Power Converters - Science topic

Hi, i think that:

-Partial scaling of the power converter, for example, involves sizing it for conditions that are less than the absolute maximum. This could be based on statistical analysis of wind patterns, historical data, and other factors.

-It allows for cost savings while still meeting the necessary performance requirements, as components are not oversized for events that might occur rarely.

But in full scaling, all components, including the power converter, are fully scaled, they are designed to handle the maximum power fluctuations, grid disturbances, and other extreme conditions, but might lead to higher costs.

Read this paper

You can use a real pole with the following transfer function:

H(s)=1/( 1 + s/w0) )

where w0=2*pi*f0, and make f0 = ( 5~10)*10 kHz

If the "Mean" is a real pole (low pass filter), set the frequency to 50 ~ 100 kHz.

Good luck!

For the implementation of a controller for DC-DC converters with a hardware platform such as dSPACE, TMS microcontroller, STM32 microcontroller, or Arduino you should first identify the operation region of the load as well as the input voltage range until reaches the suitable signal i.e, the updating of the duty cycle without saturation during the controller working.

It would be useful to read the paper entitled "Exploring the Causes of Power-Converter Failure in Wind Turbines based on Comprehensive Field-Data and Damage Analysis."

Yes sir, I'm interested..

I would like to work on Multi-level converters

Dear Xhuljo Kollcinaku:

You can benefit from these valuable Links about your topic:

@https://youtu.be/-xFq5QXYfwc

@https://ieeexplore.ieee.org/document/8358295

@https://www.mathworks.com/matlabcentral/fileexchange/

@https://www.mathworks.com/learn/tutorials/simulink-onramp.html

@https://www.mathworks.com/learn/tutorials/circuit-simulation-onramp.html

@https://www.mathworks.com/help/physmod/sps/examples.html?category=power-grids&exampleproduct=all&s_tid=CRUX_lftnav

I hope it will be helpful..

Best wishes..

The answer is you can not control the current of PMSM without controlling the SPEED ! because you are using synchronius frame(DQ) you have to consider speed control. The reason behind your model is that you are using different models for plant and controller so the speed is zero actually and you think you are controlling the current without controlling the speed.

S.M Masum

In general, the efficiency of a power converter (AC-DC or DC-DC) is assessed by comparing its input power to its output power. To be more specific, the converter's efficiency is determined by dividing its output power (Pout) by its input power (Pin).

For MATLAB, these tutorials might help:

1. https://www.mathworks.com/help/physmod/sps/ref/dcdcconverter.html

2. https://www.mathworks.com/videos/series/how-to-develop-dc-dc-converter-control-in-simulink.html

3. https://www.mathworks.com/help/physmod/sps/ref/averagevaluedcdcconverter.html

4. https://www.mathworks.com/matlabcentral/answers/522364-simscape-dc-dc-converter-efficiency

Dear Sandford Stebbings ,

Synchronization is to agree in time. So, in order to synchronize multiple sources such as the power inverters you have to choose the strongest one as the reference synchronization source. This source is better to be referenced to a crystal controlled oscillator for high stability.

Then every other inverter is to be synchronized by the reference taken from the microgrid as I proposed in the first part of my answer.

This means the local oscillator of the inverter must be synchronized to the reference by means of phase locked loop.

In summary my proposal is to take the strongest power fed to the grid is to be taken as a reference for all other source fed to the grid.

To synchronize any inverter to the grid please refer to the papers in the links:

and

Best wishes

Dear Sandford:

You can benefit from this valuable Link about your topic:

I hope it will be helpful...

Best regards....

Anup Marahatta

The frequency response, or variations in magnitude and phase as a function of frequency, is depicted using Bode graphs. This is accomplished using two semi-log scale charts. The top plot is usually magnitude or "gain" in decibels (dB). The bottom plot represents a phase, which is usually expressed in degrees.

