千方级电解槽的返厂大修：隔膜更换与长周期衰减的技术账Overhaul of Thousand-Unit Electrolyzers Returning to Factory: Membrane Replacement and Long-Cycle Degradation Technical Account

A maintenance procurement announcement has for the first time brought the reliability accounts of green hydrogen equipment into the public market. On October 29, 2026, the Sinopec Construction Engineering Electronic Tendering and Bidding Transaction Platform released the "Procurement Announcement for the Repair of Faulty Electrolyzers at Hydrogen Plant No. 1 of Xinjiang Green Hydrogen Company (Section 1)," with the maintenance scope covering two faulty single-unit 1000Nm³/h electrolyzers at the Hydrogen Plant No. 1 of the Xinjiang Kuqa Green Hydrogen Demonstration Project. The estimated contract amount for the section is 2.3642 million yuan including tax. The main work items were listed in detail: disassembly and reassembly of electrolyzers and round-trip transportation; complete replacement of gaskets and diaphragms; performance evaluation and maintenance replacement of bipolar plates and Belleville springs; electrolyzer disassembly, assembly, factory inspection, and performance testing. In a single maintenance list, there were simultaneously four types of components: diaphragms, gaskets, bipolar plates, and Belleville springs. This does not look like a single-point failure, but rather resembles the grouped aging of core electrolyzer components being revealed all at once.

Technical signal in the maintenance scope: what do the four types of components imply?

Let's first look at the most critical component: the diaphragm. The diaphragm in an alkaline electrolyzer has to fulfill two conflicting responsibilities: it must conduct ions to maintain the current, while also preventing the mixing of hydrogen and oxygen on either side. The balance of its performance determines the cell's overall power consumption and safety margin—insufficient diaphragm performance not only increases power consumption but also poses safety risks due to hydrogen-oxygen crossover. What makes it more troublesome is that the diaphragm can degrade in multiple ways: under high temperature and high pressure, the highly alkaline electrolyte severely corrodes components like electrodes, seals, and pipelines. Impurity ions in the water source and electrolyte can easily deposit and form scale on electrode and diaphragm surfaces, leading to catalyst layer poisoning and diaphragm blockage, ultimately manifesting as continuously rising cell voltage and increasing power consumption.

The gasket and diaphragm are required to be fully replaced in the same list, indicating that they most likely belong to the same batch and have reached the end of their service life at the same time. The sealing design of the electrolytic cell must withstand alkali corrosion, temperature cycling, and pressure fluctuations simultaneously. After the gasket ages, its airtightness decreases, and the leaking alkaline mist can contaminate the drying system, clog pipelines, and affect instrument readings. This type of failure has a particular characteristic: it does not cause an immediate shutdown, but manifests as a gradual efficiency degradation until a certain reading crosses the alarm threshold and is noticed.

The bipolar plates and disc springs are listed for performance evaluation and maintenance replacement, rather than direct replacement, which is a wording worth noting. The disc springs bear the clamping force, and the contact resistance directly determines the cell voltage; the flatness and coating condition of the bipolar plates also affect contact resistance and flow channel distribution. The failure modes of these two components are not simply breakage or rust, but performance drift, which is why evaluation is required instead of direct replacement. A maintenance list covering all four types of components reflects that domestic 10,000-ton projects have begun to enter an operational stage where core components need major overhauls according to a cycle.

Acceptance Test Accident: What It Means When Ignition Occurs During Controlled Emission

The implications of another case during the same period were completely opposite. According to the progress announcement of the MFE220 test accident investigation released by Clean Power Hydrogen plc on October 19, 2026, its 1-megawatt membraneless electrolyzer experienced a system ignition failure during the third and final stage of the factory acceptance test on May 28, 2026, causing severe damage to the equipment. The preliminary conclusions listed in the announcement include four points: the gas flow rate and particle impact near valve V705 may have generated enough energy to trigger ignition during standard controlled emissions; some check valves, hoses, and pipe elbows had deviations in specifications and configuration; supplier quality and the environmental characteristics of the outdoor test facility affected the overall cleanliness required by the unit, resulting in particle contamination inside the unit; and the applicability and compliance of this new technology design with the relevant guidance standards of the European Industrial Gases Association are being assessed as part of the remedial design process.

The announcement clearly pointed out that no single definitive ignition source has been determined, and no evidence was found to suggest that the incident originated from its proprietary membrane-free stack or supporting separator container.

