Behind the collapse

The grid disturbance in north India can be traced to overloading of the transmission lines and poor load management.

IT is now amply evident that the grid collapses of July 30 and 31 were not due to overdrawing of power by the northern States, as has been widely reported by the media. Violating grid discipline and overdrawing may be regular phenomena, but the National Load Despatch Centre (NLDC), in coordination with the power generators and the Regional Load Despatch Centres (RLDCs), is expected to take prompt action by isolating some lines or have more power pumped into the grid to restore the balance between supply and demand.

The grid frequency is a reflection of this balance at a given instant, and the frequency would drop if there is an imbalance caused by overdrawing somewhere along the grid. This can affect the quality of power supply and the stability of the grid. To maintain the security of the grid, the permissible frequency band specified by the Indian Electricity Grid Code (IEGC) of the Central Electricity Regulatory Commission (CERC), effective May 3, 2010, is 49.5 Hz to 50.2 Hz.

Media reports, perhaps inspired by official briefings which may have trotted out this argument, have constantly harped on the grid frequency dropping, an indication of large-scale overdrawing from the grid. But record of grid frequency before the collapse, which occurred at 2-33 a.m. on July 30, suggests that there was no such alarming drop in frequency. In fact, the stability of the frequency was extremely good. According to the flash report of Power System Operation Corporation Limited (POSOCO), on the day of the first collapse when the northern grid totally tripped, the antecedent frequency of the grid at 2-32 a.m. was 49.68 Hz against the ideal frequency of generation of 50 Hz.

This, as any power engineer would tell you, is extremely good. But, more importantly, the grid frequency had been extremely stable for hours before that. According to grid frequency data recorded by the Dadri Thermal Power Station of NTPC near Delhi, and accessed by Frontline, the frequency at midnight of July 29-30 was 49.98 Hz and, between midnight and the tripping event (excluding the frequency at the time of tripping), the frequency maximum was 50.09 at 12-50 a.m. and the minimum was 49.72 at 12-25 a.m. This range is well within the tight norm specified by the IEGC.

The standard procedure to restore balance in case of overdrawing, and consequent drop in frequency, is to shed load on the grid by tripping some line or get the generating stations to supply more power to the grid. In this case, however, evidence shows that the observed frequency stability was not achieved by having generators pump in more power to the grid to restore the supply-demand balance and increase the dropping grid frequency. For one, in such a case the frequency would have been seen to be fluctuating before it attains a stable situation, which the data do not show. And, two, generation data from Dadri during that time suggest that generation was actually being reduced. Grid frequency was good and stable on July 31 as well. When the disturbance occurred at 1 p.m., the antecedent frequency at 12-57 p.m. was 49.84 Hz.

Also, the grid is supposed to have standard protection systems, such as relays for automatic load shedding, whenever there is a fall in the frequency or the rate of change in frequency is significant to indicate grid instability (Tables 1 & 2). The frequency variation in the data does not suggest that the situation warranted any such relays to be activated because of frequency drop, nor was any alarm sent out to utilities to take immediate action to restore the balance as, according to news reports, some of the State utilities have categorically stated.

However, according to experts, grid frequency is not the only parameter that warns one of an impending trigger for grid breakdown. There can be other reasons for grid insecurity, such as a single line overloading, which is quite distinct from general overdrawing from the grid that causes frequency instability. But the standard operating procedures for the preventive action are more or less the same; that is, load shedding by tripping a line and isolating it from the grid or making available parallel lines to divert the increased loading in that corridor.

Once the trigger for the collapse occurs, it may be just too late for any preventive action, points out Anil Kulkarni, a specialist in power systems and a professor at the Department of Electrical Engineering at Indian Institute of Technology Bombay (IITB). There may be only 15 seconds after the trigger for the collapse, he added. But protocols for contingency measures exist, and typically such a contingency analysis is done (or should be done) once every 30 minutes or once every hour based on real-time grid monitoring, according to power systems experts. Load scheduling and management for the Indian grids is done by the NLDC and such overloading of lines can occur because of poor load management.

