First SSLV flight fails to put satellites in correct orbit

ISRO Chairman is confident of resolving the problem and being ready for the next launch.

The launch of SSLV-D1 from the First Launch Pad at the Satish Dhawan Space Centre, Sriharikota, Andhra Pradesh, on August 7. | Photo Credit: ISRO

ON August 7 at 09:18 hrs, the new rocket of the Indian Space Research Organisation (ISRO) called the Small Satellite Launch Vehicle (SSLV) took to the skies for the first time from the First Launch Pad at the Satish Dhawan Space Centre at Sriharikota in coastal Andhra Pradesh. About 10 minutes into its flawless trajectory, the rocket failed to place the two satellites it was carrying in the correct orbit. Instead of the planned circular orbit at an altitude of 356 km, it injected the satellites into a 356 km × 76 km elliptical orbit.

The launch of SSLV-D1 from the First Launch Pad at the Satish Dhawan Space Centre, Sriharikota, Andhra Pradesh, on August 7. | Photo Credit: ISRO

The SSLV is specifically designed to launch small—Mini, Micro, or Nano—satellites (10 kg to 500 kg mass) up to a 500 km planar orbit. ISRO developed the SSLV to serve the emerging global market for launching small satellites. According to ISRO, the SSLV offers, among other things, a much lower turnaround time compared with the Polar Satellite Launch Vehicle (PSLV), flexibility in accommodating multiple satellites, launch-on-demand feasibility, and minimal launch infrastructure requirement.

In an interview to Frontline in September 2018, former ISRO Chairman K. Sivan said about the SSLV: “We are adding new technology. The vehicle will be more and more autonomous.… It will take the lowest integration time.” SpaceTech Asia quoted Sivan as saying: “[The SSLV] will be assembled in 72 hours instead of the 60 days [for PSLV]; instead of 600 people, it will be done by 6 people.”  

The August 7 launch was the first developmental launch, or D1, of the SSLV. ISRO has planned three developmental launches. Mission SSLV-D1 was intended to launch a 135-kg earth observation satellite (EOS-02) and a small eight-unit CubeSat satellite (built by students and called AzaadiSAT) weighing about 8 kg into a low earth orbit at an altitude of about 350 km at the equator with an inclination of about 37 degrees.

Credit: NewSpace India Ltd/Department of Space

The SSLV has three stages weighing 87 tonnes, 7.7 t, and 4.5 t respectively, which are powered by the solid propellant HTPB. The satellite insertion into the intended orbit is achieved through a liquid propulsion–based Velocity Trimming Module (VTM) to correct for the dispersion in velocity that can occur at the time of injection. The trimming operation uses sixteen 50 newton bipropellant thrusters. Thus, the SSLV effectively has four stages. According to a former ISRO scientist, the VTM can correct for a velocity shortfall of up to 172 metres/s. The rocket used for SSLV-D1 was a 34 m tall, 2 m diameter vehicle having a lift-off mass of 120 t. NewSpace India Ltd, a public sector company under the Department of Space, manufactured the vehicle apparently at a cost of around Rs.30 crore.

At the Mission Control Centre at the Sriharikota complex, right through the hour-long event, the mission director made announcements at the conclusion of every significant phase. However, the eagerly awaited announcement of a successful launch by ISRO Chairman S. Somanath did not happen, and instead, he made this statement: “The performance of all the three stages of SSLV and their scheduled separations was as expected. [However,] at the terminal phase of the mission, we had some telemetry data loss occurring and we are analysing the data. We will come back with the status of the satellites as well as the launch vehicle to conclude the outcome of the mission if the [intended] stable orbit was achieved [by the satellites] or not.”

Even before the Chairman’s statement, spaceflight trackers had noted that the terminal VTM stage had underperformed. In a tweet at 10:27 am, Jonathan McDowell (@planet4589), a well-known astronomer who tracks all space launches, said: “If the fourth VTM stage did not fire, vehicle would complete about half an orbit and fall in the Pacific near the third stage NOTAM [Notice to Airmen] zone around 138W 30S after passing over Australia and New Zealand [which would be somewhere in the South Pacific area]…. [The] [t]hird stage (and possibly [the] 4th stage and payloads) would have impacted the Pacific around 0431 UTC [about 10:01 am IST].” McDowell also posted his estimated south-eastward trajectory of the SSLV from the launch site and the final leg of the trajectory after the anomaly in the VTM stage.

Highlights

• Maiden flight of ISRO’s new rocket SSLV.
• A global market for the launch of small satellites is emerging.
• SSLV’s payloads: a 135-kg earth observation satellite and a small eight-unit CubeSat satellite (built by students and called AzaadiSAT) weighing about 8 kg.
• They were injected into a circular orbit instead of the planned elliptical orbit, which caused them to de-orbit.
• The problem was likely caused because the computer read a two-second anomaly in the accelerometer as a failure.