A Bode plot /ˈboʊdi/ is a graph representing a system's frequency response in electrical engineering and control theory. It is frequently a mix of a Bode magnitude plot, which expresses the amplitude of the frequency response (usually in dB), and a Bode phase plot, which expresses the phase shift.

In electronics, frequency response is a quantifiable measure of a system's or device's output spectrum in response to a stimulus that is used to define the system's dynamics. It is a measure of the amplitude and phase of the output in relation to the input as a function of frequency.

Hi, Samo Smodiš

Pls refer to my colleague publications:

I would recommend reading the deliverables available at https://www.h2020-migrate.eu/downloads.html.

Dear Utkarsh Singh ,

In LV distribution networks, reactive power is not efficient in modifying electrical parameters due to X/R ratio of conductors. Although works discuss the effects on voltage in LV, IM tends to generate voltage problems like undervoltage.

Indeed you could benefit from this valuable article about your topic:

"Design Space Optimization of a Three-Phase LCL Filter for Electric Vehicle Ultra-Fast Battery Charging" .. September 2020.

Just make sure you have higher value of inductor or there is critical load resistance up to which the system stays at continuous conduction mode.

Please follow the following link:

IGBT

Thanks, everyone for your time and suggestions.

From other colleagues, I have received a high-level view of the answer. The selection of power/voltage controller as the outer loop depends on control objectives and the role of power converters. I have known two terminologies that define the role of power converters: 1) grid following converter and 2) grid forming converter. If I have a strong grid with stable and regulated voltage and frequency and need to operate a power converter to inject whatever power is available, then it is necessary to select a power controller as the outer loop. In this case, the converter would be grid following converter and the outer loop would control P and Q. On the other hand, if I have a local grid or small weak grid and the control objective is to regulate the bus voltage and frequency, then it is necessary to select a voltage controller as the outer loop. In this case, the converter would be grid forming converter and the outer loop control V and f.

Now, this is a grid-connected converter example, but the principles would be similar if it was for other applications, such as internal DC buses of back-to-back converters, etc.

Hi Lambada Puneeth Raj ,

You are being deceived by an inaccurate and incomplete model. There are two capacitances associated with the diode, D: 1) the depletion/junction/transition capacitance, C_j, which only exists when the diode is zero and reverse biased; and 2) the diffusion capacitance, which stores the minority carrier charge, but only exists when the diode is forward biased. These two capacitances are both in parallel with the diode. Note the caveat given in the diagram: "assuming linear capacitances". That is referring to the fact that C_j should actually be written as C_j(V) because it is a nonlinear function of the reverse voltage difference across the diode.

When the MOSFET is turned on, it shock excites the LCR circuit into a damped oscillation. The circuit oscillates because of L and C, and it is a dampled oscillation due to the load resistor R. In order to prevent a negative voltage difference from appearing across the load R, a clamping diode D is in the circuit. When the damped oscillation attempts to reverse direction, the clamping diode becomes forward biased and essentially stops the oscillation.

Now, as you mentioned, when the MOSFET is turned off, its C_ds is charged to the voltage difference, v_g. The clamping didoe, D, will have its diffusion capacitance charged to a value somewhat less than v_g, depending on the degree of damping from the load R, when the damped oscillation attempts to reverse direction. At this point, the MOSFET is turned on, again, discharging C_ds, and reverse biasing diode D which discharges its diffusion capacitance - not its diffusion/transition/junction capacitance.

There is a lot being left out of the description of the boost converter.

Regards,

Tom Cuff

Hello Lambada Puneeth Raj ,

Please provide us with a better picture of your circuit and better pictures of your three waveforms. As it stands, it is very difficult to see any details.

Regards,

Tom Cuff

Hi,

It would be that this reference can help:

PD: Matlab implementation codes are available.

You can use the following link:

Copy the IGBT model from here and use it with your DC converter. You can use any IGBT's datasheet to update the library. Efficiency calculation and plotting is also shown in the given model. Hope it will help.