Indeed, this is what seems to have happened on the two consecutive days of grid disturbance. Power overflow in one of the lines linking the western grid and the northern grid triggered a cascade tripping of lines, pulling down the entire northern grid on July 30. The same thing happened the next day as well and this time, the northern grid being synchronously connected to the eastern and north-eastern grids, it brought them down as well. This was compounded by the fact that some eastern circuits were already overloading on July 31. Incidentally, the western grid is also synchronised to the N-E-NE grid and the synchronous N-E-NE-W grid is called the NEW grid. This means that, in principle the western grid could have also been brought down along with the N-E-NE grid. But this was avoided for reasons that will be presently clear.

According to information available from the POSOCO, during both the disturbances, there was heavy power overflow of 1,000 MW on the 400 kV Bina-Gwalior-Agra line. Though there are two circuits available on this line, one is apparently under outage since July 28 for upgradation to 765 kV level. The loading capability or the permissible limit, known as Surge Impedance Loading (SIL), on this single circuit line is only 691 MW. So there was an overloading of the line by more than 300 MW. Similar overflow was seen in many circuits in the eastern sector as well, though not as drastic. The overflows were of the order of 550 MW against an SIL of 515 MW.

Interestingly, there was a warning of an impending disaster the previous day itself when a similar power overflow occurred on the same line at 3-10 p.m. on July 29 resulting in a near-miss situation. The load despatch centre seemed not to have heeded this warning, traced the cause, and remained prepared for such an eventuality. An August 1 directive of the NLDC to the utilities clarifies the nature of this insecure line.

The directive notes that this is the only line through which power transfer between the western grid and the northern grid can take place on this corridor for the following reason. Besides the second circuit of the Bina-Gwalior line, the other two interconnecting lines on this corridor, namely the 400 kV Zerda-Kankroli and the 400 kV Kankroli-Bhinmal lines are under forced outage. This NLDC order, in fact, seeks to regulate the transfer of power between various grids following the collapses of July 30 and 31, in particular W-N transfers.

On July 30, against the post-collapse regulated Total Transfer Capability (TTC) of 1,250 MW, the total W-N transfer that was taking place before the collapse was 2,862 MW. Actually, of the 1,250 MW, the margin available for Short Term Open Access (STOA), or immediate transfer, to the market at the power exchange is now limited to 790 MW, whereas 1,000 MW was being pumped at a given instant on July 30. Similarly on July 31, the total W-N transfer before the breakdown was 2,100 MW. Both are well above the total transfer amount that the NLDC is currently seeking to limit to. In fact, providing for a reliability margin of 200 MW, the Available Transfer Capability (ATC) is now limited to only 1,050 MW. The E-N transfer on the two days respectively were 2,585 MW and 1,825 MW, which are well within the current limit of 3,850 MW for TTC that has been set.

In retrospect, therefore, the cause of overloading of the Bina-Gwalior line and the consequent tripping of the line on July 30 and 31 becomes quite apparent. It was because of excessive transfer of power, more than the allowed limit, by the northern grid from the west without due consideration to the fact that the second circuit was down and to the evident instability of the line especially after the June 29 near-miss. The inter-regional transfers should have been limited by an order after the July 29 incident itself, which the NLDC now seems to be doing after the collapses and blackouts of two consecutive days. The incident also points to the non-availability of sufficient parallel lines to take care of the power exchange market behaviour.

Even though this line was a synchronous link between the western and northern grids, how did the western grid escape the collapse? According to Kulkarni, in such situations of overflow and tripping in inter-regional connecting lines between regional grids, uncontrolled activation of relays can occur and this may have isolated the western grid and its utilities. This is also an indication of the western grids better management and inherent stability, says Prabir Purkayastha, a power and energy systems expert associated with the Delhi Science Forum.

