Experimental Study on Deflagration to Detonation

Transition Enhancement Methods in a PDE

By Philip K. Panicker, Graduate Research Associate, ARC, MAE, UT Arlington
Co-authors:
Daniel T. H. New, Lecturer, Department of Engineering, University of Liverpool,
Brownlow Hill, UK
Kin F. Chui, Associate Scientist
Her Mann Tsai, Principal Research Scientist
Temasek Laboratories, National University of Singapore
Frank K. Lu, Professor, MAE Department, UT Arlington

AERODYNAMICS RESEARCH CENTER

08/03/2007
MECHANICAL AND AEROSPACE ENGINEERING
16:39 DEPARTMENT 1
 Agenda
„ Introduction
(Pulsed Detonation Engine basics, the PDE cycle,
Deflagration vs. Detonation)
„ Previous studies and results
„ Objective of the present studies
„ Experimental Program
„ Results
„ Conclusion

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 2
 Pulsed Detonation Engines: Introduction
„ PDEs employ a detonation wave to compress and combust fuel and oxidizer
mixture. This process is repeated in an intermittent or cyclical fashion.
„ Conventional compressor-turbine jet engines, use the deflagration process to release
energy from the fuel, in a steady process.
„ In PDEs, the detonation wave does the work of compressing the flow, thus
eliminating the need for a compressor stage found in turbofan, turbojet engines. This
reduces complexity of engine, weight, cost and thus fuel requirements.
„ Detonation process is thermally more efficient than deflagration, because the energy
release from a fuel detonation is much more than from its deflagration, thereby
reducing the fuel requirement.
„ Detonation allows for rapid energy conversion.
„ Thus, Compact and Efficient Physical Systems can be designed for propulsion
applications.
„ PDEs can be used from Mach 0 to about 5.
„ PDEs can be used in supersonic or hypersonic vehicles.
„ Detonation models used for studies, includes: the Chapman-Jouguet (CJ) model- A
shock wave closely coupled to a think flame front or combustion region and the
Zeldovich-Von Neumann-Doring (ZND) model, which uses finite rate chemistry.
The shock wave is followed closely by a reaction zone, consisting of an induction zone
and a heat addition zone.
AERODYNAMICS RESEARCH CENTER
MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 3
 Deflagration vs. Detonation
Mode of
Deflagration Detonation
Combustion

Low speed
combustion Supersonic
1-10 m/s combustion
1000s of m/s

Deflagration
Supersonic
/ Subsonic
Modeled as Modeled as Combustion
Combustion
Constant Pressure Constant Volume
process Process,
Shock Wave
Followed by Reaction
Zone
Decrease in (CJ or ZND model) Quasi-
Quasi-steady Quasi-
Quasi-steady
Steady State
Density, or or Steady
Combustion
Slight decrease Intermittent Intermittent Combustion
e.g. Turbojets,
in Pressure Sharp increase Combustion Combustion e.g. Scramjets
Ramjets
In Pressure, Density e.g. Pulse Jets e.g. PDEs
Due to Shock Wave.
Pulse
Detonation
Rockets
AERODYNAMICS RESEARCH CENTER
MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 4
 PDE Cycle Tcycle= Tfilling + Tinitiation + Tdetonation + Texhaust+ Tpurge
Tdetonation is very short and quick
Tfilling and Tpurge are the longest stages in the cycle
Frequency = 1/Tcycle
Tpurge is a necessary and important stage for cooling and
Shorter Tubes => Higher Frequency cleaning the tube
Longer Tubes => Lower Frequency Texhaust is the thrust producing stage

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 5
 PDE Cycle ~ Humphrey Cycle

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 6
 Means of Achieving Detonation in the Combustion Chamber
For the PDE to work as designed, it is important to achieve detonations consistently during each
cycle, rather than slipping into deflagration mode. This is one of the main challenges of current PDE
research.

1. High Energy Arc Discharge 9

High Energy Laser Ignition 8
4 5 6 7 8
2 3
2.
3. Small Explosive Charge 8
c d e

Pre-Detonator running an easily

b f

4.
detonable mixture, transitioning to PDE Assembly

Main Combustor containing another Current work at the ARC involves the
fuel + oxidizer mix → development of a 2 stage PDE.
5. Deflagration to Detonation
Transition (DDT) 9 Pre-detonator running Propane + Oxygen.
The large combustor running on pre-mixed
liquid fuel (JP) and air.
8 Not Tried in lab
Liquid fuel is sprayed through atomizing
9 Tested in lab
high pressure gasoline injectors.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 7
 Means of Achieving Detonation in the Combustion Chamber
1. High Energy Arc Ignition to Before
Initiate Detonation
„ Requires bulky electrical circuit components.
„ Transformers, capacitor banks are heavy
and impractical for flight weight model.
„ Power generation to initiate arcs will deplete
engine’s available power output.
„ Most of the power is wasted electrically
when arc current flows to ground.
„ Ignition plugs sustain heavy damage and
have reduced life.
„ The Arc is a big source of Electro-Magnetic
noise that will drown out signals from
transducers.