Good question, follow

Yes, as you mentioned in your answer, the back-to-back converter and the switch mode rectifier + inverter are the most commonly used for the WTs coupled PMSG.

Check the following interested book:

Jamil, M., Rizwan, M., & Kothari, D. P. (2017). Grid Integration of Solar Photovoltaic Systems. Crc Press.

i think Ehtisham Lodhi's answers are enough.

Hi Zhaoshun Deng,

The method of Clarke transformation can be used for grid synchronization assuming sinusoidal grid voltages. As you mentioned, low pass filters can be used for taking out the harmonics. The method can give the correct phase information in steady-state operation. However, this method as you can see is an open-loop technique with no feedback signal about the accuracy of the calculated angles.

For a converter operating in the grid following mode, it is very essential that the phase angle of the grid is precisely tracked at all times in spite of any harmonics and in the presence of transient disturbances. The Clarke transformation method suffers from a poor transient response during grid disturbances, mainly due to the use of low pass filters in the forward path without negative feedback.

The PLL essentially does a similar job, but it uses a closed loop technique. The loop filter in the forward path of the PLL performs the job of eliminating the effect of unwanted harmonics. At the same time, the feedback from the output VCO in the PLL makes sure that dynamic response is not compromised.

Hej,

I think SMA has some comercially available inverters in their sortiment for this purpose. I ll guess you ll find also some prices.

Afaq Hussain thanks for nicely putting it up. As an additional understanding exercise, would recommend you, P. Arvind, to study the waveform of current through the inductor in an SMPC device. You will see, that during one mode (ON/OFF), the current rises and during the other mode, it falls down. Now, in steady-state, the average current through an inductor is constant. This means, the amount of rise and fall of current in the two different modes should be equal in magnitude, thereby canceling the overall change in current. Note that the rise and fall of the inductor current would be equal to the rate of rise multiplied with time of rise and rate of fall multiplied with time respectively. But, we know that rate of rise/fall of inductor current is proportional to the voltage across the inductor. So basically, the product of voltage across inductor during ON mode and rise time, and the product of voltage across inductor during OFF mode and the fall time, add up to zero, in other words, "Balance" out. Hence the name, inductor volt-second balance. Now, it is important to note that this 'balancing-out' happens only in steady-state when the average current through the inductor is constant. It might interest you to know, that during the transient state, inductor volt-second balance is not a valid phenomenon. I hope this helps. Thanks for reading.

Hi,

Regarding batteries, there is always one concern that I will share.

(a) When you use a capacitor, your voltage balancing logic keeps the voltage in check, as a result you only draw reactive power from cap.

(b) When you have a battery, you might forgo the need of a balancing logic, this means you 'might' draw active power from the battery. This also means that you might be pumping active power into the battery, right? It's a possibility that cannot be easily ruled out.

So how do we proceed? If you want to avoid caps and use batteries, you must at least ensure that you draw active power from the auxiliary batteries that replace caps. Otherwise, you might pump power into a battery, overcharging it beyond rated limits. You can refer my Article on IEEE that discusses a similar scenario at length. Thanks.

regarding

Nice Dear Hakim Abbasov

Hello Ayoub Mbogela , I'm performing the same setup in simulation and I have got the results in Inverting mode of operations but am unable to operate the converter in rectifying mode of operation and was wondering if you had any success in doing so. If yes, then please tell me the appropriate switching scheme.

The inverter is connected at a bus in the network. The grid is "seen" by inverter at that bus, and this perception is in the form of voltage and frequency. Question is, who is responsible for maintaining that voltage/frequency? Whether V/f will still be maintained if inverter under consideration is disconnected? Every type of inverter injects power into the bus it's connected, however the key deciding factor of its role as grid forming/feeding/supporting is maintaining V/f.

Technically, if inverter control structure directly incorporates voltage tracker/regulator and its frequency is auto-generated (not sensed from bus), then it's a grid forming inverter.

