Most of the rectifier content out there starts with the three boxes: thyristor, diode, IGBT. Pick one and move on.

That’s not how it works on the ground.

We’ve been doing this long enough to know that the rectifier question for hydrogen electrolysis is never really about the rectifier alone. It’s about where you’re connecting, what the electrolyzer actually needs at the terminals, and how much of your budget you’re willing to spend downstream instead of upfront.

Skip the textbook definitions. You already know what each topology does.

What the grid gives you determines half the decision

Before you even look at a rectifier datasheet, get the grid connection parameters. Not the nominal voltage. The fault level. The X/R ratio. The existing harmonic background from neighbouring industrial loads.

This matters because a plant sitting next to a 110kV substation with plenty of short-circuit capacity can run a simpler front end without causing trouble. You still need filtering, but the filters are relatively straightforward—passive tuned banks, maybe a 12-pulse transformer if the electrolyzer is large enough.

Move that same plant to a rural location with a long 33kV feeder and weak grid conditions, and suddenly the same rectifier topology becomes a problem. Not because it stops working. Because the voltage distortion at the point of common coupling starts tripping things—or worse, the grid operator sends you a non-compliance notice with a penalty attached.

We’ve seen this happen. More than once. Both times the project team had chosen the rectifier based on capital cost alone and assumed the grid could absorb whatever harmonics came out. It couldn’t.

That’s where the active front end enters the conversation. Not because it’s trendy. Because it gives you control over reactive power and harmonic spectrum that passive solutions simply don’t offer. You don’t have to oversize the transformer to handle circulating currents. You don’t need a separate STATCOM just to stay within the grid code.

Whether that’s worth the higher upfront price depends entirely on the site. In some cases it’s obvious. In others it’s a close call and you have to run a full-lifecycle loss calculation.

The partial-load behaviour that changes the numbers

Vendors love quoting efficiency numbers at 100 percent load.

Real hydrogen plants rarely sit there. If you’re running with solar, you ramp up and down every day. If you’re following wind, you’re even more variable. The rectifier might spend more hours at 40 to 60 percent load than at nameplate.

Everyone quotes thyristor efficiency at full load. And sure, the number looks good—low conduction losses, hard to beat that. But real plants don’t live at full load. Drop the load down to half or less, the firing angle shifts and the power factor goes with it. Now you’re pulling more reactive current for the same DC output. That’s not a thyristor problem. That’s a system problem. And whether that trade-off makes sense depends on one thing: are you optimising for the purchase order or the electricity bill over the next ten years?

Diode rectifiers are better on the power factor front compared to thyristors, but they don’t regulate. You need the DC/DC chopper to adjust the output voltage, and that chopper adds its own loss—typically another point or two depending on the switching frequency. So the “high efficiency” label on the diode rectifier itself is a bit misleading. The system efficiency is what matters.

AFE holds the power factor above 0.98 across most of the load range. No surprises. No reactive current spikes when the sun dips behind a cloud. That predictability alone has value when you’re optimising the plant’s energy management system.

For context, our own SCR-based rectifiers for electrolysis applications have been measured at 99.34 percent efficiency under real operating conditions. That’s not a lab number. That’s on-site with load variation. But even with that, the downstream system losses from power factor and harmonics still have to be accounted for—which is why we always tell clients to look at the whole AC/DC path, not just the rectifier nameplate.

Voltage range and electrolyzer type

Alkaline and PEM stacks place different demands on a hydrogen electrolysis rectifier—starting with the DC voltage window.

Alkaline electrolyzers typically want a fairly wide DC voltage window—sometimes 200 to 500 volts difference between minimum and maximum stack voltage. That’s because the stack voltage changes with temperature, pressure, and membrane condition over time.

PEM units have tighter voltage ranges but respond faster, so they benefit from a rectifier that can track load changes quickly.

Thyristor and diode rectifiers produce a DC output that follows the AC input voltage. If the grid sags, the DC sags with it unless you have some form of compensation. AFE rectifiers have a boost capability—they can keep the DC voltage stable even when the AC dips to 80 percent or so, as long as the grid can still deliver the required current.

This matters for plants that plan to participate in frequency response markets. If you bid into the ancillary service and the rectifier can’t hold output during a voltage disturbance, you get penalised. Simple as that.

Some developers choose diode plus chopper specifically for alkaline stacks because the chopper gives them the voltage adjustment they need and the diode stage handles the bulk power conversion efficiently. That’s a legitimate approach for certain scales. Works well in the smaller to medium range where the grid is reasonably strong.

Above a certain threshold, the chopper starts to become a bottleneck. The current ratings get high, cooling becomes more complicated, and the overall footprint grows. At that point the economics often tip back toward AFE or even a hybrid arrangement.

Harmonics and the filtering trap

Here’s something that catches a lot of project developers off guard: the cost of harmonic filtering and power factor correction for a large thyristor-based system can end up being a significant portion of the rectifier purchase price. Sometimes more, depending on the required THD level.

That’s not an argument against thyristor. It’s an argument for budgeting it properly from the start.

We’ve had projects where the owner approved the rectifier CAPEX, then got shocked by the filter bank quote three months later. They ended up renegotiating the grid connection agreement to relax the THD limit—which worked that time, but it’s not something you can count on with every grid operator.

24-pulse systems reduce harmonics significantly, but they require a phase-shifting transformer with multiple secondary windings. Those transformers are expensive, physically large, and have longer lead times than standard units. If your project schedule is tight, that alone can be a deciding factor.

