Seizing The Day (Carefully): Energyaustralia’s Proactive, Risk-focused Approach To AMI


By Ezra Beeman and Boris Kobal

Two complementary initiatives have been launched over a two-year period to test and develop the company’s original $282 million* business case. The first was the Strategic Pricing Study, which addressed uncertainties around customer price responsiveness, and the second was the AMI pilot, which addressed uncertainties around AMI costs and operational benefits.


Figure 1 – Demand trends within EnergyAustralia’s network area

Independently of EnergyAustralia’s AMI initiative, a national working group was formed in 2007 to establish minimum national functionality and to evaluate the business case for a national rollout commencing in 2009. Having the country’s largest, most comprehensive and most advanced pilot is allowing EnergyAustralia to contribute to this debate.

Responding to emerging consumption trends (Figure 1), EnergyAustralia commenced a five-year project in 2004 to rollout interval meters and time-of-use (TOU) pricing to all its 110,000 residential and small business customers using 15-160 MWh per annum. The company also made TOU the default tariff for all new and upgraded sites. Together, these two initiatives were increasing the number of customers on TOU by around 50,000 per annum.

TOU pricing was justified to the regulator on the basis of expected long-term savings from avoided and deferred network capital expenditure (capex), which is driven by growth in customers’ peak demand. Analysis had shown that rolling out interval meters to customers using less than 15 MWh per annum would be uneconomic based on network capex savings alone. By realising additional operating costs (opex) savings and other social benefits using two-way communications, an AMI solution could economically deliver TOU and other more efficient price signals to the remaining 1.5 million customers.

By demonstrating the potential benefits from a full-scale AMI deployment to society as a whole, and the limitations preventing a unilateral investment, EnergyAustralia aimed to secure the required regulatory approval to proceed from as early as FY09.


Figure 2 – Strategic approach to deploying AMI in 2004

If the regulator approved the investment, EnergyAustralia would earn its Weighted Average Cost of Capital on the estimated $284 million expenditure. As an added incentive, EnergyAustralia could ‘keep’ any operational benefits realised over the five-year regulatory period under the ‘CPI-X’ regulatory mechanism, after which time the regulator would return them to customers.

Importantly, the ‘ex-ante’ approach used by the regulator to set revenues tends to penalise companies for underestimating costs. For example, a 20% estimation error or cost over-run on a $284 million investment could lead to a $56 million loss. This greatly increased EnergyAustralia’s incentive to ensure the business case would be accurate, particularly on the cost side, and that the company would be able to deliver on it.

As shown in Figure 2, roughly half of estimated business case benefits come from opex savings, while the remaining half were expected to come from savings on capex. A Strategic Pricing Study was commenced in FY05 to reduce regulatory, customer, operational and financial uncertainties associated with the estimated capex savings. An AMI pilot was commenced in FY06 to identify and mitigate similar risk factors within the estimated capex requirements and opex savings.

The Strategic Pricing Study
EnergyAustralia’s Strategic Pricing Study (SPS) is the largest of its kind in Australia, involving 1,300 small business and residential customers. Its primary aim is to clarify the potential for modifying customers’ consumption patterns using highly targeted peak prices.


Figure 3 – Prices and information options offered

Approximately 10,000 customers were offered one of two new products, seasonal TOU or dynamic peak pricing (DPP, with both high $2/kWh and medium $1/kWh DPPs being tested), or an existing product with enhanced information (Info Only), as shown in Figure 3. The Study was planned to extend over two winter and summer seasons, until late 2008. Close attention was paid to the statistical validity of the Study and independent verification of the outcomes is being obtained.

Advanced Metering Infrastructure Pilot Programme
EnergyAustralia’s AMI pilot aimed to test currently available metering, communications and systems options, with a view to identifying the most suitable technologies for rural, suburban and urban applications (Figure 4). The pilot was expected to clarify AMI deployment and operating costs, as well as potential opex savings.