On contrary, if voltage tracker is not employed and frequency reference is obtained from sensed/measured/estimated frequency locked to the corresponding bus by mechanism like PLL or FLL, then it's grid following.

Grid support is an "additional" feature, a grid supporting inverter can be of both types: grid forming or feeding. If the references/set-points of inverter (set-points includes voltage, frequency, active/reactive power, virtual flux, virtual torque, injected currents etc.) are adjusted based on "additional" inputs from information of other buses/loads in the network, for serving extra ancillary features (apart from local v/f regulation and power injection), then the inverter is said to "support the grid". The additional support include: power quality enhancement, stability support, reserve provision, event ride-through, economic dispatch etc.

See below examples:

1) Single inverter feeding standalone network: It is alone responsible for producing and maintaining V/f at all buses in network, thus "Grid Forming". If it adjusts its V/f setpoints for other purposes (like brown-out, current limiting for overloads/faults, improving voltage profile of a remote bus, loss reduction etc.), then it is also "Grid Supporting", it is supporting the grid formed by itself! Otherwise, if V/f are regulated at constant values, it is not grid supporting.

2) Parallel connected droop controlled inverters: All inverters contribute or participate in process of grid formation by means of droop coefficients. They serve additional purpose of power sharing, hence they are both grid forming and grid supporting.

3) PV inverters always operated at MPP: Irrespective of V/f of the bus it's connected, it only injects available peak power into grid, hence only grid feeding.

4) PV inverters with derated MPP: They inject certain fraction of available peak power to grid, and power reference is adjusted as per reserve requirement of the grid. Hence, it's an example of both grid feeding and grid supporting.

My 500-page book Power Electronics-Enabled Autonomous Power Systems: Next Generation Smart Grids (Wiley - IEEE) is now available at

Enjoy reading ...

Agree with Gusius Gus

The following point briefly addresses your concerns to perform the abc-to-dq0 transformation:

1. The fundamental reason to transform the three-phase instantaneous voltages and currents into the synchronously rotating reference dq0 frame is to make computations much easier. Secondly, it allows the system operator to independently control the active (d-axis) and reactive (q-axis) components of the currents. Similarly, in machine, the flux and torque can be independently controller. This way, the coupling effect can be minimized to certain extent.

2. Additionally, in the dq0-frame, the mutual inductance is constant. Thus, it allows the system operator to achieve the desired output as the inductance dependent quantities are constant.

3. Further, the various feedback controller including proportional-integral (PI) and proportional-integral-derivative (PID) practically offers an effective and reliable control solution when the quantities are DC (i.e., constant) in nature. On the contrary, if the input quantities to these feedback controllers are of periodic or sinusoidal then the integrator term of these controllers typically fail to introduce the infinite gain. As a result, the steady-state error will not be forced to zero.

It is well known that each one of these criteria leads to a solution with certain characteristics.

as far as i remember, the ISE and ITSE lead to solutions with low damping.

That's just an example.

ITAE is most commonly used. IAE may lead to slower response, and so on.

There is the LT4320 but it only goes up to 600Hz so I guess too slow for what you need. I have also thought to use a programmable analog chip such as the Silego parts (now Dialog). They have analog comparators in them and logic so you could make a rectifier controller.

Please contact me to see if any of my new designs can be implemented.

Sharaf Energy Systems Inc. Canada

especially

green plugs

SFC filters and EE devices for pumping an d A/ C -

Air Circulation cyclical loads.

I would definitely prefer the 2nd strategy because of its control over limiting the current. This also insures that that the bus voltage under autonomous mode is regulated at 120 Vrms.

Dear Varun,

What are the frequencies for the power to be separated. If the frequency is relatively high you can use band pass passive filters. If the frequency of low you can use active power filters. For using active power filters please follow the paper in the link:

Best wishes

Dear Juan,

As both Vp and Vq are synchronized to each other, then any one of them is suitable to perform as reference for the phase locked loop.

