Discussion Forum

Hi there,

I have 6 years of actual temperature data gathered from instrument probes placed within a landfill. There are 5 probes spaced out vertically within the landfill.

I am using a 1D time dependent COMSOL model to predict these landfill temperatures. I use Point Probe Expressions in order to track the temperature at these 5 locations. In the Heat Transfer in Solids physics, I specify factors such as initial temperature, ambient temperature, solar radiation, density, thermal conductivity, heat capacity, heat transfer coefficient, etc., to predict the temperature with time at 5 different locations along the 1D line (which represents a vertical column of the landfill). In the model, I use a best guess approach for specifying the heat capacity and the heat transfer coefficient (that is, I use values from literature because I do not have actual values for the heat capacity and heat transfer coefficient for the municipal solid waste in our landfill).

I would like to use the Optimization tool in order to help predict the heat capacity and heat transfer coefficient of the waste. I would like COMSOL to find the best fit values for heat capacity and heat transfer coefficients by minimizing the error between my actual temperature data, and the point probe expression temperature.

I have added a Global Least Squares Objective in and uploaded a comma separated file with my field data (column 1 = time, column 2 = temperature). However, I am having trouble with the rest. How do I specify that I want COMSOL to take the least squared error between my actual data and the point probe expression temperature? And how can I specify that I want it to do so by changing the variables for heat capacity and heat transfer coefficient?

Any suggestions would be greatly appreciated!

Courtney