Chief Operating Officer / Australia
Decisions made before the worst day arrived
- Customer connections
- 418,000
- Wyndarra Power Networks, June 2026
- Overhead line in operation
- 74,000 km
- Including 21,000 km of single-wire earth return to isolated farms
- Annual operating and capital spend in my remit
- A$935m
- 2025–26: A$395m operating expenditure and A$540m capital delivery
Decisions & consequences
01Closing Bellcarrick on a date announced thirty-one months ahead
In early 2015 the board of Telmora Energy had to decide the future of Bellcarrick Power Station, a 1,320-megawatt coal station in the Hunter Valley whose four units were between 43 and 46 years old. The station employed 412 people and about 160 resident contractors, and it was the largest private employer in its town. Its boilers needed about A$180 million of work to run beyond 2018, which the company could not justify.
Three options went to the board. Finance proposed closing Units 1 and 2 in 2016 and running the others until a major failure, saving about A$38 million of fixed costs over two years. Trading wanted no date at all, so that units could stay available if wholesale prices rose. I proposed closing all four units together on a fixed date, told to the workforce before the market, with no forced redundancies before closure and a fund for retraining. The unions wanted a later date than mine. The chief financial officer said my proposal gave away an option the company would regret, and in 2017 he was briefly right.
The board chose a single closure on 30 April 2018 and announced it in September 2015. We gave up the early savings and the freedom to keep units running. We created a A$31 million transition fund covering retraining, independent financial advice and relocation, and kept full maintenance on safety-critical systems to the last day. When prices rose sharply in 2017, the trading team asked to keep two units open for another year. I recommended against it, because hundreds of people had planned their lives around the date and our maintenance had been planned to it too. The board agreed.
All four units shut on the announced date, with no lost-time injury in the final 22 months. Of 412 employees, 141 were redeployed within Telmora, including 63 into the site's rehabilitation programme; 102 retired; 131 were in other work within twelve months; and 38 were still seeking it. Telmora's later closure plans used the same structure.
What I would do differently concerns the contractors. Our commitments covered employees only. About 160 contractors, many with more than a decade at the station, lost their work with a few weeks' notice and far less support. I should have made their employers' participation in the transition fund a contract condition from 2015.
A Catastrophic-day switch-off rule, published before summer
After the 2019–20 fires, Wyndarra's board asked whether the network should ever switch off power ahead of a dangerous fire day. Our protection settings already disabled automatic reclosing on high-risk days, but a clashing or fallen conductor could still start a fire before protection operated. The 47 feeders with the highest fire consequence supplied about 9,800 customers in steep, timbered country where a fire that started on a Catastrophic day could not be stopped.
The recommendation was mine and the decision the board's. Our chief executive worried that switching off without a fire nearby would look like the network protecting itself at its customers' expense. District officers of the state's rural fire agency warned that de-energising would cut town water pumping, telecommunications towers and the electric bore pumps farmers use to fill firefighting tanks, exactly when people needed them. Our regulatory team pointed out that the outages could count against our reliability performance. Each of them was right about the cost.
My proposal, adopted in September 2022, had three parts. The rule and the feeder list are published every September: if a feeder's district is forecast at Catastrophic, it may be de-energised from 9 am, with 48 hours' notice wherever the forecast allows. Before any switch-off, backup supply must be ready for the 31 water pumping sites and 19 telecommunications sites on those feeders. And every registered life-support customer is phoned, not texted, and offered a portable battery unit or help to stay elsewhere. We gave up reliability minutes and committed A$6.2 million of capital to backup supply.
The rule has been used on three Catastrophic-rated days across the 2023–24 and 2024–25 seasons: 61 feeder switch-offs averaging thirteen hours and affecting 21,400 customer connections in total. No fire started from a de-energised line. On the same days, two small fires started from our equipment on feeders not on the list, and both were contained. We paid A$412,000 in goodwill payments, mainly for spoiled milk and food.
The rule is now reviewed each winter with the fire agency and the water utilities. What I would do differently is the contact data. On the first day our text messages failed to reach 12% of affected customers because their details were out of date. We should have checked every number before the season, not after it.
Holding restoration estimates for a day after the November 2024 storm
On the evening of 19 November 2024 a line of severe thunderstorms with downbursts crossed the Southern Tablelands and the eastern Riverina. By midnight 96,000 of our customers were without power. Over the next two days crews and patrol helicopters found 412 broken poles and about 1,260 spans of conductor down, much of it on rural lines behind fallen timber.
Three years earlier, in a smaller storm in December 2021, we had published estimated restoration times within hours, calculated from the size of each outage rather than the damage behind it. We met 41% of them. Farmers hired generators and moved stock around times we did not keep, and our call centre spent the week apologising. In 2024 the stakes were higher: the shareholding ministers' offices wanted a single date by which everyone would be reconnected, and so did regional newspapers and radio.
I decided that for the first 24 hours every affected area would show a status of assessing damage and no times at all; after that we would publish a range for each town and update it at 7 am and 7 pm. Our chief executive questioned the plan and then backed it. Our customer and communications team opposed it, arguing that silence would be read as incompetence and that people would prefer an imperfect time to none. We gave up the reassurance of early numbers and accepted a very hard first day of calls.
Restoration followed our published order: hospitals, water and sewage pumping and telecommunications first, then the feeders returning the most customers per crew-hour. We brought in 260 line workers from two other networks and sent 38 mobile generators to the towns facing the longest waits. Ninety per cent of customers had power within four days, and the last were reconnected on day eleven. Of the ranges we published, 74% were met.
We now hold estimates in every major event until damage patrols report, each estimate is signed by the regional field manager who must meet it, and we publish our hit rate afterwards. What I would do differently is the isolated single-wire lines. Farms at the ends of those lines were last in our order, and some were off for eleven days. We should have had generators at those properties by the third day instead of the sixth.
Paying vegetation contractors for spans made safe
In 2020, after the fires, a LiDAR survey of our high bushfire-risk areas found that 11% of spans did not meet our vegetation clearance standard at the start of the fire season. We were spending A$96 million a year with fourteen vegetation contractors, paid mainly per kilometre cleared on a cyclic programme. That structure rewarded distance, so crews moved quickly through easy country and lightly through steep, heavily timbered gullies where the fire consequence was highest. Contract crews also had an injury rate more than twice that of our employees.
I proposed consolidating to five regional contractors on five-year contracts, paid for each span certified compliant by a later LiDAR flight and an independent field audit, with a bonus for high-risk spans finished before 1 October and deductions for spans later found non-compliant. Our procurement team preferred a competitive tender on per-kilometre rates, which would have lowered cost further. Several regional managers argued for keeping the local firms who knew the terrain. Six small contractors lost their direct contracts, and two mayors wrote to our chief executive about the jobs involved. We accepted that, and an increase in annual cost to A$104 million.
The five contractors had to adopt our life-saving rules, report injuries in the same system as our employees and offer subcontracts in each region to existing local firms, four of which eventually took them. We paid for felling and electrical-awareness training for 140 contract workers.
By spring 2024 non-compliant spans in high-risk areas had fallen from 11% to 1.8%, and outages caused by vegetation were 37% below the 2017–2020 average. The recordable injury frequency of vegetation contractors fell from 9.1 to 4.3 per million hours. Payment per certified span is now the basis of our pole inspection and civil works contracts as well.
What I would do differently is the transition for the small firms. The subcontracting requirement was added a year after the new contracts began, by which time two of the six local firms had already laid off crews. It belonged in the original tender.