PLC Network Simulator: Channel model l l The employed channel models allows simulating frequency selective channels with colored noise, including (simple) impulsive noise. Intercarrier and intersymbol interference is disregarded. The effect of the interfering frames is taken into account as an additive term independent of the desired signal and the channel noise. Consider a frame transmission from node s to node r. The transmitted frame consists of M OFDM symbols and employs the set of carriers K. The signal-tointerference-and-noise (SINR) at the receiver side for the constellation value received in the kth carrier of the mth symbol is given by Power allocated by node s to carrier k Attenuation of the channel between nodes s and r for carrier k Noise at node r in the If node i is transmitting a frame while node r is band of carrier k receiving the desired frame, Ai(m)=1, Set of interfering Drawn from a RV in each nodes otherwise frame! Ai(m)=0 6

G 3 -PLC Network Simulator: Frame error estimation l l l Frames are divided into regions for error estimation. Frame regions consists of an integer number of OFDM symbols with the same modulation and coding scheme and constant SINR. Example with 4 regions: -The FCH and the payload must be always in different regions because they use different coding schemes. -Interfering frames yield new regions. 7

G 3 -PLC Network Simulator: Frame error estimation l Bit error probability (Pb) in each region is estimated from the Frame Error Rate (FER) as Number of data bits in the whole frame l The FER corresponding to a channel state characterized by is estimated using the effective signal-to-interference mapping function (ESM), where the effective SNR is given by FER in AWGN channel where β depends on the modulation and the coding scheme, but not on the channel state. 8

G 3 -PLC Network Simulator: Frame error estimation l l There is an optimum value of β for each modulation defined in the G 3 PLC system. They are computed as the solution to the minimization problem Effective SNR fulfils corresponding to channel state j Thousands of channel estates J have been used. For each channel state, the following process is followed to compute the values used in the optimization Estimated by simulations Generated Estimated by simulations Obtained from the definition 9

G 3 -PLC Network Simulator: Frame error estimation l Optimum values of β for the modulations used in the payload l Using this procedure, the error between the actual FER and the estimated one is lower than 0. 5 d. B in 80% of the cases (channel states), except for BPSK in the FCC band, where this percentage is a bit lower, but still higher than 70%. 10

Validation and network performance analysis Test description l Simulated results compared to the ones obtained in a test network deployed in the laboratory. l Tested network: 100 nodes distributed in 5 levels. Flat attenuation of 50 d. B between levels. l A line impedance stabilization network (LISN) is used to control the noise level. l The application layer emulates the DLMS/COSEM protocol. l Simulation has been executed in a Dell Precision T 7600 workstation. Results l The simulated-time to real-time ratio for the tested network is 17/60, i. e. faster than real-time. 11

Conclusion l A G 3 -PLC network simulator has been presented: l l The simulator implements the full stack using the same code that is embedded in actual Microchip G 3 -PLC devices. The implemented channel models allows simulating frequency selective channel responses and colored noise, which can be also time-varying. Frame errors are estimated using the ESM function and takes into account the possible collisions with other frames. The simulator has been validated by comparing their results to a test network deployed in the laboratory consisting of a coordinator and 100 meters distributed in 5 levels. l l Faster than real-time results obtained. Excellent match between simulations and measurements. 12