After a few seconds of Arc discharges

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 8
 Means of Achieving Detonation in the Combustion Chamber
2. Deflagration to Detonation Transition (DDT)
Another method of inducing Detonations in the combustion chamber is to ignite the fuel-oxidizer
mixture and then to rapidly accelerate the combustion to Detonation.
One of the most successful DDT enhancement devices is the Shchelkin spiral, which is a helical
spring, inserted into the tube downstream of the ignition point.
The device is named after Russian Physicist Prof. K. I. Shchelkin, who proposed the spiral to bring
about DDT in his 1965 book “Gas Dynamics of Combustion”.

•Shchelkin spiral and other

obstacles to enhance turbulent Length L D

mixing enabling DDT.

•Low energy ignition source. D1

•Vary blockage ratio, pitch,

length of spiral.
Blockage Ratio = (Area of Cross Section covered by Spring)
(Total Internal Area of Cross Section of Tube)
•Reduction in DDT Run up
distance & time.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 9
 Means of Achieving Detonation in the Combustion Chamber
Shchelkin Spiral for DDT
June 2005 PDE Experimental Study
•Propane + Oxygen
•Rotary Valve injection system
•Low energy ignition system
•Blockage ratios of 35 to 56% tested
•(Tube id = 25mm/ 1in)
•Best DDT results from 55% BR, short
length (30cm/12 in).
PDE test June 2005 in which various Shchelkin spiral
•Smaller BR 35% needs longer length configurations were tested. Picture shows huge amounts
(60cm/24in). of steam generated from the cooling water.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 10
 Disadvantages of Spirals

•Blockage causes Drag.

•Spirals disintegrate
during testing and do not
Condition of spirals after just 10 to 15 seconds of testing.
last too long.
Notice the tube getting red hot and smoking at the
•Causes heating of spiral location of the Shchelkin spiral.
tube section, impractical
for long duration testing.
AERODYNAMICS RESEARCH CENTER
MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 11
 Objective of Study 1/2
There are two main objectives to this study. They are presented on this and the proceeding slides.
1. To build a PDE platform using off the shelf commercial/automotive
solenoid valves or fuel injectors.
„ Precise control using TTL generated from the DAQ computer (LabVIEW).
„ Variable frequency of operation at the touch of a button.
„ Low cost.
„ Easy to assemble and replace.
„ Rugged, field tested in the market place.
„ Flow rates and other data available from manufacturer enabling the operator to
design and set up tests easily.
„ Rotary valves tend to leak and de-synchronize due to belt slippage. They require a
magnetic pickup or optical sensor to determine valve position for timing the
ignition spark.
TTL : Transistor-Transistor Logic (0-5V)
DAQ: Data Acquisition System
LabVIEW is a proprietary graphical programming language of National Instruments.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 12
 Objective of Study 2/2
2. To test various DDT enhancing
devices and compare their
performances against results
from Shchelkin spirals tested
now as well as with results
obtained from previous studies.
„ Conical convergent-divergent
nozzles
„ Helical grooved sections
„ Circumferentially grooved sections

Since Shchelkin spirals are not

able to withstand the heavy
thermal and pressure shocks in the
PDE, a more robust DDT device
is required.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 13
 Experimental Program
PDE Platform

ASME Schedule 80 Stainless Steel pipe 24.3 mm id, For successful propagation of Detonation
33.4mm od D > 13 λ, where λ is cell size
Total length of PDE tube = 800 mm Cell size λ of Propane-Oxygen = 1.3mm
L/D=~33 ∴D/ λ = 18.7 > 13 λ

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 14
 Fuel and Gas injectors AFS-Gs60-05-5c series Fuel Valves
Solenoid Valves Alternate Fuel Systems Inc. Calgary, Canada
•12 Vdc, 8A peak, 2A hold.
•Fast acting (2ms reaction time). Up to 35Hz tested
successfully. 50Hz max.
•Easy and precise control possible using TTL
signals from a remote computer.
•Max pressure 550 kPa (90psig)

Valve
Controller

US
Quarter
coin

These valves however cannot be operated over rated pressures and at higher temperatures (60 deg. C). Thus
water cooling was provided for the injection modules.
The manufacturer provides flow rates of various gases. Therefore, flow rates need not be measured. Flow
rates can be precisely controlled by varying the duration of valve opening and the supply pressure.
AERODYNAMICS RESEARCH CENTER
MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 15
 TTL Control of Solenoid Valves and Ignition
Volts

