Harvest your machine data before it’s too late
The key to operational efficiency
Strong competition and tight margins forces contractors to rethink their operating budgets and start looking for new ways to cut costs and optimize on operational processes - become more competitive. Most contractors are concerned about how to reduce operational costs and a big chunk of these are accounted for by wasted fuel and unnecessary service expenses, due to lack of insights.
Because one hour of construction equipment activity most often consists of less than one hour of productive time, it is critical for contractors to boost their operational efficiency. Traditionally, operational efficiency (ŋ) is defined as the number of minutes per hour that the equipment is used to do productive work, often expressed as a percentage.
Nonidle time refers to the number of productive hours where the equipment is doing a construction job. In contrast, idle time refers to the amount of time that the equipment’s engine is running, but the equipment is not doing a construction job. Idle time therefore refers to non-productive equipment time, and nonidle time refers to productive equipment time. Naturally, contractors would want all hours to be productive, but often the reality is far from this utopia. All non-productive hours translate into a lot of wasted fuel and increased emissions. In addition, for contractors with full service contracts, non-productive hours imply unnecessary routine servicing that will quickly add up as critical hidden costs.
Besides these metrics, excessive idling also accelerates wear of Tier 4 technologies, burn through warranty hours and plummet residual values. Decreasing idle time should therefore be a key focus point for contractors and all actions should be taken to reduce it.
A scary but simple equation
How much can idle time influence your operating budget, you may ask? I have made some calculations showing you how small reductions in idle times make an enormous impact on your budgets. The calculations are made with a representative machine that logs 2,000 hours per year for five years. In addition, the average diesel pump price in Europe of €1.17 per litre and a price of €5 per hour for a full-service contract have been assumed. Moreover, logged engine data from a medium-sized excavator with a Tier 4 Interim / Stage 3B engine has shown an average fuel burn of 3.79 litres per hour of idling. In the table, you can see how an idle time of 10% has a cumulative cost on a five-year period of €3,790 for a single machine. If that machine has a full-service contract the impact rises to €8,790 since it now accounts for unnecessary routine servicing.
Idle time vary between various construction equipment, construction jobs, and operator behaviours, but it is estimated that most heavy construction equipment has an average idle time of 30-40%. Imagine a contractor with 100 machines. It is fair to assume that the average machine will not idle excessively, but there will be groups of machines that do. Let us assume that one out of ten machines idle excessively with an average idle time of 40%. Reducing idle time from 40% to 30% on these 10 machines will cause direct fuel savings of €37,900 on a five-year accumulated period. That’s €7,580 in savings per year or just above €20 per day. If a contractor manages to cut idle time by just 2-3% on the overall fleet, this would translate into huge cost savings.
Machine-learning models to the rescue
It’s common knowledge that actions like restricting morning warm ups, turning off equipment at lunch time and using automatic shutdowns decrease idle times. However, very few have an overview of which machines or operators are idling excessively, which makes it difficult to accurately gauge the effectiveness of such anti-idling policies. In addition, contractors rarely know the magnitude of idling and the direct benefits of reducing idle times might therefore not always be clear.
Machine data enables contractors to identify groups of machines or operators that are idling more (or less) than the average. This can be done by implementing unsupervised machine learning algorithms that can be embedded in simple dashboards notifying the relevant stakeholders whenever machines are performing below (or above) fuel efficiency targets. In addition, machine data can be used to explain why some machines are idling excessively and helping contractors taking corrective actions by suggesting highly probable root causes. This branch of statistics is also referred to as descriptive modelling, where the primary purpose of the model is not to estimate a value (as in predictive modelling) but instead to gain insight on underlying and unknown patterns in the data.
The good news is that from a technical point of view, implementing such models does not require a particularly high data maturity - such models only need a handful of CAN Bus parameters and a little data history. The bad news is that to get this capability right, it requires creating a data-driven organization with the structure, culture, and problem-solving mindset to reveal the actionable insights that contractors need to drive growth. The contractors that start to think about machine data as strategic assets will become industry leaders and get ahead of the curve.