The technical reasoning presented in these four points, considered together, is more worthy of study than a single-point conclusion. The incident occurred during the 'controlled venting' phase rather than during normal operation, and this location itself indicates where the problem lies—the controlled venting is a routine action designed as a safety procedure, precisely in this action the system needs to safely release the gas inside the equipment. If ignition had occurred during normal operation, the issue would most likely be related to the electrochemical process or the main system; occurring during the venting phase indicates that the triggering conditions come from the flow details and the random impact of particles at the end of the piping. Particulate contamination can be considered an independent factor because it has two characteristics: first, it is ubiquitous, being introduced through supplier materials, assembly processes, and open-air test sites; second, the energy release depends on the impact speed—the higher the gas flow velocity, the greater the kinetic energy carried by particle impacts.

There is another amplification mechanism when particulate matter falls into a hydrogen work environment: the surfaces of particles often carry metal oxides, and metal oxides are effective catalysts for hydrogen oxidation. This means that a slow oxidation process that originally requires a relatively high activation energy can be significantly accelerated due to particle catalysis. Therefore, the engineering requirements for gas system cleanliness are essentially about managing a non-gaseous reaction kinetics problem. Public information did not disclose the specific ignition energy and the composition of the test gas in this accident, so the mechanistic chain can only be qualitatively reconstructed based on the factors listed in the announcement, and cannot yet be verified item by item.

Three definitions of decay rate: Why a single number cannot be directly compared

Putting the issue of degradation on the table, the level of chaos in the industry becomes apparent. On November 6, 2026, when Dongfang Electric's 2000 Nm³/h alkaline water electrolysis hydrogen production equipment was launched, it publicly stated that the long-term operational energy consumption degradation was less than 1.5% per year; during the same period, the MIIT draft for normative conditions provided a different standard: after 2000 hours of cumulative hydrogen production by an alkaline electrolyzer, the increase in DC power consumption at the rated current density should not exceed 2%, with the increase during the latter 1000 hours not exceeding 0.8%. The denominators of the two standards are entirely different: one is the long-term drift rate calculated per year, and the other is the cumulative increment measured in hours.

Converting this scale reveals the essence of the problem. Calculated over 8,760 hours per year, a 1.5% per year timescale is far longer than the 2,000-hour test window; and the requirement in standard conditions that the last 1,000 hours should not exceed 0.8%, when converted to an annualized scale, appears much stricter. In other words, the annualized degradation figures published by manufacturers and the hour-level evaluation values under standard conditions are not measured on the same scale and cannot be directly compared to determine which is better. This is precisely the most common type of distortion in the industry today: placing numbers from different measurement windows and statistical bases side by side, leading to a conclusion of superiority or inferiority that has no physical meaning.

The two-stage writing in the specification conditions reflects an understanding of the attenuation pattern. A total increase of no more than 2% over 2000 hours, with no more than 0.8% in the last 1000 hours, means that the allowable increase in the early stage is 1.2% and in the later stage is 0.8%. This slope of loosening in the beginning and tightening later points to a typical 'attenuation pattern'—there is a one-time irreversible loss in the early stage, and after running for a while, it enters a relatively flat phase. If the industry only uses a single average to describe attenuation, this critical pattern information will be lost, and the attenuation pattern precisely determines how to schedule maintenance windows and prepare spare parts. From construction to maintenance: the operational and maintenance accounts of ten-thousand-ton-level projects start to become visible.

The Kuqa Green Hydrogen Demonstration Project provides a reference point on this timeline. According to the "Typical Cases of Hydrogen Energy and Green Fuel Pilots in China (2026)" released on November 5, 2026, since the Kuqa Green Hydrogen Demonstration Project by Sinopec Xinjiang began operation in 2023, it has been operating safely and stably for over 1,100 days as of September 2026. The project has built a 300,000-kilowatt photovoltaic power station and a 20,000-ton-per-year water electrolysis hydrogen plant, adopting an integration model with four electrolyzers to one gas-liquid separator and eight electrolyzers to one purification unit, effectively reducing system complexity and investment, and creating a full-process digital platform to achieve production optimization and precise management. The value of this integration model lies in fixing the ratio of multiple electrolyzers to one gas-liquid separator and one purification unit, reducing the number of interfaces and adjustment dimensions within the system.