This is a conceptual question concerning the operation of the PLL and the relationship between Vp and Vq.

Best wishes

ISO224B isolated sensor from TI, with an appropriate voltage divider on the high voltage side.

http://www.ti.com/product/ISO224

In testing the MPPT controller speed of response one applies normally a unit step change of the Irradiance and see how it takes to reach steady state condition. The settling time is time taken by the processor from instant of application of stimuli till it reaches almost steady state value. The almost steady state value is considered to be reached when the power reaches its 90 percent of its steady state value.

Best wishes

I would presume you have a basic 3-phase VSI inverter, grid-tied by an L-C filter, to answer this question. In this case, choosing the inductive reaction at the grid frequency is a popular practice to be 5-10% of your base impedance rating (= Vgrid / Igrid). For example, if your grid voltage is 230 V(ph) and the rated AC inverter current is 23 A per phase, then 10 Ohm (= 230/23) is your base impedance. Now you can calculate the inductance you need. Once you have the inductance value, the capacitance value can be calculated on the grid side (usually less than 5 percent) depending on the THD requirement.

If you don't have L-C and you only have ' L, ' then the value depends entirely on the current grid THD-higher the inductance value, lower the THD value. And the inductance is going to be much bigger in this situation.

You can use the split DC link which includes two series capacitor in parallel with a large balancing resistor for each of them . This a simple way to provide a neutral point which its voltage clamped at half of the DC link voltage .

Special PWM generating schemes can be used for balancing of the capacitors voltages. But where the neutral current is large, it flows through the capacitors and, hence, bulky capacitors are needed to make the voltage ripple on the capacitors. Also the neutral point usually drifts and becomes unbalanced when it includes DC component. So different strategies for neutral point balancing are reported, usually using the Space Vector Pulse Width Modulation or MPC control.

A better way to provide a neutral line is to add an additional fourth leg, called a neutral leg, to the conventional three-leg converter. The neutral current can be controlled to flow through the inductor so that the neutral point is maintained at the mid-point of the DC link. This allows the phase legs and the neutral leg to be controlled independently, which means the three phases can be controlled independently as well.

Ref :

Control of Power Inverters in Renewable Energy and Smart Grid Integration https://ieeexplore.ieee.org/book/6381785

The P&O method calculates the duty cycle and hence the MPP point depending on the irradiation levels. However, the fuzzy method has been trained in advance which directly operates at the MPP point. This might be the reason for less/no oscillations in Fuzzy controller compared to P&O.

Build the model in MATLAB as per your state space equations,

and cheak the responce of model and give parameter values as per your numerical example.

all types of shape are possible

The best cores for high frequencies are the ferrite cores because ferrites are electrically insulating. Therefore they will be no eddy current in the core. The limiting frequency is limited by the magnetization losses.

Ferroelectric based cores whether laminated or powder will still have eddy current losses but very suppressed. So, they normally used at lower frequencies.

As an advice you can survey the commertial coils that satisfy your requirements and specifications.

Best wishes

Dear Akram

please go through this below links, this links may be helpful for you

1.

Thank you

When you load the solar module by an inverter, then for a given load you have to adjust the PULSE WIDTH of the inverter to operate the module at the maximum power point under the the specified temperature and simulation condition.Then you have to get the maximum power point Vom and Iom, then you apply Vom on the inverter and vary its PULSE WIDTH till the input current to the inverter reaches Iom.

Best wishes

I suspect high Common Mode Noise disturbances.

Check out again with the following:

1) DSAT pin 14 to be connected to VE pin 16.

2) Ensure a resistance of about 10k to 33k connected between gate and emitter of IGBT as close to device pins as possible

3) Ensure connection of Ceramic disc capacitors (not polyester capacitors) of 0.1uF (apart from about 47uF to 100uF electrolytic capacitors) across supply pins of IC to output common, ie., between pins (12+13) & 16 and between pins (9+10) & 16, as close as possible to the IC.

