Computational Simulation of Lightning Interaction with ... · German Zeppelin LZ40 (L10) Destroyed...

Computational Simulation of Lightning Interaction with Systems for Surge Protection Design

Tim McDonald, PhDtim@ema3d.comIEEE PES SPDC

EMA Simulation for Surge Protection Design

• Aircraft Lightning Phenomenology• Current Challenging in Simulation Technologies• Case-study: Rocket Lightning Protection Design using Simulation• Case-study: Wind Turbine Lightning Protection Design using Simulation• How to Model Complex Cable Harnesses on Real Platforms• The Importance of Measurements

Aircraft Lightning Phenomenology

EMA PROPRIETARY INFORMATION

EMA PROPRIETARY INFORMATION 5

A Cockpit View of Lightning BetweenBeijing and Shanghai

A View from the Cockpit:How Does Lightning Interact with this Aircraft?

EMA PROPRIETARY INFORMATION7

8EMA PROPRIETARY INFORMATION

EMA PROPRIETARY INFORMATION 9EMA PROPRIETARY INFORMATION

EMA PROPRIETARY INFORMATION11

The First Lightning Crashes of Aircraft3rd September 1915 3rd September 1929

German Zeppelin LZ40 (L10)Destroyed by lightning off Neuwerk Island, Germany.

Ford AT-5 Tri-MotorCrash of first heavier-than-air aircraft destroyed by a

lightning strike. All eight occupants died when the airplane struck ground near Mt. Taylor, New Mexico.

Description

• Stepped leader starts at cloud and travels toward the earth

• At a distance of ~50 m from earth, upward going leader begins

• Upward going leader connects to stepped leader

• Return stroke with large current travels upward to cloud

• Process may repeat

Lightning Cloud to Ground Scenario

Separation of charge by aerodynamic forces and contact electrification

Peak Power ~1013 watts

C = Speed of Light

~100 Million Volts

Image Charge

Upward Leader

Stepped Leader• I ~1000A• Speed ~C/1000

Return Stroke• I ~1000A avg.,

200kA 1% level• Speed ~C/3

Lightning Initiation by Cosmic Rays

Source:Joseph R. Dwyer, “A Bolt out of the Blue, Scientific American”, May, 2005

Click slide to animate image

One of the Few Photos of the Upward Going Leader

Return stroke draining cloud charge

The Cloud is a Charge Reservoir

Not All Lightning Goes to Earth17

Brightness (and current) decreases with altitude

Top of bright return stroke traveling upward

Some Facts About Lightning• 1% Peak Current: 200,000 amperes (100 watt light bulb uses

about 1/2 ampere)

• Largest measured: ~450,000 amperes (Sea of Japan)

• Restrikes in one flash: Up to about 24

• Most (90%) airplanes: Create their own lightning strikes (the lightning would not exist without the presence of the aircraft)

How Does Lightning Strike an Aircraft• Two fundamentally different lightning

strike scenarios for aircraft

– Naturally occurring strike

– Aircraft triggered strike

• Research programs have shown that about 90% of strikes to aircraft are triggered, and only 10% are naturally occurring

Two Fundamentally Different Lightning Strike Scenarios for Aircraft

ChargeRegion

OppositePolarityChargeRegion

Bi-directional junction leaders originate at airplane

Naturally occurringleader originates at

charge region

ChargeRegion

OppositePolarityChargeRegion

Bi-directional leaders originate at airplane

Naturally occurring strike (~10%) Aircraft Triggered strike (~90%)

NASA F-106 Lightning Research Program (1982-1989)

• Objective: Fly instrumented aircraft into thunderstorms to intercept lightning and collect data to understand it

• Experienced more than 700 strikes

• Most were aircraft triggered lightning

• EMA/R.Perala participated

24 EMA PROPRIETARY INFORMATION

NASA F-106 Lightning Research Program

(1982-1989)

F-106 work Resulted in:

• Article in IEEE Spectrum Magazine July 1988

• Best Paper Award at NASA Langley Research Center

• Impacted SAE ARP 5412 lightning environments, especially waveform H

Sweeping Strokes Were Indicated by Photography and Attachment Patterns

View Looking Aft From Side Mounted Camera

What Causes an Aircraft to Trigger Lightning?

