Tomography for large industrial equipment.
Simulation and reconstruction of muons interaction with large industrial equipment for the application of this natural radiation in nondestructive tests.
Muon Systems is a small enterprise founded in 2015. The company develops all necessary hardware and software tools for the application of muon tomography or muography in different sectors like heavy industry, borders security or mining and civil engineering. This technique uses muons, a natural radiation generated at the atmosphere and reaching the earth surface. By studying the interaction of muons with matter, it is possible to reconstruct an image like X-ray radiography.
The objective of this pilot project is to validate by simulation the use of muography to reconstruct the complex inner structure of large industrial equipment.
Muon tomography provides density maps of the inner of objects. To apply the technique it is necessary to measure the muons, a type of natural radiation, before and after they pass through the target. An algorithm takes the measurements and reconstructs the interaction of the muons with the object in a density map. Density maps provide very useful information about the inner of the object, such as wear, accumulation of materials, etc.
Muon Systems is scaling its technology to be able to scan large industrial equipment. The larger and more complex the equipment is, the more difficult is to reconstruct the muon interaction. In addition, the industrial environment is hostile, so we need to simulate both the hardware design and the reconstruction algorithms before working at the facility.
The first step of the simulation is to build the scenario with the design of the equipment structure, its environment and the detection hardware. The second step is to simulate the muon flow. In the case of a 10 m high and 7 m diameter smelting furnace, 15 minutes of exposure requires calculating the interaction of millions of muons within the scenario. This process is repeated for each value of the parameter analyzed. The data generated in this process feed the reconstruction algorithm.
With this pilot project we will reduce the time needed to simulate this type of complex systems and we will test the execution of reconstruction algorithms that are more accurate than the current ones without exceeding the operational timing of the factory.
The project has two lines of work.
In the first part of work we will study the simulator implementation and data generation for a full-size furnace. The simulation will vary a single parameter within the furnace structure, e.g. the furnace wall thickness. This parameter will be discretized in a range and we will generate data equivalent to 24 hours of exposure to natural muon radiation for each value in the range.. We will monitor the time cost of all processes and will compare it with the current generation of data.
In the second line of work we will analyze all the data generated in the simulation. The analysis will compare different reconstruction techniques, monitoring the calculation speed and the resolution of the measurement obtained for each of them. We will evaluate which technique is more efficient to measure the changes inside the furnace.
With this work plan we want to achieve:
- Improvement the speed of simulated data generation for large industrial equipment.
- A tool to measure the efficiency of different reconstruction algorithms on simulated data.
- An analysis environment that executes complex algorithms in optimal times for real field operations.