Purge Air
50% Valve
Cycle
Volts

Fuel, Oxygen
Valves
40%
Cycle
Volts

5%
Ignition
Control
Cycle

Time (s)
Tests were performed at 10, 15, 20, 25 and 30 Hz.
15 Hz proved to be an optimum frequency based on the volume of the tube and
the flow rates of the valves, ensuring full tube filling of the gaseous mixture.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 16
 Ignition Plugs
To Ignition Coil

3k Ohm

Ceramic Tube

Tungsten Rod

„ Built in house
„ Lanthanated Tungsten rods (4mm/
3/16in dia)
„ Variable spark gap
„ 3k Ohm resistor reduces spark current
and thereby reduces radiated noise.
„ (Commercial automotive spark plugs
have built in resistance 3k to 8k Ohms,
to reduce EM noise. Without that, the
car radios and other electronics will not
Ground work.)

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 17
 Modified Automotive Ignition System

150 mJ max energy per spark

A transistor control circuit built in house
enabled the interfacing of the ignition system to
TTL controls from the DAQ.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 18
 Diagnostics and Controls

PCB 111A24 Pressure Transducers

PCB 201B05 load cell
National Instruments 16 bit PXI
chassis DAQ with LabVIEW
Sample and hold 240 kHz x 5 sec.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 19
 Water cooling for PDE Setup
transducers and valves
Stoichiometric
Propane and
Oxygen
injection

Exhaust
chamber

Linear Guide System

for measurement of
Thrust
AERODYNAMICS RESEARCH CENTER
MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 20
 Safety Features

All supply lines were fitted with

check valves to protect the solenoid
valves against blow back.
Flash back arrestors were also fitted
on oxygen and the fuel lines to
prevent the flame from propagating
back into the supply lines.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 21
 PDE set up pictures

Load cell for thrust

measurements
AERODYNAMICS RESEARCH CENTER
MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 22
 DDT Devices Tested

Wire Wire Pitch

Blockage Length Outer Dia L/
Item width depth (mm/
Ratio (mm) (mm) D
(mm) (mm) turn)

Shchelkin
Spiral 50% 160 24.3 4 4 8 6.6

Circ. Groove 0 160 24.3 5 4.55 10 6.6

Helical
Groove 0 160 24.3 5 4.55 10 6.6
Blockage Length Outer Dia Min.
Ratio (mm) (mm) Dia
15 deg. CD
sect. 50% 60 24.3 17.2
30 deg. CD
sect. 50% 60 24.3 17.2

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 23
 Test Results

Success is counted when stoichiometric Propane Oxygen theoretical CJ velocity and

pressure are obtained.
Average TOF velocity = 2360 m/s
Average Shock pressure (Von Neumann spike) = 36.7 bars

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 24
 Test Results (contd.)
Shchelkin spiral test, 15Hz

Average TOF velocity profiles of various DDT devices for 15Hz tests.

Success is counted when stoichiometric Propane Oxygen theoretical CJ velocity and

pressure are obtained.
Average TOF velocity = 2360 m/s
Average Shock pressure (Von Neumann spike) = 36.7 bars

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 25
 Test Results (contd. 2) „ Although spirals and the grooved tubular
devices are similar in shape, the Shchelkin
spiral has smooth undulations that may be
beneficial to creating a spin and improving
the turbulent mixing. Unsteady detonation
waves are known to have a spin along with
the cellular nature inherent to detonations.
The Shchelkin spirals of the right pitch
and inner dimensions are able to provide
these conditions enabling detonations to
propagate and accelerate.
„ Unlike the spiral, the grooves have more
abrupt geometrical variations, creating
impedances to the flow.
Average TOF velocity profiles of various DDT devices
for 15Hz tests. „ The CD nozzle types are intended to
create a Mach stem in the flow that would
provide good flow compression similar to
that of a normal shock. However, the
reverse was observed, with the throats
(PT= Pressure Transducer).
constricting the flow causing the
detonation wave to decouple.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 26
 It is observed that the velocities do increase
Test Results (contd. 3) „
from PT1 to PT2 before starting to die out
on all of them. The fuel-air mixture near
the exit is lean as the mixture could vent to
the ambient and get contaminated with air.
Thus, the detonation is not able to derive
the energy from the combustion to sustain
the detonation, causing the wave to slow
down and also to decouple.
„ It is observed that for the clean tube
configuration and the Shchelkin spiral case,
the TOF velocity increases beyond CJ limit
denoting over-driven detonations.
However, the velocities fall after PT3-PT4.
This could be because this part of the tube
has enriched fuel-oxygen mixture, after
which the mixture may become lean as the
location moves away from the injection
points.
„ The TOF velocities for all devices stayed
above 2000 m/s for the initial part of the
Average TOF velocity profiles of various DDT devices for 15Hz tests. combustion chamber tube length, with the
velocities dropping off towards the exit.
„ It could also be concluded that the length
of the DDT devices could be too short to
enable DDT. The length factor was not
tested in this study.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 27
 Test Results (contd. 4)
„It is significant that the peak pressures
reach 40 to 60 bar or higher in cases, even
though the ignition source is a low energy
(<150mJ) system. This is in agreement
with studies by others (such as Cooper, et
al.) who used a 50mJ system and achieved
40 to 50 bar peak pressures. Li et al. also
achieved similar results from a single shot
propane-oxygen study where pressures
ranged form 40 to 120 bar with a 470mJ
ignition source.
„There is some ambiguity in the peak
pressure levels due to ‘aliasing’ of
pressure signals due to the low sampling
rate (240 kHz). This is why there are
fluctuations in peak pressure levels from
on PT to another.
„Clean tube and the Shchelkin spiral
configurations seems to have consistent
pressure levels. The helical groove has
much lower pressure levels and low TOF
velocities, indicating that they are not as
good as the spirals at bringing about
DDT.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 28
 Test Results (contd. 5)