Answer: The local electric field at an aircraft extremity becomes large enough to cause air breakdown

• Aircraft local E field, directly related to local charge density Q, has two components:– Aircraft net charge, caused by normal P-Static (precipitation static), including engine

charging

– Aircraft polarization charge, caused by the background thunderstorm electric fields

P-Static Charging by Triboelectrification• Collision with atmospheric particles

• Snow, hail, graupel, dust, snow, ice crystals, rain, water droplets

• Charge transfer at contact, also called contact electrification

• Charging rates proportional to aircraft speed and frontal area (collision cross section)

Engine Charging• Sodium impurities in exhaust

• Sodium ions transport charge away, leaving net charge behind

• Requires an ambient electric field (fair weather ~100 V/m; or thunderstorm field)

Example: Net Charge Distribution+ + + +

+ + + + + + + + + + + + + + + + + + + +

+ + + +

QsC/m2

S∫Qsds = Qnet

distance

+ + + + + + + + + + + + + + + + + + + + + + + + +

+ ++ + + +

+ ++ + + ++ + + +

+ + + + + + + +

+ + + + + +

Thunderstorm Static Electric Field

QsC/m2

S∫Qsds = 0

- - - - - - - - - - -- - - - - - - - - - - - - + + + + + - - - - - - - - - - - - - - + + + + + - - - - - - - - - - - - - - + + + + + + + +

- - - - - - - - - + + + + + + + +

distance

+ ++ + + ++ + + +

- - - - - - - - - - - - -- - - - - - - - - -- - - - - - - - - -

- - - - - -

-- - - - - -

+ + + + + + + + + + + + + + + + + + + ++ + + + + + +

- - - - - - - - - - - - -- - - - - - - - - -- - - - - - - - - -

- - - - - -

+ + + + + +

- -- - - -- - - -

- - - -- - - -- - - -

- - - -

Example: Polarization Charge Distribution

Lightning protection strips

Example: Triggered Lightning Simulation on 265m long Airship

.mkV10=axE

90 kV/m

Enhancement Factor = 9

Electric Field Enhancement FactorsAxial Background Electric Field

Triggered Lightning ExampleIn-Flight Measured and Computed Surface B-dot Probes

• Comparison of in-flight measured and modeled (based on non-linear air chemistry) results for F-106 triggered lightning

• Axial ambient field 200kV/m

• Net aircraft charge 0.9mC

• The longitudinal B-dot probe is on the top of the F-106 fuselage

Interpretation• An initial breakdown at the nose• A later breakdown at the tail

Air Breakdown Model and Prediction of Triggered Currents

• Air conductivity nonlinear function of the total electric field

q is the electronic charge, 1.6 x10-19 coulombs,

ne is the number density of electrons [m-3],

n- is the number density of negative ions [m-3],

n+ is the number density of positive ions [m-3],

µe is the electron mobility [m2/volt-sec] and

µi is the ion mobility [m2/volt-sec].

𝜎𝜎 = 𝑞𝑞(𝑛𝑛𝑛𝑛 𝜇𝜇𝑛𝑛 + 𝑛𝑛 + 𝑛𝑛 𝜇𝜇𝑖𝑖) 𝑛𝑛 = ne + n-- +

Ionization Rate Equations for the Creation andLoss of the Three Species

GntQnnnnvnt

nnnvnt

tQnGnvnt

+=++•∇+∂∂

=+•∇+∂∂

=−++•∇+∂∂

−+++++

−+−−−

)(][)(

αe is the electron attachment rate (sec-1),G is the electron avalanche rate (sec-1),β is the electron-ion recombination coefficient (m3-sec-1),δ is the negative-positive ion recombination

Comparison of Measured and Computed Responses in-flight on F-106B Measured

Calculated

Pulses Measured on F-106B for Typical Triggered

Lightning Strike

(400 Hz to 100 kHz)

FIN CURRENT

BOOM CURRENT

FIN CURRENT (DC to 400 Hz)

LIGHT SENSOR

RECORDER TRIGGER DISCRETE

Temporal character of typical lightning strike.

100 MILLISECONDS

Pulses Measured on F-106B Vertical Fin Cap

for Typical Lightning Strike

Temporal character of typical lightning strike.

