Polydisperse particle-laden jet flows at high Stokes number

Particle-laden turbulent flows are extremely complex due to the intrinsic chaotic nature of the carrier-phase turbulence and the random character of the dispersed-phase distribution. In many applications, however, additional complexity often arises from interactions of such flows with other complex phenomena, such as evaporation, melting, and chemical reactions. Although a profound insight into many features of particle-laden turbulent flows has been gained from previous laboratory and numerical studies, characterization of such flows still remains elusive. This lack of characterization is partly because of the aforementioned, intrinsic complexity of the system and partly because of a large number of controlling parameters, namely, Reynolds number, Froude number, particle Stokes number, and mass loading.

The main objective of this project is to advance our current knowledge of relevant features of polydisperse particle-laden turbulent flows and to synthesize this new knowledge in the optimization of additive manufacturing processes. In this respect, the ability to accurately predict the carrier- and dispersed-phase behavior is of significant essence. For this purpose, we employ numerical simulations and laboratory experiments in a cooperation with the Bremen Institute for Applied Beam Technology (BIAS) to investigate polydisperse particle-laden jets with the focus on the parameter space that is relevant for additive manufacturing applications. The main characteristic of this parameter space is the moderate Reynolds number, the large particle Stokes number, the large mass loading, and the large Froude number.
For the laboratory experiments, BIAS utilizes the particle image velocimetry to accurately measure the distribution and the velocity of particles at different levels away from the nozzle exit. For the numerical simulations, we employ the Eulerian-Lagrangian method, in which the carrier phase is described as a continuum and governed by the compressible Navier–Stokes equations and the dispersed phase is described in a Lagrangian way with equations of motion for each particle. In addition, the soft-sphere model is used to describe particle-particle interactions. The numerical code has been already implemented in the open source CFD software package OpenFOAM. However, an extensive validation study using previous numerical and experimental literature is performed to assess the capability and the accuracy of the code.
In the first phase of this research, a set of numerical and laboratory experiments is designed to systematically cover the parameter space of interest for polydisperse particle-laden free jets. An analysis of the momentum and the turbulence kinetic energy budgets is performed to rationalize the results. In the second phase, deposition of particles issuing from discrete coaxial nozzles on a flat surface and the melting process through a laser beam are investigated. An optimization of several aspects of the additive manufacturing, including positions of nozzles and the laser and the mass flow of both phases, is of particular interest.

Heat conduction and ventilation in a moon habitat laboratory module

Human spaceflight demands systems that provide a livable environment for humans i.e. in habitats for long duration missions on other planets. This challenge involves the design of several subsystems of the life support system according to comfort criteria, including air management and ventilation of supply air. To ensure a comfortable room climate with fresh air while being able to remove waste heat of the habitat's systems are main goals of the ventilation system.
In this project a ventilation design for a habitat laboratory is created. The ventilation flow performance in different configurations and boundary conditions is evaluated via numerical simulation for an incompressible transient turbulent flow with heat transfer. The design of MaMBA, the Moon and Mars Base Analog, is used as an example geometry, consisting of an octadecagonal room with racks in alternating heights and stairs as obstacles. The maximum heat load (8.8 kW) produced by scientific instruments, habitat systems and humans inside the room is modelled. Since testing of a cooling system with the room ventilation being the only cooling loop does not provide an acceptable solution, a secondary cooling loop is proposed. It absorbs the heat of the electrical devices like the scientific instruments, that are located inside the racks, and releases heated air at the rack's bottom in the direction of the room's exhaust vents. Comfort criteria according to literature values are set to ensure a good mixing of the supply and ambient air with a low probability of drafts. In combination with the secondary rack ventilation, the cooling loops can fulfil most of the comfort criteria. It can be shown that a secondary cooling loop for the rack system at maximum heat load is required and should therefore be integrated in the design of the habitat's ventilation system as a minimum configuration.

