When considering the life of any asset, the question arises: what types of interventions should be planned to keep it operating or to restore operation in the event of its failure?
Organizations that manage critical infrastructure rely on a wide range of asset types to provide their services. These asset populations can include millions of assets of varying complexity and cost. Modeling asset condition and the related probability of asset failure is non-trivial and costly. Therefore, assets that require modeling are those whose failure will cause significant expense or risk to an organization, or where the cost of replacing the asset is high.
The goal is for organizations to maximize the value of their capital and operational expenditures over time. To do this, the risk of an asset failing to operate as desired must be modeled. Risk is defined as the probability of failure of an asset at a given point in time multiplied by the consequence(s) of such a failure. Models are classified into two categories:
The benefit realized by performing an intervention on the asset at a given time is the predicted reduction in risk achieved by improving the asset condition, compared to not doing anything. The benefit is compared to the cost of the intervention to better understand the cost of action vs. inaction. The time at which an intervention provides the best value considers possible constraints such as limits on expenditure, performance KPIs, resources, etc.
A simple risk model is one where the reduction in risk of the asset is achieved by replacing it with an identical one. This is unrealistic because in all but the simplest cases, it makes more sense to repair the asset to restore operation. For instance, we don’t replace the family car when the starter motor fails, assuming that the rest of the car is in good condition; whereas if the garden hose leaks, it may well be justifiable to get a new one!
To better model diverse types of intervention such as maintenance, repair, and replacement, we must understand the effects of the interventions on an asset’s condition.
Some assets may be adequately modeled as a single unit if the intervention affects the overall condition of the asset. However, complex assets may require modeling of major subcomponents if a maintenance activity only improves the condition of a subcomponent.
For example, a car may be modeled as the overall vehicle, or composed of an engine, transmission, and bodywork. If we maintain the brakes, this will reduce the safety risk but will not increase the reliability of the engine. A model that treats the car as a single unit will sacrifice information; the safety risk may be understated at the expense of the reliability risk or vice versa.
For this reason, it’s important to treat asset subcomponents as separate assets with their own asset types. In the car example, the asset might be represented by an engine, transmission, and bodywork — each with its own condition, probability of failure, and cost models.
Different approaches can be used to express the relationship between age, condition, and probability of failure. These include using health and probability of failure curves as well as industry-proven models, such as the Common Network Asset Indices Methodology (CNAIM) that was developed in the UK by Distribution Network Operators and adopted by the regulator Ofgem.
One important consideration is the trade-off between the fidelity of a model and the cost of creating and maintaining the model. High fidelity requires complex modeling of components, detailed and accurate data, along with higher costs associated with model development, data acquisition, and maintenance for both the software and data. Simpler models are cheaper to develop and require less data to be collected and maintained.
Once the granularity of the models required to represent an asset type is established, the next issue to consider is the nature of interventions that can apply to the asset. These can include:
To determine the best approach, organizations should define multiple intervention types and apply different strategies to identify the optimal mix of interventions and their timing. For example, the same intervention can be repeated, enabling the planner to understand the impact of a single intervention type on the asset over time. Alternatively, different intervention types can be repeated on a fixed schedule.
Leading Asset Investment Planning solutions, like the Copperleaf® Decision Analytics Solution, enable organizations to:
Want to learn more about how to create optimal asset intervention strategies? Check out this on-demand webinar: Leveraging AI-Powered Optimization to Bundle Asset Interventions and Reduce Outages