Diode rectifiers produce lower harmonics than thyristors to begin with, but they still need filtering if the plant is large or the grid is weak. The filter design for diode systems is usually simpler—fewer tuning requirements—so there’s some cost advantage there.

AFE eliminates the need for bulk harmonic filtering altogether. The switching pattern controls the harmonic spectrum directly. You might still need a small sine-wave filter to handle the high-frequency switching noise, but that’s a fraction of the size and cost compared to a passive harmonic bank.

Ancillary services and the revenue angle

This is the part that’s changed most in the last couple of years.

Grid operators in Europe and parts of North America are starting to pay for dynamic reactive power support and fast frequency response from hydrogen plants. If your rectifier can provide those services, you’re not just complying with the grid code—you’re creating another revenue stream.

Thyristor cannot do this independently. It draws reactive power from the grid. To supply reactive power back, you need additional capacitor banks or a STATCOM. That’s capital that could have gone into the rectifier itself.

Diode is slightly better because it has a natural conduction angle, but it still doesn’t give you active control.

AFE is inherently bidirectional in reactive terms. It can generate or absorb VArs on demand, within its rating. This doesn’t mean every AFE-equipped plant will automatically earn money from ancillary services—you still need the right control software and communications with the grid operator—but the hardware capability is there from day one.

For projects where the grid connection agreement includes reactive power penalties, the AFE premium often pays back well within the plant’s operational lifetime just from avoided charges. In some cases we’ve reviewed, the break-even was in the two-to-three-year range.

Liyuan Haina’s IGBT rectifier series already supports standard communication protocols including PLC, Ethernet, and RS485, so integrating with grid operator dispatch systems is straightforward. That’s not a differentiator on paper—it’s a differentiator when the grid operator asks for a response in milliseconds and your system can actually deliver it.

A different way to choose a hydrogen electrolysis rectifier

Instead of picking a topology first, we walk through four questions with every client:

How strong is the grid at the exact point of connection?
Not the nearest substation. The actual PCC with the transformer impedance, cable length, and neighbouring loads factored in. If this turns out to be marginal, AFE moves up the list.

What does the load profile actually look like over a year?
Not the idealised daily curve. The real one with clouds, maintenance shutdowns, and grid curtailment events. If partial-load operation dominates, the system-level efficiency difference starts to outweigh the upfront cost gap.

How strict is the grid code on harmonics and reactive power?
Some operators publish clear limits. Others have internal guidelines that aren’t fully documented. We’ve learned to request the formal grid connection study report before committing to a rectifier supplier. That report tells you exactly what you need to achieve.

What happens when something fails?
This one gets overlooked too often. Thyristor fuses are standard parts. Diode modules are widely available. IGBT stacks are more specialised and usually come from one of three manufacturers. If your plant is in a remote location, the lead time for an IGBT replacement might be weeks versus days for a thyristor fuse. That difference matters for availability-based revenue models.

Answer those four and the topology choice becomes fairly obvious. Not because one technology is universally superior—because the site conditions eliminate the others.

The rough guide we use internally for hydrogen electrolysis rectifier selection

For a smaller plant with strong grid, stable voltage, and moderate harmonic limits: diode with a chopper or thyristor with 12-pulse filtering. Both work. Which one you pick usually comes down to this—are you trying to keep the purchase order low, or are you looking at the efficiency number over the next ten years?

For medium-scale plants with weak grid or tight harmonic limits: diode plus chopper becomes a stretch. You start needing 24-pulse transformers or active filters. At that point AFE is at least worth a serious comparison, not just a line item.

For large-scale facilities, or any plant that expects to export reactive power or participate in frequency markets: we’ve stopped recommending passive front ends for new designs in most cases. The engineering effort to make them compliant and efficient ends up costing more than the AFE premium, and you still don’t get the dynamic response.

That’s not a universal rule—there are exceptions for plants with exceptionally strong grids and very stable load profiles—but it holds for the majority of projects we’ve reviewed.

For the hydrogen electrolysis market specifically, Liyuan Haina offers both SCR and IGBT-based rectifier platforms. The SCR series handles output voltages up to 1500V DC and currents up to 100kA, with configurable 12/24/36-pulse transformers depending on harmonic requirements. The IGBT series delivers input power factor above 0.95 and supports the control interfaces that modern grid operators expect. It’s not about pushing one or the other—it’s about matching the platform to the site data.

One last thing

The rectifier suppliers will all tell you their solution is the right one. That’s their job.

Your job is to check their claims against your site conditions, your load curve, and your grid connection agreement. The data is what matters, not the brochure.

We’ve seen thyristor plants run beautifully for a decade and AFE plants have early-life failures because the cooling wasn’t designed properly. We’ve also seen AFE plants outperform everything else in locations where the grid was considered borderline unbuildable.

There’s no single answer. Just a set of trade-offs that shift with every site.

Get the site data first. Everything else follows from that—including the right hydrogen electrolysis rectifier for your specific connection point.

If you’re currently defining the rectifier specification for a hydrogen project, contact us.

⚡ Liyuan Haina Rectifier

Professional Manufacturer of IGBT and SCR Rectifiers

Our products are mainly used in electroplating, electrolysis, electrochemistry, anodizing, electrophoresis, smelting, surface treatment, photovoltaics, energy storage, hydrogen energy, and other applications. They are also widely used in industries such as hydrogen electrolysis, copper refining, copper electrowinning, electrolytic copper foil, and related industries.

📧 CONTACT US