AulaPRO: Adaptive Powder Feed for Additive Manufacturing Processes

Due to the extensive developments in the past decades, numerous different system technologies for powder feed or powder nozzles for laser powder build-up welding have established themselves on the market. In general, the powder feed systems are divided into lateral and coaxial systems. In the case of lateral nozzles, the powder is supplied from a powder nozzle and only from one direction, so-called off-axis nozzles. The process is therefore direction-dependent. The nozzle shape can be designed as a round tube or in the form of a wide slot nozzle. Lateral powder feed systems are a widespread, established technology for powder feed when coating and repairing geometrically simple components such as rollers, shafts, etc. using LMD.
With a powder feed arranged co-axially to the laser beam, the LMD process can be carried out in any direction. The coaxial nozzles are therefore often used for additive manufacturing. In these systems, the powder is supplied from several directions co-axially to the laser radiation in relation to the lateral nozzles. With the discrete powder feed, the powder is fed into the process from several co-axially arranged individual nozzles. In the case of the continuous coaxial nozzle, we speak of annular gap nozzles. The powder focus is located at a fixed distance from the nozzle opening to the substrate surface. In near-net-shape additive manufacturing using laser powder build-up welding, the precision can be increased using very small weld beads. Very large weld beads are used to maximize build-up rates.

Currently, processing heads are used in which the powder focus is constant and the diameter of the laser spot is adjusted to vary the bead geometry in addition to the powder mass flow. However, a constant powder focus leads to a reduction in the degree of powder utilization when using a small laser spot and, on the other hand, poor precision is achieved when welding with a large laser spot. The aim of the planned research project is therefore to adapt the powder focus diameter to the weld bead geometry through the adaptive positioning of individual powder nozzles to increase the degree of powder utilization and precision in near-net-shape additive manufacturing using laser powder cladding. The adaptive positioning of the powder nozzles to one another, to the laser beam and to the substrate surface results in interactions and challenges that need to be investigated in the planned research project. On the one hand, the adjustment of the nozzle positions leads to a changed path of the particles in the laser beam and, on the other hand, the adjustment of the powder mass flow influences the particle speed, so that the interaction time between laser beam and powder can be too short and the particles are not sufficiently melted. There is the hypothesis that the interaction times can be influenced in such a way that a complete melting of the powder is ensured through suitable angle of incidence of the powder nozzles depending on the nozzle position and adapted conveying gas quantities depending on the powder mass flows.

Modelling electric Discharges/Arc with PIC methods

The research on thermal electric arcjet thrusters has been the subject of several projects of the Thermofluid Dynamics division at ZARM ranging from experimental investigation to the numeric modeling. One challenging aspect in the modeling of these type of thrusters always has been the description of the arc, which is responsible for the heating of the propellant through the process of Joule heating. Fundamentally, electrons originating from the high current arc burning in the partly ionized propellant, collide with atoms and thereby increasing the thermal energy of the neutral species.
In previous modeling approaches based on differential equations of the propellant gas the electric power has been modeled using basic physical principles and empirically acquired data. For a general description of the converted electric power in the thruster a more detailed look at the arc is required. This is achieved by modeling the arc using kinetic methods more precisely with the help of the Particle-In-Cell method frequently used in plasma research. The kinetic description of gases is generally free of assumption and allows in the case of plasma for a more detailed look on plasma sheaths occurring in the vicinity of boundaries. These regions of non-equilibrium state in the plasma play an important role in the behavior of the electric arc. Therefore the goal of this project is the development of an numeric tool based on the Particle-In-Cell method which is fitted for the description of electric arcs as they occur in thermal electric arcjet thrusters. In addition insights into the electric arc’s behavior gained by this tool are ultimately intended to create an improved and more generalized model for the description of electric arcs in numeric approaches based differential equations.

Former Projects

Grindball: Hydraulic Driven Grinding Tool

This project is a cooperation work together with the laboratory for micro cutting (LFM) and institute for electric drives (IALB) at the University of Bremen. Central idea of the projects is the development of a microscopic axis-less grinding tool. The central element of the grinding tool is a microscopic ball, which is driven by hydraulic forces of a surrounding liquid bed. Additionally the ferromagnetic ball position is controlled by an electromagnetic field. Because of the high liquid mass flux, needed by the hydraulic driving system, the Reynolds number of the flow inside the sphere gap is increasing over the critical Reynolds number of transient turbulent flow. Therefore the hydraulic shear forces depend on the thickness of the turbulent shear layer at the ball surface. Primary topic of this research project is the fluid/structure interaction of transient turbulent shear flows in an electromagnetic controlled positioning system. Changing the position of the grinding ball the flow character of the surrounding liquid flow changes depending on the minimum gap size between the ball and its bed.

