Digital Transformation in Logistics. Part 1. How to Monitor Fuel Consumption on 200 Cars for a Penny
At large enterprises, where a large volume of transportation is carried out using automobile transport, a significant part of the costs is fuel. Considering that the number of cars simultaneously on the line can reach several hundred, managing this process becomes very difficult.
If the vehicles are completely different (small dump trucks, BelAZs, special equipment, buses, passenger cars, flatbed trucks) and each vehicle performs completely different work from shift to shift with different drivers… you'll be scratching your head wondering how to systematize this process.
There are actually two ways: to accept the complexity of the task and let the process take its course, or to come up with a way to collect and structure the data and what conclusions to draw from it.
So, the task. Given: each shift, 100 to 200 vehicles go out on the line; shift assignments are different (from one-time movement of inventory items between workshops to cyclic work on bulk cargo); vehicle types are different; drivers change from shift to shift; conditions are inconsistent. What is needed: to find logic in a huge data flow, determine deviations, identify causes, and inform those responsible. Everything, of course, is automatic. Budget – 0 rubles.
There is slightly more than 1 million data per day. 72 thousand lines for 15 parameters.
Impossible? Then don't read any further.
What was already there: each vehicle had a GPS system that tracked the movement, and each tank had a fuel level sensor (FLS). Both values were transmitted constantly.
But the volume of this data is so huge that analyzing even one vehicle per shift would take about an hour. How can we analyze all of them? We need about 30 people full-time.
Now a demonstration of the capabilities of a Lean manager.
We grouped all the vehicles by type of work: transportation of bulk materials, transportation of individual inventory items, transportation of passengers, special work (for special equipment: loaders, rubble breakers, excavators).
Each group was divided into subgroups with reference to workshops. This allowed us to describe the patterns of equipment behavior. And as soon as the data did not fit into the patterns, a point analysis was carried out.
They quickly identified the place and time of the night's afternoon nap. By the way, lunch was also reduced from 3 hours to 30 minutes. They caught idle and semi-idle runs. And how many places they found where drivers hid from work – that's a separate story. And only at these events, the need for vehicles (and drivers) was reduced by 40% in several areas.
Now this logic had to be sewn into 1C UAT, which would look for deviations itself. The potential is huge, but since the task is to do it without expenses, the solutions were chosen simpler.
We identified three scenarios that were triggered by deviations:
1. If all drivers of a certain vehicle burn fuel the same way and constantly, then most likely the reason is in the technical condition. We send the vehicle for diagnostics of the engine and fuel system.
2. If only one driver of a certain vehicle burns too much fuel, the reason is most likely in the driving style of the specific driver. We conduct a point analysis.
3. If all drivers of a certain vehicle burn fuel, but one driver burns much more, then this may be related to both the first and second scenarios. First, we discuss this with the driver, then we send the vehicle for diagnostics.
Driving style should be understood as both incorrect operating modes (driving at low revs, idling, abrupt acceleration) and the possibility of theft.
All the logic was embedded into dashboards and displayed in the control room.
What we discovered along the way:
1. Potential risks associated with the delivery of fuel in tank trucks by a contractor. We worked out the risks of theft.
2. Blind spots for managers in terms of: Vehicle running/not running (especially at night).
3. Unjustified routes through the territory of the enterprise (even to the point of individual cases when they drove to the other end of the enterprise in a BelAZ because “the canteen there is tastier”).
4. They gave drivers a tool to justify excess fuel consumption. And this is an invoice for revising the methods of standardization. Previously, they could hit the bonus and it was difficult to prove anything to the manager.
There are now a huge number of offers on the Internet for installing similar systems. For one vehicle, several tens of thousands of rubles. It is important to find a fuel level sensor with the lowest error. We worked with an error of 1%, and even this was inconvenient.
In Russia, a program for the transition to electronic waybills is currently being implemented. And such systems allow data to be transferred automatically without the need to manually fill out endless waybills every day (I hint at increased labor productivity).