The pilot is the largest and most comprehensive in Australia, involving as many as 10,000 customers, four communications technologies and four vendors. Technologies being trialled are as follows:

  • Iskra DLC (Amis), GPRS and RS485
  • EDMI DLC (Echelon), GPRS, and RS485
  • PRI DLC (Echelon), GPRS and RS485
  • Ampy-Email DLC (Amis)

Figure 4 – AMI pilot technology scope

GPRS is used for typical back-haul from DLC concentrators or RS485 gateways. Each technology is subjected to an in-depth testing process prior to go-live.

The pilot is being deployed across several geographic areas within EnergyAustralia’s network territory to ensure that the experience and results are robust.

EnergyAustralia’s risk focused approach to deploying AMI has been instrumental in reducing the chances of under-estimating the costs or over-estimating the benefits. Experience gained from piloting the technology at scale with customers across the network has also given EnergyAustralia valuable insight into how best to move forward on the road towards a full scale implementation.

The results from the SPS to date have successfully demonstrated customer acceptance of new products and services, and customer price responsiveness in line with assumptions (Figure 5), and provided invaluable data on how best to market these new products to customers. The study has also helped EnergyAustralia assess the impact of in-house displays, detailed online data and enhanced billing information on customer energy efficiency, price responsiveness and satisfaction.

Useful insights were also gained when things didn’t go exactly as planned:

  • Using DPP to target weather driven wholesale energy or network peaks can be difficult
  • Sending dynamic messages to many customers in under two hours takes practice
  • Many customers may not use or even plug-in complimentary in-house displays, and
  • There is often very little room on the meter board for innovative new hardware.

Figure 5 – Preliminary summer residential response results

Lessons learned from the AMI pilot to date have included the planning, specification, procurement, testing, and installation of the technology. The relative operating costs and performance of the various vendor and technology options will be collected over the course of next year.

Preliminary insights yielded to date include:

  • Previous deployments may not be an accurate indicator of performance. Localising or enhancing ‘proven’ first generation technology may not be much easier than deploying ‘unproven’ next generation technology. Modifying a stable system can create instability, and significant vendor support may be needed to address technical problems. Establishing a transparent and close working relationship with the vendor is critical for the timely resolution of technical issues and development of desired enhancements
  • Commercial arrangements should be robust and allocate risk effectively. Businesses may need to adjust their contracts to reflect the reality of the new technology, which is still relatively immature. Commercial terms such as warranties, turnkey contracts, and bonuses should be considered to encourage on-time, defect-free delivery where possible. Full time vendor experts on site, visibility of the vendor supply chain, and a formal defect management process were all of considerable help to EnergyAustralia
  • Implementation is harder than it looks due to systems complexity. Each component (meter, communication and application) is relatively mature, but the integration of these components is not. The testing approach adopted during early phases is critical for mitigating technology and vendor risks that would otherwise arise during deployment. Spending money on a good lab setup for stress and scale testing can reduce significant rework in the field, where it is much costlier.

Figure 6 – Illustration of potential AMI pilot benefits

Figure 6 illustrates the potential value realised to date from the AMI pilot in terms of reducing capex estimation errors (approximately $80 million of this being due to a change in AMI deployment scope to include larger customers) in the AMI business case and identifying options for $72 million in potential cost reductions to target in Phase II.
Among these are:

  • Evaluation of other communications methodologies such as mesh radio and WIMAX
  • Investigation of plug-in meters to reduce the installation cost and future technology risk
  • Meter rationalisation to avoid the need for multiple meters and integrated load control (currently a typical installation for off-peak electric hot water involves two electricity metering elements)
  • Identification of alternative back-haul options to reduce communications charges
  • Optimisation to minimise data haulage and maximise ondemand service performance
  • Evaluation of application, communication and component standards (e.g. Cosem/DLMS – IEC62056, ANSI C12-22 for meter protocols, ZigBee for in-house display, TCP/IP, etc.)
  • Evaluation of auto-commissioning functionality for DLC systems to reduce maintenance costs
  • Evaluation of off-the-shelf AMI application capability, maturity and integration costs, and
  • Evaluation of market for turn-key and managed solutions to reduce costs.

Results from Phase II are expected to follow a similar course, leading to revised AMI business case estimates, and identification of additional strategic options for development. By adopting an iterative development approach, EnergyAustralia expects to optimise its business case, technology solution and implementation approach in advance of a full-scale AMI deployment.