AC to DC Line Conversion – It’s time to think about it

Speculation over conversion of ac lines to dc goes back almost as far as HVDC itself. It’s an idea easily quashed by the fact that you pay for terminals rated at total DC MW and gain only the incremental capability over the ac case. Furthermore dc conversion idled a third of the ac transmission asset. And even if you’d cleared those hurdles, engineers were reluctant to take full advantage of dc’s control capability – a necessary step to make full use of dc’s capability in an ac system context. In any case construction of new ac lines were a vastly less inexpensive way to expand transmission capability.

So what’s new?

  • New transmission is extremely expensive or can’t be built at all.
  • FACTS, reconductoring, temperature monitoring are helping stretch ac capability but within limits  and often at a high cost in losses.
  • Communication and control technology has made huge advances. Wide Area Management Systems (WAMS), coupled with very sophisticated FACTS options, are polite to talk about.
  • An HVDC configuration has been introduced which makes full thermal use of all three phase positions of an ac line.

It’s time to look again at the conversion prospect; perhaps with a mind-set that considers dc as the ultimate FACTS device –  one that controls flow and boosts transfer capability of existing, in-place assets by up to 4:1.

DC on all three phase positions…without earth return current?

The principle is very simple.  In fig. 1a, suppose Pole 1 and Pole 2 operate as a bipole, at the maximum allowed conductor rating. Pole 3 is idle. Power is 1.00

In fig. 1b let pole 3  relieve a portion of the positive current in pole 1, then a portion of the negative current in pole 2; alternating back and forth every five minutes or so to prevent conductor temperature rise of more than one or two degrees C.  Power is still 1.00 but the average heating in all poles is below rating.

So increase the current (and power) and adjust the ratio of current levels until all three poles are fully loaded thermally (Fig. 1c). Power is now 1.37

No ground current flows in any of the above cases. Upon loss of either a pole in the tripole configuration, it reverts to a bipole with a rating of 1.00; or considering overload capability, something like 1.15 – a loss of just 16% of full power rating compared to a drop to 50% (57% with the same overload assumption) – a drop of 43%.

The “tripole” HVDC configuration shown in fig. 2  is basically a parallel bipole and monopole configuration except that (a) all thyristors and cooling systems are rated 1.37 times the bipole rating (Note 1); other equipment remains the same and (b) pole three has back-to-back antiparallel valves.  Those differences also give the tripole an inherently higher short term overload capability.

The above characteristics translate into higher (n-1)-constrained loading of parallel ac lines and greater total path transfer per kW of terminal rating. They also permit a higher allowable DC rating within (n-1) constraints.

While the tripole configuration will cost approximately 15% to 20% more per kW than a bipole, the gain in power boost means that the cost per incremental kW over the former ac case will drop.

DC conversion vs. AC system fixes

It’s relatively inexpensive to install switched capacitors, FACTS devices, or phase-shifting solutions, thereby optimizing both pre- and post-contingency load distribution among ac circuits; certainly compared to new line construction. However a realistic evaluation of the net present value of the increase in losses resulting from boosting ac loading on an already heavily loaded system will show that cost can be an order of magnitude greater than the cost of hardware installed to achieve the increase. And, irrespective of ac system fixes, fast or slow, (n-1)-constrained dispatch will always be limited by post-contingency thermal capability of one or more circuits. Increasing those thermal limits or adding new circuits soon becomes the frontier – the only way to increase (n-1) constrained dispatch.

There are relatively inexpensive ways to extend thermal limits e.g. dynamic conductor rating systems, sag monitoring, and sag mitigation devices. But where does one go when the last trick is played?  Reconductoring? – An attractive option but if it increases only maximum thermal rating and not conductivity (e.g.  low-sag conductor options) the gain may again come with heavy penalty in losses, particularly for longer lines.

HVDC achieves a major boost in path capability

Conversion of an ac circuit to HVDC may not be as expensive as it seems. It achieves a very high boost in thermal rating, modest increase in losses, and provides the flow control necessary to take full advantage of that increase.

DC operates at peak voltage, giving a √2 advantage in power level. But, that advantage is reduced by 2/3 for the bipole since one phase position is rendered inactive by conversion. The resulting multiplier is just .94.  By using the “tripole” system, the multiplier becomes1.3. Neither boost is heroic in its own right, but there are four additional multipliers to consider:

  1. The DC voltage sustainable by an ac line is often significantly higher than line-to-ground crest of ac
  2. DC power flow is controlled independent of the AC system dispatch and can therefore contribute the equivalent to its maximum thermal rating to total dispatch
  3. If DC is configured to assume a high temporary overload, it can increase the (n-1) constrained loading on parallel ac circuits, thus multiplying the effectiveness of the converter terminal investment
  4. Power factor limits utilization of an ac for real power transfer. While DC requires reactive support (Note 2) , the full capability of the transmission line conductors is used for real power flow.

