Electricity distribution wildfire risk management with data quality
Andrew Gatland outlines a structured approach to managing wildfire risk in electricity distribution networks through data quality monitoring — linking asset data directly to risk model requirements to provide near real-time assurance that risk models can be relied upon.
The growing wildfire risk
The risk of wildfire is increasing in New Zealand due to climate change, with a marked increase in the frequency of significant wildfires in recent years. Wildfires can cause serious property and infrastructure damage and lead to loss of life.
According to Fire and Emergency New Zealand, approximately 4,500 wildfires occur in New Zealand annually, 98% of which are caused by people. Three percent develop into major incidents. Stats NZ found that very high and extreme fire danger days increased at 12 sites across New Zealand between 1997 and 2019.
Electricity distribution businesses face a unique challenge of managing extensive network assets across thousands of square kilometres in a variety of climatic and geographical conditions. Their overhead networks are often in close proximity to vegetation and communities, creating the threat of ignition events in the right conditions. As the climate continues to warm, the risk of wildfires will increase, requiring electricity distributors to develop detailed mitigation strategies and plans.
Strategies for electricity distributors to manage wildfire risk
The geographic extent and thousands of individual assets comprising electricity distribution networks means that daily or even weekly surveillance of equipment to identify and assess wildfire risks is not practicable. Quantitative models are therefore required to enable the identification and assessment of the most critical risk areas.
Inputs to such models include data and information about the assets — their location, type, installation configuration, age, and condition. The quality of the data must meet acceptability thresholds to ensure the outputs of the models can be relied upon.
The use of quantitative models to support risk management and decision-making is a well-established practice in New Zealand electricity distributors. For example, the EEA developed the Asset Health Indicator Guide, which provides a common way for the industry to approach the preparation of asset health indicators — intended as a strategic tool for asset management governance discussions.
A range of methodologies for forecasting wildfire risk and modelling risk factors for distribution networks have been developed in Australia, many of which could be modified for use in New Zealand. Some New Zealand electricity distributors have independently developed models that they are using to manage wildfire risk.
Often the biggest obstacle to producing valuable insights from models is not the modelling approach itself, but the quality of the data held by the organisation. Reference models can be developed and continually improved for application across an entire sector, however the ability to apply a model within a specific organisation depends heavily on the format and quality of the data that organisation holds.
Information governance ensures that organisations can realise full value from their data and information assets — in this context, the ability to identify and assess the risk of electricity distribution assets causing or being impacted by wildfires.
Managing wildfire risk with data quality monitoring
For wildfire risk models to be effective, data inputs must be of acceptable quality. A structured and systematic approach is needed to provide assurance that data is fit for purpose and, where issues exist, to identify corrective actions.
The key principle for effective data quality management is that data must be precisely linked to the information requirements of the business. For managing wildfire risk this means that data and information inputs for the selected modelling approach must be identified.
The diagram below provides a way of representing the information requirements for managing wildfire risk within the context of an electricity distributor's asset management system. A decomposition of the organisation's business capabilities is identified, from top-level Subject Areas into Capabilities.
Within the "Maintain" capability a Sub-capability "Manage Wildfire Risk" is established. This sub-capability includes the business processes, systems, competencies, and data and information the organisation requires to implement its wildfire risk management strategy. Four information requirements for managing wildfire risk are identified.
Figure 1 - Establishing wildfire risk management as a business capability
For each information requirement a range of data quality requirements have been identified, representing the specific data inputs to risk models. For each data quality requirement the specific dataset within the organisation's information systems is located — creating the precise linkage required between the data and information needs of the organisation and the data that is held.
Figure 2 - linking the wildfire risk management business capability to required data inputs
Once these linkages are created, it is possible to evaluate the quality of the data to meet the information requirements. Acceptability thresholds for data quality must be established, considering the criticality of the data to the modelling approach and the sensitivity of the model to the data.
These thresholds can be translated into database queries that enable programmatic data quality checks at the desired cadence, providing information users with near real-time feedback on data quality. This provides the basis for identifying data quality deficiencies that could compromise the ability of wildfire risk models to properly identify and assess wildfire risks posed by network equipment.
Figure 3 - feedback on data quality
The diagram above shows that data quality issues exist with aged conductor data — meaning the organisation's understanding of risk in some conductor assets may not be sufficient to implement its wildfire risk management strategy.
Establishing Data Quality Monitoring
Data quality monitoring is a critical business capability for supporting information governance and asset management. As the frequency and severity of wildfires increases, electricity distributors need assurance that the data underpinning their risk models is fit for purpose.
DataFrame provides a proven and rapid means of establishing real-time data quality monitoring in the infrastructure sector, creating the linkages between data and business requirements that enable wildfire risk models to be applied with confidence.
Want to improve your organisation's data health?Asset Dynamics works with electricity distributors to establish data quality monitoring that underpins effective risk management and decision-making.
Learn about DataFrameGet in touch