At about 11:45 a.m., ISRO stated: “All the stages performed normal. Both the satellites were injected. But, the orbit achieved was less than expected, which makes it unstable.” At 2:48 pm, ISRO provided an update on the mission outcome with a couple of cryptic tweets: “SSLV-D1 placed the satellites into 356 km x 76 km elliptical orbit instead of 356 km circular orbit. Satellites are no longer usable. Issue is reasonably identified. Failure of a logic to identify a sensor failure and go for a salvage action caused the deviation. A committee would analyse and recommend. With the implementation of the recommendations, ISRO will come back soon with SSLV-D2. A detailed statement by Chairman, ISRO will be uploaded soon.”

S. Somanath, Chairman of the Indian Space Research Organisation.    | Photo Credit: K. MURALI KUMAR

In his video statement on the ISRO website, Somanath elaborated on those remarks but not with enough detail to allow a full understanding of the problem. He said that when satellites are placed in an oblong or elliptical orbit, the orbit will not be stable for a long time owing to atmospheric drag and will come down. Significantly, he confirmed that the satellites had already de-orbited and were therefore not usable. He added that, but for this anomaly, the performance of the entire vehicle was good and no other anomaly could be seen.

“Every other new element that has been incorporated into the rocket has performed very well including propulsion stages, its hardware, aerodynamic design and new generation low cost electronics, control systems, new separation systems, the entire architecture of the rocket—everything has been proven in the very first time and we are very happy about that part. What we are going to do is to identify this specific problem why this isolation [of the software logic and the sensor data] happened and why it failed to place the satellites in the proper orbit,” Somanath said.

The Chairman’s interview to The Hindu (published on August 12) provided a clearer picture. Basically, “an anomaly” for a mere two seconds in one of the accelerometers (which measure the vehicle’s acceleration in the forward direction) in the second stage caused the SSLV’s guidance, navigation, and control (GN&C) software to drive the launch vehicle from the nominal “closed loop guidance” (CLG) mode—when there is constant feedback from the sensors—into an “open loop guidance” (OLG) mode when the sensor/accelerometer data get totally isolated from the GN&C software. This led to the underperformance of the mission though the exact how and why of it remains to be unravelled and understood.

When a subsystem of the launch vehicle, such as a sensor/accelerometer, is not working nominally, the logic of the launch vehicle has a salvage option, which means that it is designed to put the satellite into orbit despite a subsystem failure, even if this happens at lift-off. “In this case,” Somanath said, “what happened was that the measurement of the accelerometer showed some anomaly just at the point of separation of the second stage… [and] the internal computer felt that the accelerometer had failed. Then it triggered… the salvaging operation.”

The planned flight profile of SSLV-D1. As indicated in the graphic, the figures are based on the Flight Readiness Review (FRR) and are not the final planned flight profile. | Photo Credit: ISRO

From this point to the final salvaging, Somanath explained, the vehicle worked in the OLG mode and the GN&C software just followed a trajectory that was already loaded in the computer without the reference of the accelerometer data. According to him, once it is on this predetermined path, the ability to put the satellite in the correct orbit is slightly diminished. In the salvaging mode, the computer waits until the ignition and burning of the next stage—in this case the third stage, which also is a solid motor—are completed and “salvages” the mission by separating the satellites and injecting them into orbit before the VTM comes into play. The tracking plot of the flight, as seen in the mission video, shows that this occurred 653.5 s into the launch, that is around 9:29 am. Thus, the vehicle software did not call the VTM into operation and, therefore, it did not fire at all.

While the salvaging operation did what it is supposed to do, that process resulted in a velocity shortfall of about 60–70 m/s. The satellites should have been injected into orbit at 7.3 km/s, but the launch vehicle had attained only over 7.2 km/s, which shortfall became critical because the perigee came down to the extent that its height was less than atmospheric height. As a former ISRO scientist pointed out, if it had not been for the salvaging operation, the VTM could have easily rectified the velocity shortfall and the mission would have succeeded.

No issue with the accelerometer

Interestingly, however, there was really no issue with the accelerometer, according to Somanath. It apparently became normal after the two-second anomaly. “Why the computer found that the accelerometers had an issue [within two seconds] is something we do not understand,” he said. “There could be an actual problem in the sensor. Or there could be a logic problem in the sensor. But this is a standard system that exists in all ISRO rockets.... Whenever rocket stages are separating, there will be a transient. A small jerk will be there during which time the accelerometer level is slightly exceeded. For two seconds this anomaly existed.”

Unfortunately, the computer declared that it was a failure within those two seconds. Somanath believes that if the threshold for isolation was a little higher, say three seconds, it would not have happened. “There could be many reasons [for the isolation after two seconds]. There could be a hardware failure, a software glitch, an external trigger or the shock of the transient was slightly higher than what we expected because it is a new rocket.… So we should relook at whether the approach, the logic, to isolate [the sensor from the software] or not to isolate can be suppressed,” he added.

A Failure Analysis Committee will, of course, go over all the issues and make recommendations. However, since the problem has been reasonably well identified, the ISRO Chairman is confident that it will be resolved and the organisation will be ready for the launch of SSLV-D2 soon.