FLEET MAINTENANCE COSTS
For many large companies overall annual maintenance spend runs into several hundred million dollars or more. For few large clients we have seen these costs exceed a billion dollars a year. The size and age of the fleet, complexity and cost of the assets, repair parts, labor and regulatory and safety requirements contribute to maintenance costs. It is wise to look at the per-asset annual maintenance costs for comparative purposes. Complex assets such as Aircrafts and Locomotives are very expensive to maintain compared to a less complex asset such as railcars and trucks. Application of Analytics can help reduce these costs significantly.

Maintenance Cost Per Asset ($/year)
Railcars$1 – $3 K
Trucks$10 – $15K
Locomotives$100 – $200 K
Aircrafts including Engines$3 – $ 6 M
Source: Internal Kavi Associates estimates, composite figures (2012)

Table 1. Typical Annual Maintenance Spend per Transportation Asset

 

MAINTENANCE COST SAVINGS OPPORTUNITIES
For the discussion we have categorized fleet maintenance into three areas; Scheduled Maintenance, Unscheduled Maintenance and Predictive Maintenance. Primary maintenance cost drivers are part costs, inventory costs, labor costs, warranty costs, facility costs and supply chain costs associated with scheduling and routing assets for repairs. In the table below we have summarized the cost drivers with the applicable analytical approaches that can be leveraged to deliver business value. Each of these cost drivers are described in details in the following sections with the real life examples.

 

Cost Savings Opportunities Characteristics Analytical ApproachesBusiness Value
Scheduled
Maintenance
Preventive
Maintenance
Fix it before it breaks
Proactive in nature
Mandatory checks
and Overhauls
Determine optimal time
intervals between scheduled
maintenance events
Determine optimal repair
scope for each scheduled
maintenance event
 Maximize component
life
High reliability
Reduced overall repair
spend
Avoid over-maintenance
Unscheduled
Maintenance
Field failures
Reactive in nature
Fix only if it is broken
Root-cause analysis
Repair versus replace
analysis
Quality of fix comparisons
Recurrent event models
Reduced field failures
and repair spend
Minimize out-of-service
time and improve asset
utilization
Higher customer
satisfaction
Predictive
Maintenance
Condition monitoring
Predictive in nature
Fix if the evidence
points to an
impending failure
Real-time big data analysis
Failure signature detection
Rule-based AI Systems
Association and Sequence
analysis
Prevents costly failures
Extends the life of
components if they are
in good condition
Reduced repair spend
Higher customer
satisfaction
Parts Inventory
Management
Higher inventory
levels mean better
service but higher
costs
Pooling across
operators
Hierarchical part forecast by
location
Multi-echelon inventory
optimization across location
with service level constraints
Free up unproductive
capital
Reduce supply chain
costs
Maintain optimal service
levels
Warranty AnalyticsManufacturer to
determine terms and
conditions to
minimize exposure
Owners and
operators to identify
opportunity to replace
parts under warranty
Statistical techniques and
data mining for early warning
and detecting emerging
issues
Minimize exposure for
manufactures
Reduce warranty
leakage for operators
and owners
Labor PlanningIdentify labor needs
Consider sourcing
options
Allocate available
resources optimally
Hierarchical forecast of labor
needs
Optimization considering
overtime and outsourcing
Higher degree of labor
utilization
Improved repair quality
Ability to identify training
needs
Ability to generate
resource plans
Repair Capacity
Planning
Plan for the location,
capacity and
capabilities based on
the planned repair
needs
Statistical analysis of
historical repair data
Minimize supply chain
costs
Consolidate repair
volumes and negotiate
favorable pricing
Repair Scheduling
and Routing
Determine the
potential repair scope
Match the scope with
shop capability
reducing the need for
shop transfers
Check the availability
of capacity
Optimize supply chain
costs
Operational analytics to
forecast predictive repair
scope
Optimization to minimize
supply chain costs and delays
Minimize supply chain
costs
Reduce cycle times
Reduce out of service
delays and associated
costs
Reduce costs due to
shop transfers
Take advantage of the
lower labor rates outside
US for aircraft
maintenance
Reliability AnalyticsStudy component life
in terms of time to
failure or with respect
to other usage
parameters such as
miles, take-offs, hours
operated, etc.
Weibull analysis or fitting
statistical distributions for
non-repairable systems
Parametric and nonparametric approaches for
repairable systems
– Wayne Nelson’s Mean
Cumulative Function
– Homogeneous Poisson
Process (HPP) and NonHomogeneous Poisson
Process (NHPP)
Extend component life
by determining optimal
replacement strategies
Ability to improve design
of the components
Ability to source from
manufactures with
higher reliability scores

Table 2.  Maintenance Cost Savings Opportunities

We hope to provide relevant insights and a fresh perspective on how analytics can make a profound impact on asset-intensive organizations.  So stay tuned for our next post.