(dc-400 Hz)

Transient recordertrigger

Pulses Measured on F-106 Vertical Fin Cap with 40 ns Sampling Rate for Typical Triggered Lightning Strike

Time, 𝜇𝜇s

Vertical fin cap current corresponding to figure 3. (Recorded at 40-ns sample interval.

0 320 640 960 1280 1600 1920 2240 2560

It, kA

Peak Current Level Distributions for Aircraft Triggered

and Natural Lightning

A View from the Cockpit:How Does Lightning Interact with this Aircraft?

Answer: Aircraft charging by P-Static and External Background E-fields creates arcs and streamers around

the aircraft. Some are swept over the windshield.

The Problem

• Design teams are expected to use 3D CAD information

• The design cycles are shorter• Teams must do more work in less

PLM and Electromagnetics

• Currently experiencing universal adoption of product lifecycle management (PLM) systems that require a high-level of interoperability and a focus on change management.

• EM engineers must learn to use 3D geometry in a practical way in their work progress without letting the complexity of handling the full CAD overwhelm their design cycles

Challenges for EM• Though improvements in simulation speed and new

acceleration methods are always desired, they are not the limiting factor

• More than 90 % of the EM effort is dominated by non-solver issues:– CAD to CAE, cleaning and defeaturing– Interpretation of results and design modifications– Requirements trace, system engineering and test

specification• Remainder of this segment of the talk will focus on

these non-solver concerns

Automating geometric manipulation tasks

• CAD to CAE dominates the simulation work cycle

• Current state-of-the-art technologies automate common tasks

• The trend of automation will continue to include machine-learning geometry operations

Automated Simplification

• The problem is not lack of detail but too much detail.

• Current mechanical CAD is heavy and complex.

• New advances in automatic simplification allow a reduction in CAD size and complexity without sacrificing accuracy.

Joining Objects to prepare for Simulation

• Shrink-wrap generators join bodies in contact, ignore small gaps and removes small holes to prepare for successful automatic meshing

Machine Learning for CAD Simplification

• Frontier of this realm is machine learning that is trained by analyzing human processing steps

• Requires database of “before” and “after” geometry to train models

Falcon Rocket – Margin

Issues

Lightning Strike to

Catenary

Integrated harness and full-wave software for the correct coupling

Diode Sizing

• Candidate TVS load placed in series• Current through TVS calculated• Power dissipated is current time clamping voltage• Compare power to de-rated power dissipation limit

of selected part• Repeat if necessary

EMA3D CAD Import

Cable Routes

New Cables in EMA3D CAD

FDTD Mesh (50 mm)

Cable Cross Section

Cable 1 Cable 2 Cable 3

EMA3D Cable Cross Section Tool

• EMA3D has the unique capability of Co-simulating with a full-featured Transmission Line Solver

• Ability to quickly calculate the accurate capacitance and inductance matrix for cable cross-section

• Easy drag and drop cable cross section tool

• EMA3D full-wave simulation completed in minutes

• Built-in gradual permittivity scaling speeds up time-domain simulation by a factor or more than ten

• Built-in parallel simulation capabilities allow for speed up over 100 times on multi-core workstations and computational clusters

Magnetic Fields

• EMA3D can create color picture probes, slice probes, and animations of any fields or currents as necessary.

• The tools are easy to use and visualize the results

• The results are in ASCII and are readily exported to other packages with minimal manipulation

Magnetic Field Slice Probe

Waveform 1 Overall Cable Bundle Current

Waveform 1 Individual Pin Current

Waveform 1 Individual Pin Voltage

DC Simulation

EMA3D built-in gradual permittivity scaling allows for a time-domain calculation of a ramp up to DC as described.

