Interview: K. Radhakrishnan

GSLV MkIII, the next milestone

Print edition : February 07, 2014

K. Radhakrishnan , Chairman of Indian Space Research Organisation (ISRO). Photo: K. BHAGYA PRAKASH

Interview with K. Radhakrishnan, Chairman, Indian Space Research Organisation.

FOLLOWING THE SUCCESSFUL LAUNCH OF GSLV D5/GSAT-14, Frontline caught up with K. Radhakrishnan, Chairman, ISRO, on January 10 in New Delhi. In the conversation, Radhakrishnan discussed the various problems faced in the development of cryogenic systems and how these were overcome to prepare the organisation for its next big challenge, a success with GSLV MkIII, which will be powered by an entirely indigenous cryogenic engine and stage, and is targeted for 2015. Excerpts:

What elements of the Mark II cryogenic engine and stage, which fired GSLV-D5, still retain the legacy of Russian engine technology and design? How much of it is truly indigenous and how much of it relies on the Russian heritage?

Basically, both engines use the “staged combustion cycle”. That is one approach compared with gas generator cycle, which we are using for the C20 engine to be used in GSLV MkIII. There are several other differences, conceptually also, especially the igniter system that we are using, which is totally different from what has been used in the Russian engine. [In MkII liquid oxygen (LOX) and gaseous hydrogen (GH2) are ignited by pyrogen-type igniters in the pre-burner as well as in the main and steering engines during initial stages, as against pyrotechnic ignition in the Russian engine.] But in the staged combustion cycle, similarities can be found in the way the engine is started and the steering engines are used for controllability.

But the staged combustion cycle itself is quite complex.

The staged combustion cycle is complex but it gives slightly improved performance in terms of the specific impulse [Isp]. But in the gas generator cycle, you have the ability to test the individual elements. So if you look at the reliability aspects—establishing a reliable system and the time required for that—we can work in parallel. The issue is relevant in the context of GSLV MkIII, for which we were working on the engine and stage elements in parallel. The turbo pump, which has something like 5 megawatt of power, has already been tested and it has logged about 1,400 seconds on the ground. We have tested the thrust chamber along with the injector, igniter and the nozzle. We did two tests, and the third test is being done today [January 10]. [This test, which lasted for 50 seconds, was as predicted and was successful.] Now, when we have sufficient knowledge about the ignition characteristics, the combustion instability aspects and performance in different regimes of [LOX+liquid hydrogen, or LH2] mixture ratio, then we can start with engine test and then the stage test. So the time required from now to qualifying the stage becomes less. This is the main advantage. The flexibility that is available in a gas generator cycle is much more because individual systems can be tested from the input/output point of view and they can be qualified in parallel. In the previous situation, the stage process was started after the engine qualification.

The second aspect of the GSLV MkIII engine is that we are gimballing the nozzle for thrust vector control [and not the two using vernier (steering) engines as in the Russian engine and the cryogenic upper stage (CUS) of the indigenised MkII]. So we are only concerned about two ignitions, that is, the main engine and the gas generator. In the case of GSLV MKII’s CUS, the two steering engines have to ignite before the main engine ignites and that feed has to come from the main line. Unless the right temperature and pressure conditions are there at the beginning of that process, the steering engines will not function. In fact, in some of the ground tests, we noticed the problem of a two-phased flow. The most important part of the cryogenic engine is the sustained ignition and the termination of the turbo pumps. It should not give any undue rate for the spacecraft. So in this launch, the engine start was as predicted; the four ignitions were as predicted; similarly, the termination was also as predicted. If you look at the performance of the subsystems, the turbo pumps—the main turbo pump, the oxidiser turbo pump and the fuel booster turbo pump—in both the regimes—the uprated regime and the nominal regime—plus the temperatures, all have been well within the specifications. All the components that it employs, too, have performed well.

In retrospect, considering that it has taken such a long time, could we not have used the gas generator cycle?

No. At that time, before 1992, we were working with a one-tonne engine. We were learning cryo at that time; the learning started in 1982. If you trace the history, in the early 1970s, when the Space Science and Technology Centre was there in Thiruvananthapuram, in the pre-SLV3 stage, we were trying to understand all these propulsion systems, including hyper-propulsion, semi-cryogenic propulsion, and cryogenic propulsion. But then the essential focus was on the SLV-3 programme, which had solid propulsion in all the four stages. And that was essential for the programme. So, in 1992, the approach was technology acquisition. We followed a path and we continued with that.

What is the gain in Isp in the staged combustion cycle?

It is only of the order of 10 seconds.

The one-tonne engine and the subsequent indigenous work were all based on the gas generator cycle. So it would seem that just for that little gain we seem to have embarked on a highly complex technology that has taken such a long time to absorb.

It was complex but at that time it was the only one that was available. We now know that the flexibility available in the gas generator cycle is enormous. It is a learning process.

The GSLV has seen several failures and there must have been a lesson to be learnt from each one of them.

