Optimize your fleet by looking at a single metric: Operating Hours.
One of the main concerns of mixed fleet owners is the low amount of telemetry data received via the standard AEMP 2.0 (ISO 15143-3) interface.
This article shows how to be more efficient just by tracking a single metric, operating hours, available via AEMP. Following analysis of this metric will allow right-size your fleet's capacity to your business needs.
Fleet usage patterns
The operating hours metric is the most simple and widely available metric for machinery in the construction industry. This metric is relevant for machine owners as it drives the aging of the fleet, the need for maintenance, and a proxy for representing work progress.
The following data is based on the operating hours of real machines operating worldwide in the construction industry. The figure shows the monthly average working hours of machines and the average number of days in a month the engine has been operating. Each dot represents a machine.
Figure 1: Fleet utilization patterns
Machines in the first column of the plot are seldom utilized. Those machines are active only one day each month and for very few hours on average. On the other side, if we look at columns on the right, we see machines that are used every day of the month. And as we move our eyes upward, we see machines that, apart from being used daily, are used for long hours each day. The machine at the top right of the plot is used every day for an average of 22 hours.
Completing the view on your fleet
Before going deeper into the analysis, let's point out a drawback to this representation: more than one machine can overlap simultaneously, and we would only see one point in the plot. So it is hard to know the usage pattern of all the fleet at a glance. We can complement the previous plot with a heat map showing how many machines fall in each dot.
Figure 2: Fleet utilization heat map
By looking at the heat map, we can see that most machines work between 20 and 26 days per month, between 3 and 9 hours a day.
Diving into your fleet data
Let's go back to the previous plot and do some further analysis. Now we can cluster machines by their usage patterns. We reflect these clusters in the colors of the dots:
- Green machines are low utilization machines
- Orange machines are high-utilization machines
- Purple machines are outliers
- machines without a clear usage pattern that do not fit in the two main groups.
Eventually, we can draw some boundaries in the plot that let us look into more fine-grained groups of interest. We use color bands to define these boundaries.
Let's start by breaking the plot into two broad areas with the horizontal pink boundary. In the bottom area, we have machines that work less than 9 hours a day. The top area holds machines that operate more than 9 hours a day and are assumed to be working as part of a shift operation and exposed to considerably more use.
Now let's divide the plot into vertical sections with additional color bands. These bands separate machines into groups of machines that work
- up to 2 days a week (orange),
- up to 5 days a week (green),
- up to 6 days a week (red) and
- up to 7 days a week (blue).
Looking into a specific machine model
Until now, we have been looking at the whole fleet. Let's start by looking at machines of a single-mode: Model Type 1.
Figure 3: Model Type 1 utilization patterns
The plot for Model Type 1 is using the same representation we have explained above. Instead of showing all fleet machines, this plot analyzes a subset of the fleet, only machines of Model Type 1.
We have also calculated a few aggregate values for the two machines: green (low utilization machines) and orange (high utilization machines). We see that Model Type 1 machines from the green cluster are used close to 3 hours a day, 6 days a month. And they age at a rate of 202.94 hours a year.
Model Type 1 machines in the orange cluster are used nearly 6 hours a day, 22 days a month. And they age at a rate of 1566.52 hours a year.
It seems like machines in the green cluster are underutilized compared to those in the orange group.
Analyzing excess capacity
Let's see what is the excess capacity of the green cluster compared to the orange cluster. Machines of Model Type 1 in the green cluster have an excess capacity of:
Given we have 48 machines in the green cluster, the total excess capacity is of:
If we wanted to express this excess capacity in number of machines, we could divide the excess capacity by the average aging of the machines in the orange cluster:
So we have 41 machines in excess.
Optimizing the fleet size
Just by looking at operating hours, we can start making decisions about the sizing of the fleet. Of course, other questions are relevant to make the final decision to resize the population of a particular model of machines:
- Is this excess capacity a result of worksites distribution and need for redundant machines, lack of operators, and demand?
- Do we want to maintain this extra capacity to deal with peaks in demand?
- Should we reduce the number of machines in this model?
- Should we buy a lower-end model or rent to cover the low utilization machines?
Analysis of machine operating hours is just a starting point to drive good decisions about your fleet capacity and sizing. It is not going to give you direct responses. Still, it is the basis for building a decision framework that your business expertise would assist.
Get the best insights from your fleet data.
As this article shows, you can start getting value out of your fleet data now. You don't have to wait to get all your desired metrics, as the most simple ones already deliver high-value outcomes when looked at with care. You require the right expertise.
We are experts in telematics, data analysis, and mixed fleet management. If you need help getting the right insights on your fleet data, please do not hesitate to contact us.