These factors can push the dc/ac multiple to as high as 4:1 and do so without any physical modification of the line other than live-line replacement of ac insulators with dc units of the same length.

Conversion of a critical line to HVDC can be achieved without removing it from service by constructing terminals abutting existing substations, temporary off-hour switch-over for commissioning tests…then final switchover to DC operation.  If one simultaneously reconductors, the leverage achieved by ac reconductoring can be multiplied by the ac to dc power ratio. It makes sense to initially switch over to bipole HVDC, leaving one phase position idle for reconductoring, sequentially freeing the other two phase positions for the same purpose, and then going to tripole operation.

Conversion of a critical line to HVDC can be achieved without removing it from service by constructing terminals abutting existing substations, temporary off-hour switch-over for commissioning tests…then final switchover to DC operation. If one simultaneously reconductors, the leverage achieved by ac reconductoring can be multiplied by the ac to dc power ratio. It makes sense to initially switch over to bipole HVDC, leaving one phase position idle for reconductoring, sequentially freeing the other two phase positions for the same purpose, and then going to tripole operation. Conversion may have other important advantages:

  • DC reduces rather than increases short circuit duty

  • DC and can send more dynamic support per unit of rating than ac, and can supply it faster. AC synchronizing power flows as a consequence of a worsening situation (greater angle separation) while dc acts upon recognition of the problem.

Making full use of dc’s capability requires closer integration of control logic with ac system operating data than is usual, but this is exactly what the move to more sophisticated FACTS controllers and WAMS in general call for.

Is there an economic case for conversion?

If one of four identical ac lines is converted to tripole at a voltage no greater than ac line-to-ground crest, the resulting (n-1)-constrained path flow in increased as much as it would be by construction of a fifth ac line. If the dc voltage can be increased 25% above line-to-ground crest, the increase is equivalent to construction of 1.4 new circuits. If, in addition to increased voltage (made easier by reconductoring) new conductors of twice the original thermal rating are installed as a part of the conversion project, the (n-1)-compliant result would be equivalent to almost four new ac circuits.

It is worth doing the arithmetic in specific cases, particularly where lines are relatively long…or made to be long by abstracting one of several parallel circuits between successive busses, making an “express” dc circuit and feeding intermediate loads from the lines remaining in ac service. (fig. 4)

Today’s planning context suggests that dc conversion alternatives should be included in studies to extend the usefulness of existing transmission line assets.

Conversion Example

Suppose a 138 kV line with a maximum current rating of 900 amperes is convertible to a +/- 138 kV dc triple line. Suppose further that n-1 constraints limit ac flow to 70% of the line’s thermal limit and that the ac power factor is 0.9. Then the ac MW flow is:

The tripole power with the same conductor system, operating at the thermal limit is:

That’s a boost of 2.5 times, neglecting the increase in ac loading made possible by the high emergency pick-up capacity of the tripole system. The tripole would actually operate below 340 MW, e.g. 280 MW, shifting the difference (60 MW) onto the ac dispatch to minimize losses but maintaining the same n-1 compliant total path flow achievable with the 340 MW operation.

Notes:

  1. Since thyristor valves come in only two standard sizes, 100 mm = 2.1 kA and 125 mm = 3.5 kA, the penalty in valve cost is significant only when the 1.37 factor drives one to the higher option.
  2. While conventional HVDC, bipole or tripole, is difficult to apply to systems of low short-circuit ratio, VSC-based systems, can also be adopted to tripole by using thyristors on the modulating pole, thus gaining the advantages cited above even with service to a dead load. [3].

References:

  1. L.O.Barthold, H.K.Clark, D. Woodford, “Principles and Applications of Current-Modulated HVDC Transmission Systems,” IEEE T&D Conference, Dallas, Tex. May 21-26 2006.
  2. L.O.Barthold, “Technical and Economic Aspects of Tripole HVDC” Powercon2006; 22-26 October, 2006 Chongqing China
  3. L.O.Barthold, D. Woodford, “Application of Voltage Sourced Converters to Tripole HVDC Transmission,” 9th FACTS Users Group & Task Force Meeting, Montreal Quebec, Sept. 6-8, 2006

For questions, comments and further discussion, contact us at mailto:info@pterra.us or Lionel at imod@adelphia.net