What is Lessons Learned?

What is H2LL?

This database is supported by the U.S. Department of Energy. The safety event records have been contributed by a variety of global sources, including industrial, government and academic facilities.

H2LL is a database-driven website intended to facilitate the sharing of lessons learned and other relevant information gained from actual experiences using and working with hydrogen. The database also serves as a voluntary reporting tool for capturing records of events involving either hydrogen or hydrogen-related technologies.

The focus of the database is on characterization of hydrogen-related incidents and near-misses, and ensuing lessons learned from those events. All identifying information, including names of companies or organizations, locations, and the like, is removed to ensure confidentiality and to encourage the unconstrained future reporting of events as they occur.

The intended audience for this website is anyone who is involved in any aspect of hydrogen use. The existing safety event records are mainly focused on laboratory settings that offer valuable insights into the safe use of hydrogen in energy applications and R&D. It is hoped that users will come to this website both to learn valuable lessons from the experiences of others as well as to share information from their own experiences. Improved safety awareness benefits all.

Development of the database has been primarily supported by the U.S. Department of Energy. While every effort is made to verify the accuracy of information contained herein, no guarantee is expressed or implied with respect to the completeness, causal attribution, or suggested remedial measures for avoiding future events of a similar nature. The contents of this database are presented for informational purposes only. Design of any energy system should always be developed in close consultation with safety experts familiar with the particulars of the specific application.

We encourage you to browse through the safety event records on the website and send us your comments and suggestions. We will continue to add new records as they become available.

How does H2LL work?

If you have an incident you would like to include in the H2LL database, please click the "Submit an Incident" button at the top of the page. You will be asked for a wide range of information on your incident. Please enter as much of the information as possible. In order to protect your and your employer's identities, information that may distinguish an incident (your contact information, your company's name, the location of the incident, etc.) will not be displayed in the incident reports on H2LL.

Lessons Learned Corner

Visit the Lessons Learned Corner Archives.

Key themes from the H2Incidents database will be presented in the Lessons Learned Corner. Safety event records will be highlighted to illustrate the relevant lessons learned. Please let us know what you think and what themes you would like to see highlighted in this safety knowledge corner. You can find all the previous topics in the archives.

A single-stage diaphragm compressor failed during boosting of high-pressure hydrogen ground storage banks. The compressor sources hydrogen from a 44 MPa storage bank as suction and discharges it at a stop set point of 85 MPa. The compressor capacity is 0.71 m3/min (25 scfm).

The original notice of failure was through an inter-diaphragm pressure indication and alarm. There should not be any pressure build-up between the layers of the diaphragm. Upon opening, hydraulic oil was found, leading to the assumption that the hydraulic-side diaphragm was leaking. Although spare diaphragms and seals were available for on-site repair, difficulty was encountered in attempting to remove the compressor nut above the diaphragms. Similar difficulties were encountered when the unit was returned view more

Two fitting failures were experienced for fueling equipment filling systems. Both fittings were installed in the system thermal chamber experiencing ambient temperatures of -40C to +50C. They were connected in high-pressure lines used for 70MPa hydrogen fueling.

The first fitting, a 0.25-inch NPT hose connection, was in service for approximately one year with no signs of leakage. The failure was noticed when the system was pressurized during a filling sequence. The failure was discovered by an audible hissing noise during leak checking. The system was depressurized and the fitting removed and replaced. The system was re-pressurized with no further leakage.

When attempting to reconnect a second fitting, a double-ferrule high-pressure connection, the fitting in question view more

The subject needle valve was used primarily for manual filling to control the flow rate of hydrogen from storage banks to the 70MPa test system. The valve was installed on the exterior of the thermal chamber in ambient temperatures of -5C to +30C. The gas flowing through the valve was at conditioned temperatures of -40C to +50C. The valve was in service for approximately two years and 400 fill operations.

Failure occurred during a test under an open valve condition. When attempting to close the valve, the turning force increased and the technician was unable to completely close the valve. An upstream ball valve was closed to isolate the flow.

