Видео

You are quite welcome!

You are welcome and I am glad you are finding the series worthwhile.

Thanks! Unfortunately I don't have any material on that topic to post right now.

I don't have a paid course version of my classes.

Distribution topics are in playlist: ruclips.net/p/PLZKNjI5MfCm2PTr0NEdRAxbu0qaktyhe6

Hmm, I'm a child of the 60s so that might be another good topic.

I do not have any material on that topic to post.

You are very welcome.

Hmm, I just looked at it and appears to run properly.

It has to do with how the transformer core model is represented. In the inrush case with model set to "ideal", I model the core saturation as a current injection source on the primary side that emulates the core saturation characteristics. So this current injection effect is essentially added to a transformer model with no core to get the total effect. In the ferroresonance case, I set the transformer model to "nonideal" and utilize a more detailed nonlinear core model. I'm by no means an EMTP transformer modeling expert, but there is more detail on this in the PSCAD program documentation.

If system frequency is over 62 Hz or under 56.5 Hz, DER can remain engaged up to 160 msec mark. But after 160 msec, when these limits are exceeded, the DER needs to trip. DER does not need to wait 160 msec in these ranges though, it could trip instantaneously. Not sure if that was what you were asking though.

@@dr.davidlubkeman284 Yes thats the answer to my question. Thank you very much.

When applying a pi-equivalent model, accuracy depends on the line length. We would want the line length to be much smaller than the wavelength associated with the highest frequency component. But I don't have a quantitative rule I can give you for this. Note we can break a line model into multiple series pi-equivalent segments to improve accuracy. So if you were uncertain about line model accuracy, just keep breaking a line into smaller segments until simulation results stay the same.

Thanks!

Did you refer to my original PSCAD file on the Github site or build the model from the screen shots in presentation? Wondering if there is something going on in the other parts of the circuit model. Maybe compare to my pscx file at: github.com/DavidLubkeman/Power-System-Transients-Course/blob/main/Lecture%208c%20-%20PSCAD%20SinglePhTransformerInrushEx.zip

@@dr.davidlubkeman284 Thank you very much, Very helpful. I initially modeled it based on the screenshot, but after reviewing your PSCX file, I noticed that the voltage input time constant in my source model was left at the default value (0.05s). After changing it to 0 or a very small value, the inrush current now matches yours. One follow up question on this: I am modeling a generating unit along with a step up power transformer in PSCAD. But I do not have the values for air core reactance, magnetizing current, or knee voltage, as the project is in early stage and these values are obtained through transformer testing reports. I do have other rating parameters such as X. Is there a way to estimate them with reasonable accuracy?

@@maryamd99 I don't have any material on that topic ready to send. I would suggest you go to the PSCAD web site and search their Knowledge Base. Once you do that, then you could also contact their technical support.

@@dr.davidlubkeman284 Thank you!

You are welcome.

Both V and E can be used in general to refer to voltage. For nonideal transformer modeling, I use V to refer to the transformer source or load voltage. Then given we have primary and secondary winding resistance and leakage reactance, use E to refer to the voltages associated with the ideal transformer voltage component, which is embedded in the nonideal transformer model. It would help if you could give me a specific lecture number and slide number to refer to.

@ This is common in questions from the textbook “Power System Analysis and Design”, where problems would provide E and ask to solve for V in the same problem. I just found it confusing, but treated it the same.

@ Thank you for the answer.

@@BillLehman-qg8jk Yeah it can be confusing. I would tend to use V to represent a measurable voltage at terminals of a device, while I would use E to represent an "internal" voltage associated with rotating machine or transformer winding models. You will also see U used to represent voltage. But I'm sure I get sloppy with the usage depending on what textbook I might have been referencing at the time while working on my notes.

High voltage does not necessarily translate to core saturation, since the current energizing the core is limited by the magnitude of the chopped current. I guess there would be frequency dependency in the core model, but I'm thinking that if you set up a PSCAD model, that the next step for more accuracy would be to model cabling and resistive damping due to the core model, loads etc. But I haven't actually gone through this exercise myself.

That capacitance is due to stray capacitance seen at the breaker. So the value is relatively small and time to discharge would not be an issue.

If you have any questions as you go through the videos, please post in the comments.

@@dr.davidlubkeman284 Thank you. Question I have is in the first problem, step a shows a complex conjugate for S. I show I’m my textbook, S=IV* which I can understand, but I don’t know how I=S*/V* is derived.

@@BillLehman-qg8jk For single-phase complex power we have S=VI*. So I*=S/V. If we take the complex conjugate of both sides of the equality we have (I*)* = I = (S/V)* = S*/V*.

@ Thank you - It makes sense. I can see it now.

You are welcome!

This is not a computer simulation, but a real field recording as taken from a substation relay showing the impact of PV plant energization. I don't have the information on what percentage of the plant's inverter transformers were energized in this scenario.

