Operating an alkaline electrolyzer below 20% of rated power carries a risk of hydrogen-oxygen cross-permeation, and wind and solar output cannot be exhaustively enumerated, while offline simulation cannot connect to a real controller. Yuankuan Energy uses the MT8060 to bring this closed loop into the lab, and on a Chifeng project prototype it has run the full process from wind and solar fluctuation through to synthetic ammonia.
Wind and solar output is random, yet an ALK alkaline electrolyzer requires relatively steady input, and operating below 20% of rated power carries a hidden risk of hydrogen-oxygen cross-permeation. The industry's answer is to use energy storage to smooth fluctuations, divide electrolyzer groups by role, and use hydrogen tanks for long-cycle buffering, but whether such strategies hold up requires closed-loop verification. Offline simulation falls short in three places: extreme conditions are hard to enumerate, millisecond-level switching and hour-level hydrogen storage states cannot be handled with a single time step, and a real EMS controller cannot be connected. Yuankuan Energy's approach is to connect the controller to a virtual hydrogen production system and reproduce the full time-sequence of operating conditions in the lab at a real time scale. This means the test object shifts from the model itself to the control strategy and communication timing.
Using the Inner Mongolia Chifeng wind-solar-storage hydrogen-to-ammonia demonstration project as a prototype, Yuankuan built an HIL system on the MT8060 multi-core CPU plus FPGA platform. The CPU runs models for wind and solar, energy storage, hydrogen production, hydrogen storage, and green ammonia synthesis, while the FPGA simulates the electrolyzer's NPC plus three-phase interleaved parallel Buck hydrogen production circuit at a 1 μs step, with the two sides interacting over an internal high-speed bus and the EMS controller connected to the closed loop via the Modbus protocol. The official statement is that this architecture allows microsecond-level switching transients and hour-level hydrogen storage states to be closed-loop on the same platform, avoiding the mutual constraints between accuracy and speed under a single architecture. For engineers working on boards and interfaces, Modbus access means the controller under test requires no dedicated hardware changes.
The test forms a full time-sequence cycle from strong sunlight and weak wind, weak sunlight and strong wind, strong wind and clear skies, to no wind and no light, with the wind and solar system's frequency and voltage remaining stable and lithium battery storage responding quickly to load changes. The electrolyzer group ramps up from standby to the minimum hydrogen production power point, and ALK and PEM rise to their optimal power at their respective variable-condition rates until full load; when wind and solar drop slightly, ALK backs off while PEM stays at full output, and when wind and solar fall to zero the group smoothly reduces load. In the hydrogen storage and ammonia production stage, the ammonia production system's total power rises to 0.578 MW, and the synthesis tower outlet composition stabilizes at 45.1% ammonia, 45.2% nitrogen, and 9.8% hydrogen. These figures come from test results published by Yuankuan and can serve as a reference baseline for strategy verification, but they still need to be recalibrated for a different project.
This solution targets EMS controllers and hydrogen production system integrators; it does not directly produce boards or interface IP, so domestic teams working on high-speed interfaces will not find short-term use for it. The indirect impact is in the testing phase: Modbus access reduces the amount of change required on the controller side, and if a project uses a domestic EMS, the verification cycle may be shortened; however, the MT8060 and StarSim IEHLib are both developed in-house by Yuankuan, and the model library is tied to the simulator, so switching platforms requires rebuilding models. For engineers watching domestic substitution, what is worth noting is whether such HIL platforms can open up IO and bus expansion, and whether the electrolyzer power electronics model supports user-supplied topologies.