The tool consists of an abrading sphere which is mounted inside a spherical gap using a magnetic bearing. The sphere is set in rotation pneumatically with air which is led into the gap via a duct. The optimal diameter for the duct is chosen as large as possible while it should not greatly exceed the height of the spherical gap it leads into.
Standard volumetric flow rates of up to h result in a stable sub-sonic flow using air. Turbulent behaviour is observed inside the spherical gap which justifies use of an LES model. During the simulation, pressure and momentum forces acting on the sphere are determined. Since the rate of rotation is given asa fixed boundary condition, tangential force transferred on to the sphere is equivalent to the force that would be available to the abrasion process in practice. Hence, conducting a series of simulations with varying flow rates and rotation frequencies aids in determining the momentum transfer and its relation to standard volumetric flow rate, and rotation frequency. Simulations show a linear correlation between momentum transfer and rotation frequency. Momentum transfer displays a quadratic dependence on flow rate fora stationary sphere, and the idle rotation frequency (rotation frequency for ) can be expressed as a root function of the flow rate.

Thermo-Electric Emission for Plasma Processes

Just as in an arc welder, a high-current electric arc is struck between the anode and cathode. As the cathode heats up, it emits electrons, which collide with and ionize a propellant gas to create plasma in MPD thruster. For gases, there are many circumstances where the electron temperature T_e differs from the heavy particle ( ions and neutral atoms) temperature T_h. The electrical conductivity and other coefficients in the equations of motion depend on T_e. It's therefore necessary in such circumstances to include T_e among the variables which define the state of gas, and to employ an additional equation which governs the behavior of the electron temperature.
The ZARMmpdMultiRegionFoam is a solver inherited from the Built-in heat transfer solver ChtMultiregionFoam in open source CFD Software OpenFoam, which will take the temperature deviation from LTE (Local Temperature Equilibrium) into account. Through this solver we can get the numerical results of electron number density and electric potential distribution along the electrode-plasma space and in turn the electron emission rate from the cathode to the present plasma flow is determined.

Computational Arc-Jet Modelling

Total fuel mass is one of the main economical and technical restrictions while designing space propulsion systems. Given the high costs related to the transport of mass into space, the total fuel mass necessary for accomplishment of the mission should be minimized. An optimum “thrust/fuel consumption ratio” demands the maximization of the gas exit velocity for a given mass flow. This can be achieved through the increase of propellant temperature and pressure in the propulsion system, which results in higher values for the speed of sound at the “throat” of the de Laval nozzle.

In the project PlasmAccelerator arcjet propulsion systems are investigated. Higher gas exit velocities are achieved through the heating of a micro gas flow by an electric arc inside the sub-sonic region of the propulsion system. The electrical arc induces a partial ionization of the propellant gas, thus creating a plasma flow through the nozzle. The behavior of electric arc and propellant gas is modeled by custom solvers based on the open source CFD software package OpenFOAM. The numerical results are validated via a comparison with data from experimental rigs at ZARM. The project PlasmAccelerator aims at the development and validation of numerical models for the electric arc and the partial ionized gas inside an arcjet propulsion system. The project also focuses on the optimization of the arcjet propulsion system at ZARM based on the numerical results.

Modelling Astrophysical Hydrodynamics

One of the most amazing applications of magneto fluid dynamics are astrophysical approximations. High temperatures are responsible for fusion and ionization processes in space. Furthermore, the motion of charged particles in space is part of its astrophysical description. The motion of charged particles near planets with a magnetic core or inside accretion discs in the environment of a black hole is the result from attaching forces of other particles and is influenced by electric and magnetic fields.
The velocity distribution is comparable with a fluid motion in a way similar to the relation between microscopic molecular flow and the corresponding macroscopic fluid flow. Therefore, characteristic non-dimensional numbers from fluid mechanics -like the Reynolds, Knudsen or Nusselt number- obtain an astrophysical meaning. Through this approach, the motion of charged particles is comparable with the hydrodynamics of electrically charged fluids. As a result, unsteady magnetic fields can be described by compressible magneto-hydrodynamic transport equations.

According to the parallels with fluid mechanics, the consideration of high shear rates (Reynolds number) and dilute particle flow (Knudsen number) leads to a compressible description of a turbulent flow inside an accretion disc. In addition, all turbulent effects influence and at the same time are influenced by the magnetic field near a black hole. Through this approximation, the astrophysical examination of energy fluxes (Nusselt number) can be described by turbulent, compressible and magneto-fluid dynamics.

