Impulse Noise Model

Impulse Noise Model The following presentation was part of a research project at Robert Bosch on automotive Power Line Communications Presenter: Thomas Hogenmüller (Bosch)

AE-BE/EKE-Hogenmüller | 24/04/2013 |

Power Line Automotive Noise Model

Analysis of Impulsive Noise Environment (1) Power Line Automotive Noise Model


Frequency [MHz] rpm

[min-1] P

ower

dens

ity

spec

trum

[dB

µV]

Frequency [MHz] rpm

[min-1]

Pow

erde

nsity

sp

ectru

m [d

BµV

]

Impulsive noise spectrum at battery mounted in engine compartement

Impulsive noise spectrum at fuse box in rear trunk

Different noise spectra at different locations within car

2

Analysis of Impulsive Noise Environment (2) Power Line Automotive Noise Model


Typical measurement result of a manual switch operation in cars

(e.g. terminal 15 switch operation)

Simulation of the measurment with ISO-pulse

(e.g. test pulse 3a) Am

plitu

de [V

] A

mpl

itude

[V]

Time [µs] Time [µs]

Time [µs] Time [µs]

Freq

uenc

y [M

Hz]

Fr

eque

ncy

[MH

z]

3

Analysis of Impulsive Noise Environment (3)

→ ISO-pulses are not appropriate for simulation of automotive disturbances



On the contrary to manual electrical switches, spark pulses produce disturbances up to high frequencies

Am

plitu

de [V

] A

mpl

itude

[V]

Frequency [MHz]

Time [µs]

Pow

er d

ensi

ty

spec

trum

[dB

µV]

Time-frequency domain Time domain

Am

plitu

de [V

] A

mpl

itude

[V]

Time [µs]

Time domain

Spark plug ignition 2 Spark plug

ignition 1

Trigger signal (Spark plug 1) Spark of plug 2

Spark of plug 1

Spark of plug 2

4



Conclusion of Impulsive Noise Analysis Below 100 MHz:

ISO pulses according to ISO7637 are applicable

Above 100 MHz Spark engine system is mainly responsible for pulse noise.

Furthermore the highest level of pulse noise can be observed during engine idle.

5



Modelling of Impulsive Noise Creation of a configurable impulsive noise model:

6

Automotive Impulsive Noise Modelling Power Line Automotive Noise Model


Aperiodic sequence of impulsive noise

Periodic sequence of impulse pattern

→ Generation of stochastic random variables with Quantiltransformation

Impulse sample duration

Impulse sample period

Impulse amplitude

Impulse period

Impulse duration

Duration [ns] Amplitude [V]

rel.

frequ

ency

2D-Histogram of pulse duration and pulse amplitude

Time duration [µs] re

l. fre

quen

cy Histogramm des Impulsabstands

7



Spark impulsive noise can be defined with 5 parameters: impulse distance, impulse amplitude, Impulse period and Impulse sample distance

which are stochastical values. The fifth parameter Impulse sample period,

is a deterministic value depending on the engine speed

Spark Impulsive Noise (SIN) Parameter

Impulse sample distance


Impulse amplitude

Impulse period

Impulse distance

8



Guideline for SIN Simulation 1. Generation of a switch signal depending on engine rotation speed

for identification of a spark event (→ impulse sample distance and impulse sample period)

2. Rectangular signal overlay according to trigger time aperiodic disturbance pulses within one pulse pattern (→ impulse period and impulse distance)

3. Increase of background noise for the duration of an aperiodic impulse noise (→ impulse amplitude)



Impulse amplitude

Impulse period

Impulse distance

9



Quantiltransformation Distribution function

Quantiltransformation

Quantiltransformation

X-Values

SNP -Random Numbers

Y-V

alue

s

Gen

erat

ion

unifo

rmly

dis

tribu

ted

rand

om n

umbe

rs

10



SIN Composition Block diagram of spark impulsive noise generator



Impulse amplitude

Impulse period

Impulse distance

m-sequence

m-sequence

VTbQ-Trafo Impulse distance

VTbQ-Trafo Impulse period

Impulse amplitude

Sample distance

VTbQ-Trafo sample period

m-sequence

AND

White noise Filter

Spark impulsive noise

11

Simulation of density function

Simulation of measurement data



Flowchart for Impulse Noise Model Set of Measurement data

Combinded data set

Reduced data set

Distribution table

Random number, events are midpoint of class

Random numbers with required density (impulse duration etc.)

Impuls noise signal

Random numbers with additional values

Distribution function

Estimated density function

Analysis

Synthesis

12

Channel model



Frequency Response Frequency range 100 MHz ≤ f ≤ 300 MHz With center frequency f0 = 200 MHz Sample frequency denoted as fa

SIN + additive noise

sender receiver

noise n(t)

Channel H(f,t)=const.

13



Frequency Response Frequency response is simulated with an IIR filter of order 12. Frequency band is normalized to center frequency f0=200 MHz and

sample frequency fa=200 MHz

(a) Frequency response (b) Phase response

measurement measurement

sim model sim model

14



Conclusion ISO pulses are not sufficient to model impulse noise in vehicle Main impulse noise source are spark plugs What’s available:

Measurement data base for power lines Mathematical description of noise FPGA based emulator (for automotive use)

PhD-Thesis of Thorsten Huck is available in German at Shaker publisher for 50 € (I try to find out if the relevant part can be published at IEEE)

15

http://www.shaker.de/de/content/catalogue/index.asp?lang=de&ID=8&ISBN=978-3-8322-5614-2

Impulse Noise Model

Documents

Transcript of Impulse Noise Model