01Simulation completed in a little over an hour

02Not practical with other time-domain solvers

Ramp to 800 A DC Simulation Results: Pin Currents

Material Library

EMA3D includes built-in sample properties for standard types of cables:• Diameter• Resistance• Transfer impedance parameters

Classic Cable Simulation Real World Cables

The Cable Conundrum

Certain Cable Complexity is Required to Get the Right Answer

CEM Model Details• Simplified bulk representations of

harnesses were initially used in the models

• The resulting answer diverged from a comparison to testing by 9 dB

Initial Bulk Cable Detailed Harness Model

Approximate diameter and resistance

Results from More Accurate HarnessMore accurate harness development in the model also permitted pin voltage calculation and comparison

0 1 2 3

x 10-4

20Vmax = 83t1 = 5.4e-05t2 = 1.9e-04

Time [s]

Test 17LTATM Mod 4

RF Cable Branching Effect on Coupling to Cables

• The branched cable dissipates nearly all of its energy by 100 μs, ten times faster than the cable harness without branches

• The first vehicle resonance as well as the other modes below 30 MHz have been largely reduced

Cable TopologyMaximum

Current Level(mA per V/m)

Current Level Reduction

Straight Uniformly Packed Cable 500 Baseline

With Branching 30 24 dB

The Best Solution

• Automation of simplification for cables• Rapid specification of cable harness details based on the wiring

diagram• Transmission line modeling co-simulated with 3D simulation to

resolve branches, shields, and multiple conductors• Accelerated methods to speed up calculation

Cable Routing Simplification

• The CAD description of cables can be incredibly complex

• Hand simplification takes a long time

• Over simplification can give the wrong answer

How do we get the harness details into the model?

• Many details need to be defined for cables:– Types (twisted pair, coax, etc.)– Wire gauge– Termination impedances– Circuit loads– Shield properties transfer

impedance

Cable Wiring Diagram Simulation Model

Cable Harness Definition in

MHARNESS Wizard

Concentrate on Measurements that

can be Generalized to Entire Platform

• Current approach to EMP, Lx and other E3 environments is based on pass-fail characterization of the entire platform

• Targeted measurements of critical material properties is fast and low cost

• These measurements are more useful to analysis techniques across the entire frequency range

Large Platform Validation is Enabling Technology

Computational Simulation of Lightning Interaction with Systems for Surge Protection Design

Tim McDonald, PhDtim@ema3d.comIEEE PES SPDC

Computational Simulation of Lightning Interaction with ... · German Zeppelin LZ40 (L10) Destroyed...

Documents

VISUALISIERUNG - Burg Campus Onlinecampus.burg-halle.de/id-neuwerk/wetter/wp-content/uploads/sites/47/... · Ozeanische Beobachtungen Oberflächennah Ozeantemperatur Atmosphärische

Protection against lightning Lightning Arrestor

Lightning Protection Lightning Protection Application Guide

Hart-Malone: Lightning and Lightning Protection

External lightning protection Insulated lightning

Lightning: Principles, Instruments and Applications · Lightning: Principles, Instruments and Applications Review of Modern Lightning Research Lightning represents a natural phenomenon

John Mayer - Heavier Things (Songbook)

Challenges with the Production of Heavier Crudes

Bigger Trailers. Heavier loads. Same Old Tractor

Fundamentals of Lightning and Lightning Arrester ... · PDF fileFundamentals of Lightning and . Lightning Arrester Protection Zone Determination . For Protection Engineers . Seung

Main catalog hélita lightning protection systems External ... · ABB External lightning protection | 1 Lightning mechanism and location 2 Lightning protection technologies 3 Lightning

LIGHTNING HOW LIGHTNING HURTS US

Bigger, larger, heavier

der Pfarrei Maria von den Aposteln Neuwerk · 2.1. Das Territorium der GdG Neuwerk Das Territorium der Pfarrei Maria von den Aposteln umfasst die ehemaligen Pfar-ren und seit 2013

Lightning Component × Lightning Design System

BEST PRACTICES WITH POWER - MikroTikmum.mikrotik.com/presentations/ZA13/chris.pdf · BEST PRACTICES WITH ... Lightning. Lightning. Lightning. Lightning. Power Protection Surge What

Gemeindeblatt der ev. Kirchengemeinde Neuwerk

Lightning componentとlightning design system

Density. Which is heavier, gold or feathers? Which is heavier, cup of foam or cup of beans?

Electricity Lightning Lightning, Electricity Lightning ...kestrel.nmt.edu/~sessions/phys189/ab_ch11.pdfElectricity Lightning Thunderstorms Tornadoes Electricity Fundamentals of electricity