Bidirectional Single-Phase Inverter (Simulink Model)
A bidirectional single-phase inverter model — grid-connected, dq-frame current control, 4-quadrant operation.
Paid course
Explanation and Comparison of Control Methods for a Grid-Connected Bidirectional Single-Phase Inverter
A 175-cell Jupyter notebook and a 57-page PDF that walk through every control decision for a grid-connected bidirectional single-phase inverter — in both abc and dq frames.
What you'll learn
- Model a grid-connected single-phase inverter as an averaged plant (L + C + switched voltage source).
- Design a PI current controller in abc coordinates and understand why it fails at fundamental.
- Derive Park + Clarke transforms for the single-phase case (SOGI, OSG variants).
- Design the equivalent dq current controller with feed-forward decoupling.
- Build a PLL that locks through 5 % line-voltage distortion.
- Pick between abc, αβ/dq, and resonant control based on your hardware constraints.
- Read the closed-loop Bode plot and predict the step response before simulating.
- Run 3 complete simulations: abc / dq / resonant — side-by-side comparisons of step and disturbance behaviour.
Prerequisites
You should already know: KCL / KVL, Complex-impedance reasoning, PI controller basics, Python at the level of reading a numpy script.
You don't need prior experience with: Prior single-phase inverter experience, dq-frame math (derived from scratch), Park / Clarke transforms (derived), PLL internals (built up from basics) — these are taught from first principles.
Deliverables
- 57-page PDF (structured notes, full derivations, diagrams)
- Jupyter notebook (175 cells, runs in ~30 s on a laptop)
- All simulation code + data
- Email access for questions during your first 6 months
Syllabus
1. Single-phase inverter as a control plant
- Topology + averaged model + discrete-time implications
- Why a PI on inverter current fails in abc
- Simulink model familiarisation
2. Control in abc coordinates
- Proportional-resonant controller — theory and tuning
- Tracking-error behaviour at fundamental
- Transient response benchmarks
3. Control in dq coordinates
- Park + Clarke (+ SOGI/OSG for single-phase αβ generation)
- Decoupled PI controller in dq
- Anti-windup strategies (3 options)
- Feed-forward from grid voltage
4. Phase-locked loops
- SRF-PLL basics
- Bandwidth vs distortion-rejection trade-off
- Robustness to frequency deviation
5. Comparison — abc vs dq vs resonant
- Step responses on the same plant
- Disturbance rejection (grid-voltage dip)
- When to pick which
About the instructor
Kristian Skorpen — Power-electronics engineer at Pixii (grid-connected BESS). MSc Electrical Engineering, NTNU. I publish runnable Simulink models and deep-dive courses that teach control design the way I actually use it at work.
FAQ
Can I expense the course?
Yes. The Stripe receipt includes your VAT number (add it at checkout). Most EU employers accept this for reimbursement.
Which OS does the notebook run on?
Mac, Linux, Windows. Only dependency is Python 3.10+ and the packages listed in the included requirements.txt.
Do I get updates?
Yes. Minor updates are free forever. If a major v2 ships, you get it at 50 % of the then-current price.
Refund policy?
14-day, no-questions refund. Reply to the receipt email and you'll have it back within 48h.
VAT receipt?
Yes — automatic from Stripe, with your VAT number if you provide one.
Do you offer team or student discounts?
Teams: email me (kristian.skorpen.yt@gmail.com) for 3+ seats. Students: DM me from your .edu address for a 40 % discount.
Do you offer support?
Yes — email access for technical questions during the first 6 months after purchase. Typical turnaround 1-3 business days.
Is this course specific to Norway / Europe?
The physics and control are universal. Examples use 230 V / 50 Hz but the notebooks parameterise for 120 V / 60 Hz too.
Does it cover three-phase inverters?
No — this course is single-phase. For three-phase, see my [simulation models](/simulation-models) (deadbeat + dq versions with a PLL); a three-phase course may follow in 2026-2027.
What tools do I need installed?
Python 3.10+, Jupyter, numpy, scipy, matplotlib, control. The notebook's first cell imports and sanity-checks them.
A sample before you buy
A free 5-page excerpt (the dq-transform derivation chapter) is available here. Enter your email below to get the PDF.