Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Seeking Help with Capacitive Sensor Modeling in COMSOL
Posted 24 ott 2024, 14:11 GMT-4 Electromagnetics, Geometry, Modeling Tools & Definitions 1 Reply
Please login with a confirmed email address before reporting spam
Hello everyone! I’m working on modeling a capacitive sensor using COMSOL. The sensor I’m studying consists of an FR4 substrate with 15 pairs of copper interdigital electrodes. The measurements I took with the sensor showed capacitance values ranging from 6 pF to 100 pF (using air, water, and ethanol as test media). My goal is to model this sensor using the Finite Element Method (FEM).
Initially, I developed a reduced model, the RVE (which represents only one pair of electrodes). By proportionally multiplying the results, I obtained values relatively close to the measured ones, with errors ranging from 1% to 17%.
Currently, I’m working on a full model that aims to represent all the geometric characteristics of the real sensor. Although I haven’t been able to fully replicate all the feed lines, the model seems fairly accurate. However, the errors are even larger than in the reduced model.
Does anyone have suggestions on how I can improve the model? Is there any way to import the .gerber file used in the sensor’s fabrication directly into COMSOL? Or could this be a mesh adjustment issue?
I’m feeling a bit lost, so any help would be greatly appreciated. I’ll attach some images that illustrate the problem. If anyone is interested, I can also share my current model file.
Attachments:
Please login with a confirmed email address before reporting spam
Hello Savio,
If you have access to the Uncertainty Quantification Module (you can check on that under File>Licensed and Used Products), you may want to start by checking how sensitive your results are to the inputs. My thought here is that if the actual experimental values of the inputs (material properties, dimensions, applied voltage, etc) are each off by a little bit compared to what you typed into your model, the results produced by the model could well be off by the sort of percentages you are seeing.
Best,
Jeff
-------------------Jeff Hiller