4) Instead of a single gate resistance in series with gate, split the resistance into two: 2/3 of original value remains in series with gate and 1/3 of original value be put in series with track from VE (pin 16) & common of supplies to actual Emitter of IGBT.

5) In case wires are used to connect from pcb to the IGBT gate & emitter, they should be as short as possible (preferably less than 6inches in length) and the two wires to be twisted together.

Best wishes.

Dear respected Fahad Abu Khald,

The nominal battery voltage must be less than the solar panel voltage for charges to flow from the solar panel to the battery. Since the battery is to store charges, the solar batteries should be connected in parallel in order to maximize charges stored by the batteries not voltage...

Please check my profile for more information..

Best regards.

Kifilideen.

Based on the literature, yes it is possible to achieve soft switching with MPPT. Maria Teresa Outeiro has achieved soft switching with MPPT for fuel cell with SRC application. Try to go trough the authors article:

1- Design, Implementation and Experimental Validation of a DC-DC Resonant Converter for PEM Fuel Cell Applications

2- Resonant power converters applications: LsCsLpCp circuit for renewable energy sources as case study

I think you should check your simulation status, in powerGui, in continous time, it takes time, you should descritze it, and reduce the samples. Also check if there are any zero crossings, becoz matlab hangs when perofrming differentials while executing such probs, as such kinds of probs are prominemt in switching convrts.

I agree with Mr. L.T.G. Lima

Implementing long-horizon FCS-MPC with sphere decoding is not trivial. Implementing it for a system with an LCL filter is even more difficult. I would recommend to first get it to run for a simple RL load with a voltage source. Make sure you have good debugging at hand and plot your predictions. It certainly took me many weeks to get these things running. Two weeks is much too short a time frame for this.

did you solve your problem. i am having same problem in DAB.

Dear Lokesh,

SMC was initially built for nonlinear systems, while MPC was designed for linear systems to anticipate for desired control and after that there have been developments for nonlinear system. in fact each method has a specific domain or it can gives excellent performance. Also we can add the two method in same application.

For more information about this subject, i suggest you to see links on topic.

Comparison between Predictive Sliding Mode control ... - ResearchGate

https://www.researchgate.net/.../282504170_Comparison_between...

Comparison Between Predictive Sliding Mode Control ... - ResearchGate

https://www.researchgate.net/.../311316765_Comparison_Between...

Application of MPC and sliding mode control to IFAC benchmark models

Recent Advances in Electrical Engineering and Control Applications

https://books.google.com.tr/books?isbn=3319489291

Robotics and Mechatronics for Agriculture

https://books.google.com.tr/books?isbn=1351788604

Best regards

Ioannis, you are correct. A transformer does not have a fault level input.

I provided a range of 11kV network fault levels so that you can use to model it as an equaivalent Thevenin's Voltage source i.e. Infinitly stiff voltage source behind a series impedance. This is also referred to as "External Grid" in network modelling tools.

For 1a 00 MVA fault level in a 11kV network, your Thevenin voltage source value will be 11,000/(sqrt(3) = 6350V (Ph-n) and your series impedance (usually assumed reactance) = j1.21 Ohm. Your distribution transformer - say 11/0.415kV 300KVA Dyn11 Z=5% will be connected to this series impedance in your model.

Above is reasonably accurate representation of a 11kV network and a distribution transformer model for basic load flow and short circuit analysis.

In case you are planning to investigate Z(f) - impedance of the grid as function of frequency, than you will need to add equivalent shunt capacitance to the 11kV terminal of the Thevenin impedance. However, this may not be as easy to get as Thevenin equivalent series impedance. You will also have to model your LV cable as a PI model with series impeadnce and shunt capacitance.

Trust above is helpful.

1. Oscillating Water Column (OWC)

2. Hinged Contour Device

3. Floating Device

4. Hydraulic Device

Try using improved PID like Fractional pid controller or used Anti-windup technique with pi to reduce the integral effect

thank you very much