„The combination of two nozzles, 30 deg. followed by the 15 deg. nozzle,

yielded very high pressure peaks of 40 to 90 bar, but their TOF velocities are
not impressive. This could suggest that they cause the detonation wave to
decouple and slow down the pressure wave.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 29
 Video of 25 Hz test (Clean tube configuration)

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 30
 Operational Issues
•As can be seen in the pressure history of the
test shown to the left, the pressure peaks are
not always at the same level. Thus, success is
not evident at every firing cycle. The reasons
for this lack of success may be the following.
•The PDE consumed large masses of fuel and
oxygen in short periods of time. As a result,
there was considerable pressure loss in the gas
lines, as the gas cylinders (O2 and Propane)
were not able to keep up with the flow rate.
This could have led to non uniform filling of
tube and hence lack of DDT.
•Although the valves are able to be precisely
controlled, small variations in response from
injector to injector can lead to improper filling.

Shchelkin spiral test, 15Hz

AERODYNAMICS RESEARCH CENTER
MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 31
 Operational Issues (contd.)
•The most significant cause of the unsteady nature
of the Von Neumann peaks seen is due to ‘aliasing’
of signals due to under-sampling.
•The 240 kHz sampling rate was obviously too low
and a SR of at least 1MHz is recommended for
future detonation or shock wave studies.

•Also note the drift in the base line of the pressures

due to the high temperatures impacting the
operation of the Piezo-electric pressure transducers.
This again impedes long duration testing of PDEs
despite having “low pressure” water cooling.

•Some of the DDT devices do not have the smooth

undulating nature of a spiral and instead have sharp
corners and recesses that may trap the flow and
impede it, rather than creating spin that is
experienced by flow through a spiral.
•Thus the spiral is still the best DDT enhancing
device.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 32
 Conclusions (1/2)
„ A PDE platform was built with commercial automotive solenoid
valve gas injectors that enabled computerized control of the engine
with simultaneous diagnostics. Pressure measurements were taken
yielding detonation pressures and time of flight velocities.
„ The PDE is able to be run at various frequencies ranging from 0 to
50 Hz. 15 Hz was chosen to be the optimum based on the length of
the tube and the filling rates with the valves.
„ Various DDT enhancing devices were tested and their DDT success
rates were tabulated.
„ Clean tube configuration did in fact bring about DDT and this is
ascribed to good mixing and fuel-oxygen delivery by the solenoid
valves.
„ Of all the DDT enhancing devices tested, Shchelkin spiral is still the
best choice, with blockage ratio of 50% to 55%, and L/D ratio of
more than the 6.6 used for this study.

AERODYNAMICS RESEARCH CENTER

MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 33
 Conclusions (2/2)
„ The grooved tubular devices and the CD throats also yielded
DDT but at reduced success rates and it is thought that their
geometries cause blockage of the flame propagation and tend to
retard the flow.
„ Therefore, they can be improved to build more robust, longer
lasting DDT devices.
„ CD nozzle throats in the PDE tend to decouple detonation
waves, rather than boost them.
„ Variations in the Von-Neumann pressure peaks observed were
due to the low sampling rate of 240 kHz causing aliasing of the
signals. A minimum of 1 MHz sampling rate is recommended
for detonation or shock studies.
„ Pressurized (and treated) water cooling or active cooling has to
be introduced to increase the run time beyond the 10 to 20
seconds of operation achieved so far.
AERODYNAMICS RESEARCH CENTER
MECHANICAL AND AEROSPACE
08/03/2007 16:39 ENGINEERING DEPARTMENT 34