Ionization and Thermo-electric Thruster Systems

When the difference in electric potential between electrodes with different charge is high enough, an electric discharge through the fluid medium separating the electrodes occurs. The resulting electric flux is associated with a local motion of electrons, the only mobile charge carrier inside the solid electrodes. As a result from the motion of electrons and the collisions and resistance forces inside the conducting electrodes, a local temperature increase occurs. The resistance against electron motion in fluid media is a result from similar forces. The biggest difference lies in the mobility of the charge carrier phase molecules. As in the solid electrodes, the electric conduction through a fluid phase leads to a local heating. Furthermore, higher temperature values increase the mean free path of the particles and reduces the electric resistance of the fluid. This region is characterized by the appearance of an electric arc.

The conductivity of electron inside the different charge conductors is described by the field of electrodynamics. The local change in the electron velocity in the interface region of the conductors is a research field of special interest. The characteristics of this region are responsible for the acceleration of electrons towards the region with lower electric resistance. The regions where very high absolute acceleration values are reached are called „Sheath layer“. In the „Sheath layer“, the macroscopic discharge reaches its maximum.
The modeling of the electron transport inside fluid media deals with the collisions between electrons and the neutral particles of the carrier phase. If the kinetic energy of the electrons exceeds the ionization energy of the carrier phase, the ionization process is initiated and plasma is generated.

Computational Modelling of Shock Waves in Transsonic Flow Regime

When developing machines and facilities with optimization regarding fluid dynamic behavior, the form of the optimization tools results from the operating conditions of the studied flow (laminar, turbulent, incompressible, trans-sonic, hyper-sonic, molecular, continuous etc.). If mixed conditions are present, the interactions between the sub-models has to be investigated and modeled through coupled transport equations.
The modeling of the behavior of thermo-fluid dynamic systems is achieved through computational methods. For this, the fluid domain of a complex technical flow is discretized in single cells. Furthermore, previously developed transport equations are discretized into a system of discrete differential equations. After linearization large sparse systems of linear equations are the result. The fluid and thermodynamic properties for each single cell of the discretized volume are then computed using modern solving algorithms. The computational engineering tool for the study and analysis of these problems is called CFD (Computational Fluid Dynamics).

Further technical applications such as sprays and reacting flows are characterized by their time-dependent behavior. In order to describe them, macro-scale laws are developed based on micro-scale investigations and the continuum hypothesis (usually used for rheological modeling). This way, properties associated with molecular motion are transformed into models of reactive flows or continuum mechanics. The description of molecular flows through continuous models do not satisfy the diffusive and convective behavior of dilute gas flows in micro-channels. In this case, the non-continuous motion of the dilute gas flow is characterized by the very high values of the Knudsen number.

Heat Transfer and Convective Flows

The computational description of evaporation processes requires the calculation of the heat and mass transfer exchange rate over the liquid/gas interface. The most influential parameter for the exchange rates is the relative humidity of the gas phase. The highest relative humidity gradient occurs at temperature levels slightly below the boiling point. As a result, these temperatures have to be considered when computing the dispersed phase evaporation driven by temperature increase.

The dimensionless numbers characterizing the heat and mass exchange process are the Nusselt number Nu and the Sherwood number Sh. Several approaches based on the correlation model of Ranz and Marschall, have been developed in order to model these parameters. The evaporation rate model of Abramzon and Sirignano additionally considers the latent heat flux of the evaporated liquid leaving the droplet. The correct definition of the gas phase humidity requires the mass ratio of the liquid's vapor in the gas phase Y, which is influenced by convective, conductive, turbulent and thermal diffusive effects, to be computed from an appropriate transport equation in addition to the one governing the temperature field T.
For all subcritical and supercritical thermo fluid-dynamic processes the second law of thermodynamics, non-negative production of entropy, has to be satisfied. Models describing thermo-fluid dynamic processes, which comply with the second law of thermodynamics, are called thermodynamic consistent. When describing macroscopic thermodynamic processes with stochastic approaches of molecular motion, the thermodynamic consistence must be proven by the second-law analysis (SLA). The resulting consistent entropy production, the comparison between macroscopic physical values and the molecular diffusion modeling are some of the main advantages of this approach.