But this integrated model has an implicit premise: a reduced number of interfaces means a more stable system. However, once a slot requires major maintenance, the proportional relationships must be disrupted. This also explains why the maintenance notice specifically lists disassembly, reassembly, transportation to and from, as well as factory inspection and performance testing—major maintenance is not as simple as replacing a single slot; it involves a whole process of disassembly, evaluation, reassembly, and revalidation. The estimated cost for repairing two slots is 2.3642 million yuan, which is not large compared to the project's total investment of nearly 3 billion yuan and an annual production capacity of 20,000 tons, but it marks the official transition of domestic ten-thousand-ton green hydrogen projects from the construction phase to the operation and maintenance phase, and this aftermarket expenditure had not previously appeared in any cost calculations.

Public information also shows that the electrolyzer for the Kuqa project is supplied by three hydrogen production companies—Pire Hydrogen, Longi Hydrogen, and Suzhou Cokelier—competing for the contract. The two units undergoing maintenance this time come from Cokelier. The requirements for contractor qualifications specified in the announcement are quite revealing: Since 2022, the contractor must have experience in electrolyzer manufacturing, maintenance, or servicing projects, and the hydrogen production capacity of a single unit in any single contract must not be less than 800 Nm³/h; the project manager and technical leader must have a second-level registered constructor qualification in electromechanical engineering or an intermediate professional technical title in mechanical engineering, and possess project experience on a similar scale. This means that major overhaul capability has become a scarce service with clear qualification thresholds, rather than merely an after-sales service provided by equipment manufacturers.

Assessment: The focus of equipment competition is shifting from the commissioning ceremony to the end of service life.

Combining these pieces of information can provide a clear technical assessment. The technical bottleneck of green hydrogen equipment has shifted from whether it can be built to whether it can reliably complete its lifespan under long cycles, high loads, and real operating conditions. This shift is supported by three layers of evidence: leading manufacturers can already have the actual operating performance at multiple load points verified by third parties at press events, indicating that manufacturing capability and consistency control have reached a verifiable stage; 10,000-ton projects have begun performing major overhauls of core components on a cyclical basis, showing that the operational demands due to degradation and aging are already occurring; and two different types of accidents and failures have respectively exposed the specific failure paths during acceptance testing and long-term operation.

The directions for improvement have also become clear. The first is the standardization of decay rate metrics, which requires including the measurement window, statistical baseline, and whether auxiliary energy consumption is accounted for in the testing specifications; otherwise, comparisons across manufacturers are meaningless. The second is the creation of a failure mechanism database. The draft guidelines for comments from the Ministry of Industry and Information Technology require companies to establish mechanisms for fault analysis, reliability assessment, and continuous improvement of water electrolysis hydrogen production equipment, in order to enhance operational stability and lifecycle quality. This point hits the core—turning every membrane replacement and disassembly inspection into accumulative engineering knowledge has more long-term value than any single performance improvement. The third is cleanliness management in acceptance testing. CPH2’s investigation has already listed particle contamination, open-air testing facilities, and supplier quality as three factors side by side, and the lessons are not limited to just one company.

Wei Suo, Vice President of the China Industrial Development Promotion Association and Chairman of the Hydrogen Energy Branch, stated publicly on November 6, 2026, that the main technologies, key materials, devices, and equipment manufacturing in the hydrogen energy industry chain have basically been localized. The hydrogen energy industry has entered the stages of technological innovation, demonstration applications, large-scale cost reduction, and market expansion. He also mentioned the need to strengthen collaborative innovation across the industry chain, improve the efficiency, lifespan, and safety of hydrogen production systems under fluctuating conditions, and enhance the integration of basic research, pilot testing, and engineering demonstrations so that new technologies can withstand real operating conditions and long-term operation inspections. Placing this assessment alongside the 1,100-day operational record of the Kuqa project makes the conclusion clear: localization solves the supply issue, while lifespan and real-world testing address the trust issue, and the trust issue is the prerequisite for green hydrogen products to enter international certification systems and long-term purchase and sales agreements.

As for the manufacturer's claim announced at the offline ceremony on November 6, 2026, that the long-cycle operational energy consumption of the equipment decays by less than 1.5% per year, its value depends on whether it can be replicated in the future using a unified method across equipment from different batches and manufacturers — and this is precisely the question that the upcoming standardization and testing system construction most needs to address.

The references for this article come from public announcements on the Sinopec Construction Engineering Electronic Bidding and Tendering Trading Platform regarding electrolyzer maintenance procurement; the draft for soliciting opinions on the specifications and conditions of the hydrogen equipment manufacturing industry publicly released by the Ministry of Industry and Information Technology; investigation progress announcements publicly disclosed by Clean Power Hydrogen plc; and public reports from China Petrochemical News, China Energy News Network, China Power News, and industry-specific platforms.