The reference I am using for time is that t=0 at the inception of the fault. But I cannot just have a source voltage of sqrt(2)Vm*cos(wt) driving the circuit, since in practice a fault could occur at any point on that source waveform. So I add a point on wave angle to the source and model it as sqrt(2)Vm*cos(wt+theta). So if theta=0, the fault occurs at a voltage peak, or if theta=180 degrees, the fault occurs at a negative peak.

In this case the resistance represents a 0.5 Ohm fault impedance. It was added to the Lecture 4b worked example to make the problem more challenging. If we were to model a load resistance, that value would be much higher than 0.5 Ohms.

You are welcome.

@@dr.davidlubkeman284 Dr. Lubkeman, I just wanted to take a moment to say thank you! I had the privilege of being one of your students back in 2019, and it was an absolute pleasure to take your course. I truly enjoyed every minute of your class and learned so much. Your teaching made a lasting impact, and I’m incredibly grateful for the knowledge and inspiration you provided. Thank you so much!

@@FDR49861 Glad to hear you got so much out of this course. 2019 seems so long ago now. Hope you are doing well.

You are welcome. Glad to hear you are finding the videos useful.

You are welcome. Good luck on your project.

Depends on what types of transients you are trying to analyze. For the basic first and second-order switching transients modeled by a combination of R, L and C components, you still have a linear system and don't need a small-signal model. But if you have nonlinear elements and are looking at the impact of small perturbations, that might be a useful approach. You have to be careful though, since the transient could involve a large perturbation where the linearized model is no longer valid.

You are welcome.

You are welcome.

ECE451 is a first course on power systems analysis with a focus on transmission systems which I don't have videos for.

Glad you liked it.

I am not posting the lecture slides, but there is additional material at the course GitHub site that includes written notes for some of the lectures. Go to github.com/DavidLubkeman/Power-Distribution-Systems-Course

@@dr.davidlubkeman284 Thank you

You are welcome.

Take a look at the material on RL Fault Transients in Lecture 2a and Shunt Capacitor Switching in Lecture 3a. In both cases, we have an AC source that is driving the steady-state solution and impacts the transient solution.

Did not end up posting the microgrid control lectures since that had too much proprietary material.

Great to hear you are finding the videos helpful!

I did not post the OpenDSS files for this lecture since there was some proprietary information in them. As an alternative I suggest you look at the circuits in the OpenDSS software distribution under OpenDSS->EPRITestCircuits->epri_dpv. You should be able to apply similar principles as discussed in the lecture. Would suggest you also look at Lecture 20 videos.

@@dr.davidlubkeman284 Thank you. I'll follow the lecture continuously. Thanks for posting lectures.

Take a look at Lecture 13c: Short Circuit Analysis - Examples - Power Distribution Systems Spring 2021 - Lubkeman.

@@dr.davidlubkeman284 Awesome! Just what I was looking for. Thank you so much sir.

The figure you are referring to does not show the tap. But basically when we have a nonzero tap position, we are inserting a portion of Winding 2 in series into our circuit. We can manipulate the Winding 2 connections to get either a voltage raise or drop. If we want to take the voltage regulator out for maintenance and not interrupt the circuit, then we need to close the bypass switch, which will short Winding 2 out. If you have a nonzero tap position, you will have voltage across this winding. So shorting out Winding 2 would result in a substantial fault current. This is why the regulator needs to be in a neutral tap=0 position before shorting out Winding 2.

The total number of customers on this feeder circuit is 550. When we get the SAIDI number, that would be the average sustained outage time on that feeder per customer per year. In this scenario, every time a fault causes the top of feeder circuit breaker to operate open, then 550 customers will see an interruption though.

Thank you so much for replying so quickly! Just to make sure, is SAIFI denominator then the number of customers affected by the outage? (Different to the total number of customers on the network whole network). Thank you again for all the course notes and sharing them.

@@WietseBuwalda SAIFI denominator is also the total number of customers on the circuit. So SAIDI and SAIFI are both averages over the entire connected customer count. A consequence of this is that you could have a small number of customers on the circuit that experience many outages, but once you average this out over the entire connected customer base, SAIDI and SAIFI could still be low.

@@dr.davidlubkeman284 Thank you again for your time! Very much appreciated. Learnt a lot from your course material.

@@WietseBuwalda You are welcome

The voltage is still zero right after the fault clears, according to the boundary condition for the capacitor. However voltage quickly jumps to a magnitude of 2x peak voltage at a very high frequency given by w0=sqrt(1/LC). Frequency is high since that stray capacitance is very small. In a computer simulation it may appear to change instantaneously, but there is still a small time duration associated with that jump.

Happy to hear you found these videos useful.

A really useful video ,