Modelling the thermal desorption soil treatment
Haemers Technologies is a leading company in the thermal desorption soil remediation industry with more than 25 years of experience in this field and a commitment to being at the forefront of technology. To do this, Haemers Technologies has focused most of its activity on inventing, improving and applying thermal desorption technologies all over the world, where we work with local partners.
The Smart Burners technology developed by Haemers Technologies is a thermal conduction desorption method developed in-house and patented. It consists of a system of burners and pipes inserted in the soil through which hot gas circulates. The soil is heated by conduction until a certain temperature is reached, allowing the pollutant to volatilise (typically between 200 and 250°C for hydrocarbons). The polluted gases generated in the soil are extracted by suction before being transported to a vapour treatment unit. In the case of hydrocarbon pollution, the polluted gases can be re-injected into the burner as an energy source. This method, called “reburn”, allows a significant reduction in the burner’s gas consumption.
The research and development projects within Haemers Technologies® are perfectly in line with the company’s strategy to improve the competitive position of the technology in the soil remediation market. The R&D department has a key role to play and is working on several projects to maximise the energy efficiency of the technology, thereby minimising the treatment time and therefore the cost.
One of the projects within the R&D department, called RSIM, consists of modelling the thermal treatment of soil using ANSYS FLUENT software. Previously, only very basic tools were used to understand the transport phenomena in the soils to be treated and it was therefore not possible to accurately determine the process control parameters. This led to the use of large safety factors to ensure that the remediation objectives were met. However, these safety margins are a source of high energy and time losses that undermine the competitiveness of the technology. Modelling is therefore essential in order to gain a better understanding of the phenomena that take place in the soil during thermal treatment. Modelling downstream of a worksite also helps to optimise designs and provide operational data to the site teams in order to control the treatment in the most optimal way according to the site conditions.
This large-scale project is divided into several stages to achieve its goals. In broad terms, the modelling will first take into account the properties of the soil due to the presence of moisture and pollutants in the soil and to understand and quantify the physico-chemical phenomena such as evaporation and pyrolysis of the pollutants that occur during the treatment. In a second step, the modelling will turn to our treatment facilities by simulating the combustion that occurs in a Smart Burner and by simulating the phenomena that occur in a vapour treatment unit.
In this report, the first major advance of the project is outlined. This is the consideration of the properties of a soil to be treated as a function of its initial moisture content.
The first thing that was done was to analyse what the ANSYS FLUENT tool can do. Indeed, some materials are already predefined in the program, so maybe a wet soil too? It was necessary to analyse the equation models that the tool makes available to the user, such as the energy, turbulence and momentum equations, etc.
In fact, the simulation software, ANSYS FLUENT, used makes it possible to consider soil properties that vary with temperature, vary over time or are simply a constant value. However, during our numerous projects, we have been able to observe that the thermal properties of the soil vary as a function of humidity as shown in Figure 2, which are measurements taken at a worksite.
ANSYS FLUENT allows this factor to be considered but requires the resolution of several equation models taking into account the phases change of water and therefore requires too many time-consuming calculations. It was decided to develop a formula, written in the C programming language, for these properties that takes into account the variation in humidity over time under the effect of the heat created by the heating elements. This formula initially considers a porous medium consisting of a type of soil initially containing a certain percentage of water. It also considers numerically the energy consumed by the evaporation of this quantity of water. Thanks to this proprietary formula, employing the “used defined function (UDF)” option in FLUENT, a simple model can be used which only solves the energy equation developed below.
Once this formula developed, its robustness had to be verified by comparing the results obtained by modelling with those measured on site. Temperature is a parameter that can be measured on site over time and is a major parameter of thermal technologies. Therefore, the on-site tests focused on this parameter: thermocouples (temperature sensors) were inserted (TG4, TG5 and TG6 on Figure 3) into the ground every 15 cm from a heating tube (H5-6 on Figure 3) to the coldest point (centre of the triangle formed by this heating tube (T10 on Figure 3)) and two surrounding tubes (H5-7 and H6-5 on Figure 3). In the simulation, this area was drawn in 2D and the position of the thermocouples was recorded in the software so that the layout of the site and the simulation are identical.
By juxtaposing the measured temperature curves on the worksite over time with those obtained numerically by the simulation, several observations were made.
First, the formula developed is functional: the general behaviour of the curve is similar to that observed on the worksite as shown in the Figure 4 and Figure 5.
Second, the soil heterogeneity is difficult to exploit by simulation but could highlight problems on the worksite such as a poorly operated lowering of the water table.
Indeed, as shown in the Figure 6, the temperature has never been able to exceed 100°C due to a perpetual inflow of water due to the known presence of a water table at this depth.
Finally, it appears that what was considered to be the coldest point on this zone – going from the heating tube to the centre of the triangle – was not necessarily the coldest point. As a matter of fact, the simulation revealed the influence of the two other surrounding tubes on this central point.
This comparison study also provided an understanding of heat transfer in the soil. When the temperature profile between a hot point (wall tubes temperature) and the corresponding cold point is known, it is possible to carry out a thermography at a given time. Site measurements showed that the temperature curve tends towards a parabolic profile. It could be observed that this profile became stable in time. These observations have also been obtained by modelling, meaning that for each future site, simulation could be used to provide the profile equation between the different hot and cold points, and therefore generate thermography as close as possible to reality.
Only a 2D model has been tested here, in the future we will work on 3D models taking into account any soil properties that may exist such as the effect of groundwater on the treatment. We will also take into account the temperature profile that may exist along our heating tubes by modelling the combustion that occurs downstream.
In the future, it will be possible to combine Fluent with the monitoring tool in order to have the most realistic thermographies possible. Indeed, only the temperature of certain cold spots is measured on site, so the monitoring tool extends its measurements to the other surrounding spots in an arbitrary way.