If you look at the first flight [GSLV-D1], essentially it was related to the mixture ratio used for the Russian cryogenic stage as far as the vehicle was concerned. In the aborted launch that took place [three weeks] earlier [March 28, 2001], essentially it was because of the blockage in one of the feed lines [by a lump of lead in the NO (dinitrogen tetroxide) feed line of the strap-on (S3) liquid engine L-40’s gas generator]. The latter called for tightening of the assembly and inspection processes. In the second [GSLV-D2] and the third [GSLV-F01] launches, there were no issues. In the next launch [GSLV-F02], there was again a fabrication issue. A dimension was not inspected during the manufacturing process. When that component [of the strap-on S4] was tested, a deviation was seen but that was taken as a wild point, something to do with the facility. It was too abnormal because we got a flow rate almost nine times what was expected. It was an annular gap which was supposed to be 0.5 mm. Something which was to be 17 mm was made 16 mm, and because of this dimensional change, what was to be 0.5 mm became 1.5 mm, resulting in an increase in the area [3x3] and hence the flow rate. That’s what happened.

If you look at the PSLV’s first flight, which failed, when we did the simulation on the ground, there was a wild point even at that time. Only one out of some 1,000-plus simulations, but it was taken as a wild point and ignored. So the lesson that we learnt is that wild points are not to be ignored but to be studied. They are an indication of something else that is happening. In F04, the control system of a strap-on stage failed because a gas motor stopped. This has again to do with the component and has nothing to do with the vehicle design. In GSLV-D3, it was again a component problem. All four ignitions started, but it was a pump [fuel booster turbo pump] that stopped. Why did the pump stop? We looked at three scenarios and the contamination theory was the most likely because we found the source of that contamination. It was a propellant acquisition system [PAS], which is basically a filter that ensures that the propellant gets into the outlet.

The initial theory was that there are three bearings, and a normally assembled motor is tested under standard room conditions and not in cryogenic conditions. And when the pump goes into cryogenic temperatures, there will be contraction. Since there are different materials, there would be dissimilar contractions. So tolerances are provided so that the bearing will not stop. But we found that the calculation might not have been done accurately. But the issue is, if it touches [something],will it stop, because there is a lot of power given to it? So we decided to test this in cryogenic conditions. A test facility was created. We did not take [the theory] for granted.

The second scenario is that a welding could have failed and the casing could have come out. We would have had a similar condition then. The possibility of a casing coming out was, however, very remote. But still we redesigned it. We made a casting.

The third scenario was contamination. We did not want to get locked on to the contamination theory because then we may not see the other things. The PAS is imported and is kept in a sealed cover. When we vibrated it, we found foreign particles coming out. Then we had to do a lot of cleaning and so that was the reason. So we decided to redesign it and that is what we used in D5. Otherwise, the liquid hydrogen tank itself could provide that contamination. All these three issues that we came across have been corrected. So it is a component-level problem and a not system problem. In F06, it was an inspection problem. The cryo-stage shroud, which is expected to move a little bit during the vehicle movement, is supposed to be provided with a lanyard of about 15 mm. But it was almost not there. It was only about 1-2 mm, resulting in greater tension. Two connectors are provided. But both connectors came out. And then we lost the signal from top to bottom. This is what happened actually.

The question is what did we do to address these. Compared with the PSLV, in the GSLV we have a large number of fluid components that have to work in flight. So we make the system, test it and use it and then after some time fly it. In this process, the leak rate would increase sometimes if there are very small defects. So getting a component assembled and tested properly becomes very important. We have tightened that now. In the recent PSLVs also, we have a good record in this area. We have introduced a zero-defect delivery system, which essentially boils down to the person who assembles the component. The technician who assembles the component should be aware of the impact of even a small scratch on a sheet or a dust particle coming into the system or something he might miss during the assembly. We introduced training about one and a half years ago and it is given in situ at the work spot in the local language with examples. When we say that the components have all performed to specifications during the entire campaign, it is actually the result of this. So this is the lesson from the GSLV. Otherwise, the GSLV per se, compared with the PSLV, is a better vehicle. Cryo is complex. Leave that part. If you take the bottom stages, the number of propulsion systems and control systems that come into the picture is far smaller. The only issue there is the solid motor hardware, which will have to be carried for nearly 40 seconds by the strap-ons because that does not separate. The advantage is that proven stages are being used here.

What is the next important milestone for the GSLV?

The immediate thing is GSLV MkIII, the experimental mission with the passive cryo stage.

What do you mean by passive cryo?

No engine will be burnt in the third stage. Actually, if you look at the GSLV, 50 per cent of the velocity is given by the non-cryo portion. So we will get about 5 km/s velocity, and it will be a suborbital flight. But what we want to test here is the atmospheric phase of the flight. While it is coming down, we will use it to characterise the crew module. We can measure the thermal stress when it is coming down. As far as the vehicle is concerned, its exterior will be ditto. Internally, the cryo will be passive.