Ion Flow

An electric propulsion system has been assembled on the ZARM institute. This propulsion system is actuated with an electrical power up to 500W and inert gases as propellants. During the experiments the pressure inside the test chamber is between 0.1 and 1 Pa. Still at the moment the runtime is limited to only a few seconds, because of the electric heating.
The results of the experiments are used as a referents for the development of a numerically simulation solver. This solver based on the compressible Navier-Stokes-Equations and includes a magnetic volume force field, to describe the behavior of electric charged particles in thin gases. The gases Argon and Neon were ionized inside the cyclic gap with different current magnitudes. With a larger electric current (30-50mA) the light emission and the ion density increase in the plasma flow.


Hurricanes are effected by turbulent natural convection flows, which are caused by solar radiation. To understand the hydrodynamic stability of these flows, especially in terms of the Coriolis acceleration of the earth rotation, the first step is an investigation of a simplified convection problem. Therefore, a suitable experimental setup is compared with numerical simulations. Convective flows in the atmosphere are drive by cold air streams over warmer layers on heated ground. The forced convection moves heated air in higher regionsof the atmosphere.
With the coriolis acceleration depending on the rotation speed of the earth the heated columns are forced to start a rotation. This rotation energy is dissipated by shear forces in tangential layers of the rotation. During this dissipation process inside the turbulent shear layer larger vortices (eddies) are generated and moved on the cyclic motion. These eddies are breaking in several smaller vortices. This vortex splitting process in turbulent shear layers is called energy cascade. During this project the stability of these convection driven rotational flows has to be investigated and analyzed. The complete energetic process is displayed from the thermal-driven convection to the turbulent dissipation in the radial shear layer.

Analyzing the flow inside an experimental set-up inside a capsule, which is moved on a cyclic way inside a centrifuge, the relative accelerations correspond with accelerations inside the atmosphere. By this way macroscopic phenomena are investigated in scaled experimental set-ups in laboratory. The thermo-lift module allows convective flows in relative oriented acceleration fields. This method is used controlling forced rotation flows and turbulent dissipation effects is the radial shear layer. At this time the facility is used for the project “TurboTherm” for analyzing stability and interaction of turbulence production and dissipation coupled with variable fluid density.

Autarc Lunar Infrastructures

It is proposed to establish a Collaborative Research Centre (SFB) with the title “Autonomous Lunar Infrastructures (ALI)” at University of Bremen with several cooperating partners close by. Bremen is an important site in Germany where space related activities are concentrated with more than 1500 engineers and scientists working in the fields of development, production, operation and utilisation of orbital and exploration systems, satellites and launchers.
Space companies as EADS Astrium, OHB-System AG and a number of university institutions like ZARM together with institutes from DLR and DFKI characterise the consolidation and clustering of space competences in this region. A further interesting partner for cooperation close by is AWI in Bremerhaven where polar and marine research is performed.

It is the intention of the proposed Collaborative Research Centre (SFB) in Bremen to strongly interact with the relevant local industry and scientific community, exchange ideas, exploit synergies and prepare future activities by strengthening all partners and players, giving them a decisive advantage over other European competitors. With regard to the above mentioned scientific tasks, a set of relevant research projects is defined below and thematically combined into five project groups:

    A) Base Construction and Materials
    B) Mobile Infrastructures
    C) Environment and Resources Prospecting/Processing
    D) Energy Supply & Night-time Survival
    E) Communication/Navigation

Excluded from the considerations are:

  • transport between Earth and Moon
  • medical and sociological aspects of the crew (including life support)
  • nuclear energy
  • architecture

  • Within the scope of the planned work key technologies will be developed in each of the project groups up to the level of prototypes (breadboard or down-scaled demonstration hardware) which will be tested in cooperation with AWI in later funding periods under demanding conditions in the Antarctic. Besides the low temperatures and the increased UV radiation due to ozone depletion in the southern hemisphere, the Antarctic winter over nearly half a year without sunlight (partly with temperatures as low as -90°C) serves as an unrivalled challenge for every prototype, which can hardly be found at other places on the earth.

    Convectice Cooling

    Thermally driven flows are described as a turbulent compressible Rayleigh-B´enard problem. For the numerical discretization of the Large-Eddy-Simulation the flow is separated in large and small scales, the so-called sub-grid scales. The large scales are solved directly on the discretisation grid, while the sub-grid scales are modelled with an applicable turbulence model.

