hybrid energy storage system for E-cycle

I got a model of hybrid energy storage system with boost converter and single buck-boost converter from mathworks. but i dont want buck-boost converter i want only buck converter. Can someone modifify the simulink model according to that . I want only boost converter and only buck converter... Not buck-boost converter.

[Diana Lemian]

Hello! Can you sen me the equations used for this model or the paper on my email [e-mail]. Thank you!

Answer 1

[Joel Van Sickel]

Hello,

this type of model can be modified to have any type of converter you want, but you will have to do it yourself. Here are some examples of buck converters that you can start with:

https://www.mathworks.com/help/physmod/sps/ug/buck-converter-increased-accuracy-and-simulation-speed-using-interpolation.html

If you need help with the tools, here is some free training: https://www.mathworks.com/learn/tutorials/simulink-onramp.html

Regards,

Joel

[Rajat Powade]

Hi, I devloped the model of buck converter, But the problem is that it shows different results when the powergui mode is change from continuous to discrete . Please if you know anything about it then please help

[Joel Van Sickel]

Ah, ok, that makes sense. In general, I would use variable step with the solver ODE23t as your first simulation. This will typically be the most accurate. When you switch to discrete, if you get different results, it typically means that you need to use a smaller step size. This will likely be 100 times smaller than your switching period, but for some dc/dc converters, you might needs 200 or 400 times smalller (1000 times smaller for lower power dc/dc converters, but that shouldn't be an issue for your application). The issue is that a variable step solver estimates error and adjusts its time step to try and improve simulation accuracy, while the fixed step solver assumes the step size is good enough, and does a calculation without estimating accuracy. However, in some cases, if you have lots of snubbers and parasitics in your model, the fixed step solver might appear more accurate. On average, the variable step solver is more accurate. I've posted some specifics to your other question for getting your model to work discrete. The main issue is adding a snubber to the MOSFET and making your fixed step time smaller.

For this application, I don't know how important your switching dynamics are. It would be much faster to simulate using average models. https://www.mathworks.com/help/physmod/sps/powersys/ref/buckconverter.html

This will let you pass in a duty cycle instead of a PWM cycle if you select the average option in the drop down menu. It will simulate much faster, but won't have switching dynamics.

Regards,

Joel