Switch off the feeder only after you have powered the pump

Every September I sign a document that tells about 9,800 households, farms and businesses that we may switch off their electricity on a day when the fire danger is at its worst. It lists 47 feeders by name, the districts whose forecasts govern them, the hour at which we would act and what we will do for people who depend on power to stay safe. I sign it with some discomfort. My position is that a rural network should be willing to de-energise high-risk lines on Catastrophic-rated days, but only on three conditions. The rule must be published before the season. Backup supply for water, telecommunications and life support must already be in place. And the network must count the harm a switch-off causes with the same care it gives to the fires it is meant to prevent. Without those conditions, a switch-off simply moves risk from the network to its customers.

I hold this view because I have seen both alternatives. Before 2022, our approach on the worst days was to change protection settings: we disabled automatic reclosing on high-risk feeders and made earth-fault protection more sensitive, so that a fault would trip a line and keep it off until a crew had patrolled it. That reduces risk considerably, and we still do it on every day rated Extreme or above. It does not remove the risk of the first arc, when a conductor clashes or a branch falls across a line in a sixty-kilometre wind. During the 2019–20 season one of the fires in our area began in exactly that way. The protection worked as designed. The fire started anyway.

The other alternative is to switch off whenever the weather looks frightening, decided on the morning by whoever is in the control room. Networks overseas have tried versions of this, and the lesson I draw from their experience is that unannounced, frequent and poorly supported shut-offs destroy the public consent on which the tool depends. People will accept losing power to prevent a fire. They will not accept losing it without warning to protect a company's liability, and they will not accept it at all if it empties the water mains the fire brigade needs.

What we publish before the season starts

Our rule is short. If the fire danger rating for a feeder's district is forecast at Catastrophic, the feeder may be de-energised from 9 am, and we aim to give 48 hours' notice. The decision is made by the duty executive on the advice of our fire risk team, and it is reported to the board's safety and network risk committee the following week, whatever the outcome. Power is restored only after crews have patrolled the line, so customers should expect to be off until evening and sometimes overnight.

The feeders are chosen by consequence, not by fault history. We model how a fire starting at each span would spread under Catastrophic conditions, then select the feeders where an ignition would reach homes and towns fastest and where suppression would be least possible. Forty-seven of our more than 1,100 feeders meet that standard. The list changes each winter as we underground or cover lines and as the model improves, and this year it will grow by twelve.

Publication matters for reasons beyond courtesy. A rule that lives only inside the control room can be bent on the day, towards caution when an executive is nervous or towards inaction when the phones are ringing. A published rule can be tested by the communities it affects. Councils have asked us why a feeder is on the list, and twice they have shown us that our model misread local terrain. Farmers have told us which days they would rather we did not choose, which is information we cannot act on but are glad to have.

Water, radio and the person on oxygen

District officers from the state's rural fire agency gave us the most important objection when we first proposed the rule. A switch-off that stops a town's water pumps leaves the brigade with empty hydrants. One that stops a telecommunications tower leaves people without warnings and without the means to call for help. One that stops a home oxygen concentrator puts a life at risk before any fire has started. They were right, and our rule now says that no listed feeder is de-energised until backup supply is in place for every water pumping site and telecommunications site it serves.

That condition cost about A$6.2 million in capital for permanent generator connections and standby units across 31 water sites and 19 telecommunications sites, with a further annual cost for fuel, testing and crews. It also required negotiation with water utilities and carriers about who owns each generator, who tests it and who answers the phone when it fails. Those conversations took longer than the engineering.

Registered life-support customers receive a phone call, not a text message, at least a day ahead. We offer a portable battery unit, help to stay with family or at a community facility, or, in a few cases, a generator delivered to the house. On our first switch-off day, text messages failed to reach 12% of affected customers because their contact details were out of date. Every one of our calls to life-support customers connected, because those numbers had been checked. That difference told me where to spend the following winter.

Counting what a switch-off costs

The rule has now been used on three days, with 61 feeder switch-offs averaging thirteen hours and affecting 21,400 customer connections in total. No fire started from a de-energised line. That sentence is easy to write and impossible to claim as a proven benefit, because we cannot know which of those lines would have started a fire. On the same days two small fires started from our equipment on feeders that were not listed, and both were contained. I report those beside the switch-offs, because they are the best evidence we have about whether the list is right.

The costs are easier to count, and we count them in detail. We paid A$412,000 in goodwill payments, mostly to dairy farmers who lost milk and households who lost food. A small business in Bombala lost a day's trade in the week before Christmas. Two aged-care homes ran on generators for longer than their plans assumed. Our reliability figures were worse in both seasons. I do not regard any of those as reasons to abandon the rule. I regard them as its price, and I want our board and our customers to see the price in full.

There is a version of this argument that becomes an excuse for spending less on the network. A utility able to switch off on the worst days might be tempted to defer covered conductor or undergrounding on exactly the lines that matter most. I think that would be a serious mistake. The list should become shorter as lines are made safer, and our board measures it that way. A switch-off rule is a bridge to a safer network, not a substitute for one.

My advice to any operator considering a similar rule is to begin with the people who will lose power, not with the fire model. Find every water pump, radio tower and life-support customer on the feeders you might switch off, and make sure each has power before you decide when to remove it. Publish the rule before the first hot day. Count every cost. Then, on the morning the forecast says Catastrophic, follow the rule you published.