    A rectangular, air-filled container is chosen as test case. The walls are smooth and the vertical ones are heated homogeneously with a constant temperature difference. For the lateral walls different boundary conditions are investigated. Due to a non-slip-condition the velocity field at the walls is zero in all cases. The Rayleigh-number Ra=1.58e9 and the Prandtl-number Pr=0.71 are always constant. The vertical positioning of the heated plates parallel to the direction of gravity effects a steady state flow. Two recirculations zones in the vincinity of the temperature-controlled walls can be seen. A variation of the thermal boundary layer becomes apparent. The profiles of the temperature and the vertical velocity between the temperatured walls are plotted against the experimental data.
    The thermal driven flow in the container is effected by the decreasing density and the resulting pressure gradient in direction of gravity. The intensity of the turbulence increases with the shear flow interaction of hot lift streams and cold sinking streams. Singular hot gas plugs break through the cool sinking flows. This unsteady behaviour dominates the vertical heat flux. This is illustrated by the instantaneous temperature iso-surfaces for 1.92e8. The plots of the mean horizontal and vertical velocity of the xy-midplane at one time step show the typical convection cells which are developed between the temperature-controlled walls.

    Debris Flow

    This set-up forces a debris flow at the ground of the cylinder. Additionally the interface between debris and liquid media is destabilized. The new steady shifting form depends on debris material and the mean shear stress at the surface of the granular bed. Inside the debris flow cylinder a steady shear flow controlled by a rotating disk is built up. The shear stress inside the layer between forcing “upper wall” and the granular bed induces interaction forces between the moved fluid and the solid particles at the ground.
    Modeling fluid flow and granular motion is made with continuum mechanical transport equations. The granular media are moved by the shear stress attaching at the bed surface. The motion is comparable with the shear motion of continuous media. The particle diffusion works similar to the deformation of a non-Newtonian, incompressible fluid. Because of the interaction forces at the interface the granular bed surface is deformed and holds the structure also after stopping the flow, because of the plastic components of the modeled material properties of the quasi-continuous dispersed medium.

    Double Cone

    The subject of interest is the validation of a 3-D numerical computer model of a hypersonic flow around double cone geometry. The double cone geometry represent a generic space vehicle which enters an atmosphere with very high velocity. This leads to a complex flow phenomena around the space vehicle. In this poster results of the investigation are presented.
    The experimental data are measured by different double-cone geometries which are mounted inside a hypersonic wind-tunnel. During the experiments the Mach number is equal to 9. Three different geometries and four different operating conditions are the subject of this study. Because of the short test period of less then 200ms a measurement of temperatures and local velocity’s during test is not possible. Therefor the computational model is used.

    Space Blast Pipe

    The objective of this study is to prove the feasibility of the suggested technical concept and validate the elementary performance calculations. Besides gas- and thermodynamic analyses, this includes basic examinations on some options for the supporting structure; the piston and the outlet mechanisms as well as the constructional constraints of the tube itself. The unsteady flow and the time-dependent pressure distribution inside the blast pipe will be modelled, computed and simulated. This is necessary to predict the required power and to define the different phases of the blasting process.
    Here it is assumed that the pipe is inclined with an angle α against the horizon. The pressure variation bases on aerodynamic forces, expansion waves and convective influences depending on the aerostatic conditions inside this inclined pipe.

  • During the INITIALISATION PHASE in the beginning the chamber below the piston platform is pressurised until the reference pressure is reached. The pressure over the platform has to decrease on a minimum to reduce the drag during the later acceleration phase. For the de-pressurizing this upper volume during this initialization phase the upper outlet of the pipe has to be closed.
  • In the ACCELARATION PHASE the piston platform is unlocked and accelerated by the difference pressure. The local pressure change inside the blast pipe is predicted by the running shock waves and the aerostatic pressure distribution inside this very high tube. While the pressure “under” the platform decreases because of the expanding volume, hot gas is induced by local gaps in the side walls of the inner volume of the blast pipe. During this process the pressure in the upper chamber increases by the permanent volume reduction.
  • When the pressure in the upper chamber reaches the atmospheric pressure behind the still closed outlet at the top of the pipe, this shot is opened and the ATMOSPHERIC PHASE starts. During this time interval the gas of the upper chamber is pushed into the atmosphere until the platform reaches the outlet. After this point the phase outside the blast pipe starts only with drag and gravitation forces while payload is shot into space.