Simulation of premixed flames with FDS

Application to the hot smoke testing system Izar

Blond Hernández, Juan José
Basler & Hofmann AG
Rios, Oriol
Centre for Technological Risk Studies (CERTEC). Universitat Politècnica de Catalunya Barcelona, Spain

Introduction

Don't worry, this section is not long…

  • This work is based on the hot smoke testing system, Izar
  • It was developed by the swiss company Basler & Hofmann AG
  • A premixed combustion is the source of energy for the test
  • The results presented in this work are part of its development process
    • The entire work was presented as master thesis within the IMFSE program in 2013

Before we begin, what is Izar?

I am very happy with your question…

Izar is a system with… a gas burner

150416-3871

… a gas supply system…

130815-6609

… and some fog generators

Our clients wanted to see the smoke,… and my boss too 151117-9498

It sounds interesting, but what makes Izar something special?

Well Izar is definitely a cool machine…

  • It doesn't generate soot
    • Nobody likes to clean…
  • Live control of the HRR
    • It is possible to follow fire curves like the t2-curve
  • A validated FDS model is available

(Oh, a validated FDS for a testing system? It seems interesting, I will keep attending what this guy is saying…)

How is that you decided to combine FDS with Izar?

It is not the best idea to do a smoke test an burn the room… I needed to figure out how to avoid it

Yes, that is definitely a good reason

  • We do the test in rooms in use or just before commissioning
  • The room limiting factor must be considered (sprinklers, lights, protected ceilings)

Calculate the plume temperatures originated by Izar

  • The temperature is one of the main factors to design the smoke tests
  • Realistic test = high temperatures
  • Necessary temperature data from Izar for different powers and room heights

How to calculate temperatures from premixed flames?

  • The combustion efficiency is the most important factor for premixed flames
  • What happens afterward? –> Not enough studied

A new framework is developed with FDS to calculate the plume temperature

And how did you create the FDS model for Izar?

  • The combustion itself is “not important”, we want to model the plume
    • The combustion will not be modeled
  • We model the combustion products which conform the plume, the “smoke”
  • Three initial factors
    • Species
    • Temperature
    • Origin surface

The initial species (1/2)

  • The combustion takes place under stoichiometric conditions
  • Well known which the combustion products are

C3H8 + 5 O2 + 5⋅3.76 N2 = 3 CO2 + 4 H2O + 5⋅3.76 N2

We define the mass ratio fuel/product

Specie Molecular weight Amount of products Mass ratio
(g) (g)
C3H4 40 - -
CO2 44 132 3.3
H2O 18 72 1.8
N2 28 526.40 13.16

The initial species (2/2)

  • We calculate the necessary mass flow rate for a desired HRR with the fuel heat of combustion
  • Example for 500 kW
    • Propane heat of combustion: 46 kJ/g
    • Mass flow to achieve 500 kW
      500 kW / 46 kJ/g = 10.85 g/s
Specie Amount of fuel Mass ratio Amount of products
(g/s) (g/s)
C3H4 10.85 - -
CO2 - 3.3 35.80
H2O - 1.8 19.53
N2 - 13.16 142.78

Proper initial values for the subsequent system mass balance

The initial temperature

  • The initial temperature of the gas is related with the flame temperature
  • The combustion takes places under stoichiometric conditions
  • The adiabatic combustion temperature characterizes the flame temperature
  • For our case: Flame temperature = 1.995 °C (Propane adiabatic temperature)

Grid mesh and geometry

  • The mesh definition is a critical value in a FDS model
  • The combustion product must be introduced in kg/m2⋅s
  • The combustion surface defines the initial moment in the system
  • Our system has a geometry of 122 cm (length) x 17 cm (width)
  • A 6.5 cm cell size was chosen after a sensitivity analysis

Radiation

  • High combustion efficiency –> low radiation loses
  • The radiation fraction can be calculated considering the partial pressures
  • The radiation loses of the system are around 3 %

Well you can find the equation in my paper…

radiacion y mano

Validation of the FDS model

  • You are probably thinking
    • Nice method
    • The theory is interesting, but I want to known if it works
    • Does he have more cool pictures?
  • Well, let's show the result
    • And yes there are some cool photos

Slice Data

Temperature 0_1 MW

Temperature 1_2 MW

Let's begin with the plume in FDS

  • Using the described initial inputs, Izar was modeled in FDS
  • A grid was programmed to measure the temperature
    • Temperature sensor every 20 cm in the X and Y axis
    • Repeated each 19.5 cm in the Z-Axis
  • This way of simulating the system avoid possible uncertainties related with the combustion

FDS grid

Input values in FDS…

  • The gas burner is rectangle 130 cm (200 cells) x 19.5 cm (3 cells)
  • The amount of combustion products is in kg/m2⋅s
  • The boundaries are open –> “free plume”

input pyrosim

… and it looks like that

imagen 800 kw

The next step is to validate the results

And that means?

  • Switch on Izar
  • Do some real fire
  • Measure the temperatures in the plume

Grid sensor full scale

Therefore, some action out of the office

130815-6499

And it turns out that the results match at 1.200 kW…

Temperatures plume_1200

… but also with Izar working at 400 kW…

temperatures_plume_400 kw

… or just working at 100 kW

temperatures_plume_100 kw

Next step, a full scale test

  • The plume model works
  • Therefore, it should be possible to model a real scale
  • Is it really possible to simulate that ?

Full scale test

130815-6504

The full scale test

  • We measured the temperatures within the smoke layer
  • We carried out tests at different powers


Manegg measure sensor

We modeled the full scale test with our FDS model…

bild_smokeview_manegg

… and these are the results

alle sensor Manegg

I know that you cannot see too much here…

here three examples

Sensoren Manegg

they look definitely similar

temperature_Manegg

(a)

130815-6597

(b)

Conclusions

  • The methodology efficient tool to model the system Izar
  • The main inputs are:
    • Combustion products
    • Flame temperature
    • Combustion region geometry
  • FDS resolves properly the turbulence and entrainment around the plume
    • The centerline plume temperatures confirm this point

Future work

  • Different configurations
    • Different fuels
    • Different geometries
  • Validate the model in tunnels
  • Development of plume equations
  • Test and validate new premixed burning submodel

The future of the FDS Simulations?

  • The model can be used to validate FDS geometries “a priori”
    • We can carry out a test with Izar
    • We take the necessary measures
    • We use the validated FDS model from Izar to calibrate the simulation
    • We program the design fire
  • This way we reduce the uncertainties related with the geometry
  • Reduce the safety factors

  • Optimize the smoke extraction system

Thanks for your attention

150416-3880