Let me please continue a series of “green posts” on transients. One of the main reasons for installing surge arresters (SA) is to protect switchgear from lightning surges caused by lightning discharges to connected overhead lines (OHL). In a switchgear, the most valuable equipment is a power transformer, and it is next to it that we always try to place SA (say SA1). There are various options for placing SA1:
➡️ on the transformer tank;
➡️ on the portal above the transformer;
➡️ in front of the transformer (bottom photo).

Each of options is characterized by:
✅ the distance between SA and transformer (Lt);
✅ the length of the current path from phase wire through SA to the ground (Lsa), which also depends on the length of the SA connecting “leads”.

Let us give example of statistical calculations performed in EMTP using the methodology described in the book “Lightning overvoltages in high voltage switchgear”. Suppose we have a 500 kV switchgear, and consider lightning strike to OHL phase wire, at the input to the switchgear. The parameters of such lightning vary statistically, as described in the book.

Two cases were calculated in EMTP:
👉 Lt = 0 (SA1 on the tank or above it).
👉 Lt = 20 m (SA1 in front of the transformer).

Each of two graphs has:
➡️ abscissa axis: the voltage on transformer (Ut);
➡️ ordinate axis: the probability (P) that voltage (Ut) will exceed the value from abscissa.

For a 500 kV transformer, the voltage level of 1500 kV can be considered as acceptable, and then the intersection of curves with this this level is the probability (P) of dangerous overvoltage. For example, P=0.1 means that there is danger in one out of every 10 strikes (here we will not recalculate this to the number of years of operation).

The results of processing the two graphs are shown in the lower left corner. It can be seen that:
✅ at Lt=0, increasing the length of SA leads (Lsa) significantly increases the probability (P) of damage to the transformer;
✅ at Lt=20m, the effect of SA leads (Lsa) is no longer so noticeable;
✅ the probability of damage at Lt = 20m is greater than at Lt = 0m.

What does it mean? This means that lengths (Lt, Lsa) really affect the probability (P) of damage to the transformer. The closer the SA is to the transformer and the shorter its leads, the better. For example, it should be avoided to hang the SA above the transformer (since it gives great Lsa).

Sometimes we see that the SA is simply placed on the transformer tank or close to it, because it gives minimal Lt and Lsa. However this solution can be dangerous in case of SA failure, and the question arises – is it possible to keep SA at the distance from transformer?

The answer is YES, if we use so-called “cascade protection” where there is a second SA (SA2) installed at the input of the OHL to the switchgear. Arresters SA1 and SA2 together provide excellent protection of transformer, regardless of the specific length of Lt and Lsa.

For example, if Lt=20m and Lsa=10m, then installing SA2 (L12=100m) reduced the probability of damage to the transformer from 0.25 to 0.05, that is, by 5 times. The probability P is even less than if the only SA1 was installed directly on the transformer (curve “Lt=0m”).

Why cascade circuits are so good for protection of power transformers is described in detail in the book. Knowing this circumstance, it is strange to observe how people put arresters directly on the transformer tanks, and then get upset when failure if one of these arresters leads to the serious damage to transformer bushings.

I would also like to point out that a certain value of the post lies in the fact that there are almost no known cases in the world where EMTP has been used for statistical calculations of lightning processes. Usually, engineers who use EMTP simply consider a lightning strike with specific parameters, which leads them to get an incomplete picture of the processes and even incorrect conclusions above surge protection.

Let me please recommend you the book “Lightning overvoltages in high voltage switchgear”. It is in the list of books. Thank you