Transmission Loading Relief procedures, as they are called, are standard, which are routinely applied to prevent overflow in real time and contingent cases, according to experts. In this particular case, the real-time loading itself was far excessive and the load despatch centre concerned did not seem to have responded as per laid-down procedures. In cases of overloading, these, as mentioned before, include making other parallel lines in that corridor available for diverting the increased flow or load shedding or asking some utility to increase generation so as to restore the balance. Clearly, none of these was implemented in time, which reflects on the poor load management system in place especially when an increased loading by over 300 MW would not have happened suddenly. It would have taken at least 15 minutes for this build-up to occur, triggering a trip of the line, which would have been sufficient to take corrective action before the line could trigger a trip and consequent grid collapse.

To give the NLDC the benefit of the doubt, Kulkarni points out that even though the despatch centre may have issued instructions for appropriate actions on the part of utilities, they may not have been complied with or it may have taken longer than required at that critical juncture for the instructions to trickle down to personnel in the lower rungs to activate manually the appropriate relays and other preventive measures. It is not clear if the RLDCs, after their transfer to Power Grid Corporation of India Limited (PGCIL), a corporate entity, have the legal authority to ensure that the instructions are promptly complied with, points out Purkayastha.

Interestingly, these two instances are not the first such to occur on the Bina-Gwalior line. This line seems to be having some problem for some time and seems to have been acting like some kind of a blockage in the network. A similar near-miss situation had occurred on November 28, 2009, when the same circuit carrying around 1,000 MW tripped, having developed a fault in one of its phases. This had led to subsequent cascade tripping of other parallel lines in the corridor. However, the system survived because of support from generators and quick action by operators. In the present case, this could not be managed to prevent the collapse. The reasons for this are what need to be established. However, as observers have pointed out, the committee appointed by the government includes officials of PGCIL and POSOCO and thus may not be a true fact-finding one. The committee needs to have some independent power-system professionals from the industry or academic institutions to ensure an impartial inquiry.

August 1 order

The following from the August 1 order also needs to be highlighted. It says:

In several incidents over the past few years, the following concerns have come to the fore in several disturbances involving high antecedent line loadings in the system.

1. Surprises in the form of protective relaying mis-operations either due to incorrect settings, load encroachment or use of distance relays for power swing blocking. These have led to cascading failures in the past.

2. Sustained high loading for a period of 10 minutes or more in daytime high ambient temperature conditions have led to line tripping on transient faults possibly due to increase in sag.

3. High reactive power consumption by the transmission line under heavy loading conditions which leads to voltage dips in the system. Lack of dynamic reactive power consumption adds to the constraint.

It appears that line overloading has been a common problem in the Indian network. In particular, the inherent instability in the Bina-Gwalior line was evident in 2009 itself. But it is also evident that lessons were not learnt and appropriate measures not taken to prevent recurring incidents of this kind.

I think the picture of a weak transmission system is the reason for the grid failure, says Purkayastha. Incidentally, according to him, the Twelfth Five-Year Plan Draft document admits that a grid that is required for open access and trading is different from the grids requirements earlier when its only purpose was to transfer surplus power to deficit regions and shore each other up. Clearly, he adds, one key reason why the grid is not robust enough is the large amounts of wheeling required for the power market. It might have worked for earlier requirements but is not suitable for new market requirements.

PGCIL

When PGCIL earlier known as National Power Transmission Corporation Limited was created in 1989 as part of power sector reforms, the responsibility of the RLDCs, which hitherto was with the Central Electricity Authority (CEA), was transferred to it. In February 2009, the NLDC was established. Now the NLDC and the RLDCs constitute the POSOCO, which is a wholly owned subsidiary of PGCIL. These grid collapse events point to the need for reverting these load despatch units to the CEA, which is where they rightly belong, as otherwise there is an apparent conflict of interest between the two functions of PGCIL, namely, operating the transmission network and load monitoring and scheduling.

The events also underline an important policy aspect. The governments focus in recent years has been only on increasing generation through private investments, ignoring the need to improve the transmission and distribution aspect to match the increased generation. More importantly, unregulated access has been given to the market for inter-regional power trading. Hopefully the sentiment expressed in the Twelfth Plan document is translated into some action.