An explosion at a coal-fired power plant killed one person and injured 10 others. The blast killed the delivery truck driver who was unloading compressed hydrogen gas, which is used to cool the plant's steam generators. Hydrogen deliveries are routine at the plant, occurring about once a week. Evidence pointed to the premature failure of a pressure-relief device (PRD) rupture disk, which had been repaired by the vendor six months prior to the explosion.

Hydrogen and chlorine concentrations at a certain plant are measured once each shift. On the morning of the explosion, the hydrogen concentration in the chlorine header leaving the cell bank was 0.47 percent. After passing through the chlorine coolers and liquid/gas separators, the hydrogen concentration of the gas streams increased to 2.5-3.2 percent H2, i.e., 63-80 percent of the lower flammability limit.

About 5 hours after the measurements were made, the DC power to the electrolysis cell bank was shut down because of intermittent power supply problems. At that time, a low-order explosion was heard from the chlorine dryer area of the plant. Thirty seconds later, chlorine gas began escaping from the chlorine header pumps, and another explosion occurred in the electrolysis cell view more

Overview: A pipe end containing fuel oil corroded at the outlet of a heat exchanger on the outlet side of a desulfurization reactor. The corroded pipe end leaked hydrogen gas, which exploded, causing oil to leak from the heat exchanger. The leaking oil developed into an oil fire, and the damage spread. The causes of the pipe end corrosion include the following:

There was a high concentration of corrosive substances in the process injection water.
The concentration of corrosive substances increased due to re-molding the heat exchangers.
The shape of the pipe cap was dead end piping.

Incident: During normal operations at a fuel oil refinery, a pipe end in a desulfurization unit developed a hydrogen leak, which led to an explosion. The pipe end was located on view more

Overview

The catalyst in a dehydrogenation reactor, which was usually operated under a hydrogen atmosphere, was changed while the reactor was isolated from the peripheral equipment by closing a 20-inch remotely controlled valve. The hydrogen pressure in the peripheral equipment was set at 20 KPaG, and the reactor was opened to the atmosphere. Anticipating some hydrogen leakage, suction from the piping was accomplished with a vacuum device and, nitrogen sealing was performed. When the piping connections were restored after changing the catalyst, flames spouted from the flange clearance and two workers were burned. One cause of the fire was poor management of the catalyst replacement process.

Incident Synopsis

A catalyst exchange was carried out in a dehydrogenation view more

Overview

Hydrogen leaked from the outlet piping of a hydrogen heating furnace at a fuel oil desulfurization cracking unit during normal refinery operation. The leaking hydrogen caused a localized fire. Dilution water for cleaning polythionic acid collected in the drain nozzle after a turnaround shutdown. The chlorine concentration in this dilution water was high because its concentration in the industrial water was originally high. The chlorine in the industrial water was concentrated by the high temperature, after the plant was restarted, and stress corrosion cracking occurred. Hydrogen leaked and was ignited by static electricity or heat.

Incident 

A fire occurred at the fuel oil desulfurization cracking unit of a refinery 257 hours after startup of the plant, view more

The bulkhead between a liquid hydrogen tank and a liquid oxygen tank failed due to a series of events. Air services to the building were shut down for repairs and the facility had switched to an emergency nitrogen supply. Failure to switch back to service air when it became available, led to the mishap.

The emergency supply became depleted and two valves in the normal nitrogen purge system failed in the open position, releasing the high-pressure nitrogen gas from the manifold into the liquid hydrogen tank. The gas flow raised the liquid hydrogen tank pressure to 4.5 psig. That was sufficient to rupture the bulkhead wall.

Overview
A solution of potassium carbonate was being drawn off to an inventory tank for a turnaround/shutdown maintenance activity at a refinery's hydrogen production unit. On the day of the incident, the solution level in the tower wasn't checked as it should have been, which resulted in hydrogen gas flowing back into the tank until the increased pressure caused the tank to explode. The direct cause of the incident was the workers neglecting to check the solution level in the tower. It is not known whether the potential for backflow of hydrogen gas into the inventory tank was understood beforehand or not.

Incident Synopsis
An explosion occurred due to unexpected backflow of hydrogen gas while a solution of potassium carbonate was being drawn off to an view more