Failure history of the Russian launch industry

Please consider donating to Behind the Black, by giving either a one-time contribution or a regular subscription, as outlined in the tip jar to the right or below. Your support will allow me to continue covering science and culture as I have for the past twenty years, independent and free from any outside influence.

Doug Messier has compiled a detailed and what appears to be a complete list of all Russian/Soviet launch failures going back to 1988. As he says, “Launch failures are not a bug in the system, they’re a feature.”

What struck me most about his graph is the number of Soyuz rocket failures. For decades, various versions of this rocket have been used to bring astronauts up to either Mir or ISS, and because there have not been any launch failures during those manned launches, the impression given is that the Soyuz is one of the most reliable rockets in existence. Messier’s table proves that impression false, and also tells us that the Russians, and the United States, have been very lucky that no lives have been lost in the past three decades on any Soyuz launches.

The table also illustrates why commercial customers have been so quick to shift their business from the Russians to SpaceX. The Russians have not provided a very good or reliable product. Since 1988 there have only been two years, 2001 and 2003, in which the Russians had no failures. And the table indicates that their failure rate has increased in the past decade.



  • Localfluff

    Crewed Soyuz launches have all been survived. During 50+ years. Off the top of my head it has had three failures, all on the launch pad and including one of the uncrewed earliest test launches.

    At that empty test launch the main engine started but not the boosters, so it remained on the launch pad. However, that activated the launch escape tower rocket system. It was programmed (electro-mechanically) to abort if the Soyuz leans more than 15 degrees while below a certain altitude. Thus it fired half an hour later, when the rotation of the Earth had turned the gyroscope by that much and ground staff were at it to defuel it. The then surprising launch of the abort tower caused the rest of the rocket to explode with casualties.

    When a crewed Soyuz started to burn on the launch pad in 1983 they were successfully evacuated by the launch escape tower. Although it took 22 seconds to activate it by radio because the communications cables were the first thing destroyed by the fire. And the radio was in another building than the control center and required a pass word (ehum, is “jalousie” spelled with G-E- or C-H…?) But kerosene burns modestly in air, until the oxygen tanks explode. That makes it inherently a bit safer than hydrogen rockets, when there’s a launch escape system available.

    Another time, also before 1988, a crewed Soyuz ignited only for a very short moment, but it was enough to give the crew another launch escape tower ride. I’ve heard it gave the crew the highest G-forces ever survived uninjured by human. So Soyuz’ abort system has been tested successfully in heat. But it cannot help upper stage failures and that’s a big concern with all of their non-crewed failures, the seven latest failures of the Soyuz according to his list were in third stages.

  • David

    Saying “Crewed Soyuz launches have all been survived” is both accurate and misleading. It’s like saying “STS only had one launch failure.” It covers up the fact that the craft has had multiple mission failures after launch, some of which seriously injured crew and at least one of which was fatal. When evaluating the safety of the system and the risk of continuing to use it for crewed flight, of course you have to count the whole mission, not just the launch. Of course from a pure commercial satellite perspective, the capsule part can be discounted, which does change the data, but it still doesn’t look good in comparison to the other options.

  • Localfluff

    The Soyuz capsule landings are another story. Landing in China. Being dragged by the parachutes across the landscape. But Russians are used to deal with the wilderness.

    Anyway, commercial crew isn’t far off now. Soyuz will have served for way way too long, but done its crewed launches very well all in all, on the bottom line. Russians are traditionally superstitious, they believe in luck and it seems to have helped the Soyuz for as long as it was needed.

  • From the table I notice there doesn’t appear to be any pattern to the failures: rockets failed in a variety of ways. As Robert has pointed out on many occasions, this suggests endemic organizational problems.

    Out of curiosity I looked up the history of manned space flight incidents on Wikipedia. Interesting reading.

  • Edward

    Quality assurance seems to be a widespread problem, judging by the causes in Douglas Messier’s list. Ingestion of debris, cleaning rag [left] inside propellant feed system(!), software error, payload fairing disintegrated in flight (twice), stage-separation fired before launch(!), and cetera.

    Blair Ivey wrote: “From the table I notice there doesn’t appear to be any pattern to the failures

    We should expect that there not be a pattern, if the manufacturers and launch teams are learning from previous mistakes. However, Ariane 5, Atlas V, and Delta IV show us that the organizational culture can be changed in such a way that it learns more than specific lessons and becomes more careful in overall design, test, and operation methods; overall quality assurance can be improved so that there are very few failures. These three rockets have had a very low failure rate from the beginning, and even the Delta II developed a very low failure rate.

    One spacecraft manufacturing company that I worked for took on the task of teaching its employees how to be as careful with quality as the US airline industry had become. This company was implementing a similar philosophy that the airlines now use. It did not work perfectly, as one of the spacecraft that they built had a failure due to some carelessness during manufacturing. However, it seemed that overall errors were being reduced, so the philosophy and new corporate culture was beginning to take hold, as it eventually took hold in the airline industry.

    Russia probably does not yet have the same philosophy, as their culture has long been very different. Major failures have been caused due to this different culture. While performing tests, they collided a Progress freighter into their Mir space station, a few months after a near collision under the same circumstances, and they had an infamous meltdown at their Chernobyl power plant. A root cause to both of these accidents is the way they present incentive payments for their tests; encouraging the person in charge to take unnecessary risks during the test in order to get the incentive pay for completing the test.

    Russia destroyed a hydroelectric power plant, because providing power was more important to them than keeping a misbehaving turbine offline until it could be repaired. That kind of culture risks disaster and tragedy, and could be a cause of some of the rocket and spacecraft mishaps that have happened in Russia and the Soviet Union.

    If they were running a turbine with known deficiencies, in essence, they’re putting economic concerns before human-life safety factors

Leave a Reply

Your email address will not be published. Required fields are marked *