Problem with liquid temperature readings in SimScape

Hello,

I am trying to understand how the physics of the thermal liquid library works. I have a specific case that I am trying to simulate but SimScape does not appear to be reproducing the physics that I would expect.

Consider the following simple example shown below. If you heat up a closed loop of liquid that is circulating, I would expect that the temperature readings on T1 and T2 (sensors) will be different. My SimScape model correctly shows this and the heat input does deliver a corresponding temperature difference. The problem is that there is no time delay between the two sensors and no feedback effect that you would expect in a closed loop. I have shown a simple graph of the behaviour that would be expected with a closed loop starting at an initial temperature. I would like my model to be able to vary the time shift of the feedback effect depending on the length of the pipe or the flowrate.

At present, SimScape seems to be modelling it all as a single "lumped" volume. I have attached a more complex model that has some dynamic pulsation behaviour and I would expect to see a time shift between the spikes. I have also attached real data that shows this behaviour. Note in the real data graph that there is a time shift between the sensors and also that there are secondary and tertiary oscillations caused by recirculation of the fluid. This dampens out over time.

EDIT: I believe that the effect that I am trying to describe is correctly called a "transport delay". Some previous (unsuccessful) attempts at this problem can be found here.

I can easily implement a variable delay here for reading the temperatures. This will give me the horizontal time shift that I discuss below but it will do nothing to address the feedback and real physics of the model. It only impacts the reading of the signals and not what's truly going on within the pipe.

Transport Delay

How do I implement this transport delay/feedback effect in SimScape?

The above graph is the expected behaviour that I would expect to see and the graph shown below is real experimental data showing the time shift between sensors.

Here is what my SimScape model shows. Notice that there is no time delay at all between the two sensors.

[Andreas]

There is a very small time delay between input and output (5 s). Remember: the pipe block is indeed just one single "lumped volume". Try to discretize for better results.

Maybe this helps a bit: http://www.mathworks.com/matlabcentral/answers/216358-simscape-are-there-ways-to-build-segmented-discretised-models-with-n-as-a-block-mask-parameter