Fuel consumption evaluation CA3500

A number of fuel saving initiatives were implemented in later years on Dynapac machines, however no controlled, practical verification on an actual job-site was ever made…


A number of fuel saving initiatives were implemented in later years on Dynapac machines, however no controlled, practical verification on an actual job-site wasevermade. Eco mode, which controls the engine RPM to only deliver as much power as is needed at any given point in time,saves fuel and reduces both noise, emissions and wear on the machine. In addition, Dynapac now also offersthe Seismic system on selected CA rollers, this system further reduces the energy consumptionby allowing the machine to operate at a lower vibration frequency than a traditional vibrator roller.The machine measures and evaluates thenatural frequency for the current material, state of compaction and the machine. This is done automatically at a rate offive times per second. In turn, italso reduces the emissions, fuel consumption and noise in andaround the machine. The Seismic system also improves the compaction performance on certain materials; this means that the number of passes can be reduced compared to a traditional roller. This will further reduce the running cost of the machine. However, this is not taken into account for this specific test.There are a number of compaction performance tests available that verifies this claim. The Seismic system was evaluated in a doctoral thesis conducted at the Royal Institute of Technology (KTH) in Stockholm by Tekn.Dr. Carl Wersäll. Further studies on the benefits of Seismic (or resonance compaction) have been carried outunder the umbrella of the Swedish Transport Administration (Trafikverket).Please see Appendix B for a summary of the articles etc. published on the subject.The testscovered in this reportweresupervised by researchers from VTI (Statens väg-och transportforskningsinstitut, The Swedish National Road and Transport Research Institute). The supervision included in-person monitoring of running hours, operating cycle and fuel replenishment. A separate, independent report has been published byVTI.

Test set-up

The testing was performed on a land reclamation area consisting of blasted granite in various fractions, the material was hauled and dumped by articulated dump trucksand spread by a dozer before compaction. The job site is part of the port expansion of Karlskrona Baltic Port on the island of Verkö in Karlskrona.

Overview of the test area with Dynalyzer data plotted
SEISMIC roller, dozer and ADT at the test site

The roller used was a CA 3500 D, Seismic S/N 10000188TMA031719. The machine was delivered from Dynapac in August 2021. Running hours 333 at the start of the test and 374 at the end. The machine was rented from a Swedish rental company. A small change to the machine software was implemented in order to evaluate the accuracy of the fuel consumption data sent via the Dynapac telematics solution, Dynalink. After the tests were concluded, the onboard software was rolled back to the original release.

Fuel was supplied from a portable filling station on-site; access to the fuel filling point on the machine was not possible during operation. The engine hood was locked with a pad lock during operation and the key was in possession of the VTI test supervisor until the end of every day.

A CAN logger recorded fuel consumption data from the engine management system. In addition to this, fuel consumption data was also logged in Dynalink. Actual volume of fuel filled at the end of each day was also measured and logged. A Dynalyzer system recorded, among other data, compaction meter values, number of passes and the compaction area per day.

As no standardized drive cycle exists for evaluation of fuel consumption of soil compactors, we decided to create our own based on Dynalink data collected from a number of comparable machines. Data from machines on several markets was used to calculate a representative operating cycle.

The result is a one hour long cycle with the following distribution:

  • 9 minutes static operation/transport, 4-5 km/h
  • 14 minutes vibration, low amplitude, 3,5 km/h
  • 7 minutes vibration, high amplitude, 3,5 km/h
  • 9 minutes static operation/transport, 4-5 km/h
  • 14 minutes vibration, low amplitude, 3,5 km/h
  • 7 minutes vibration, high amplitude, 3,5 km/h

This one-hour cycle was repeated six times per day for a total of 6 running hours. Every test day started with a ten minute warm up in idle each morning. Two 15 minutes breaks were also part of the test schedule, one in the morning and one in the afternoon. The machine was left idling during these breaks. A half hour lunch break, during which the machine was turned off, was also part of the schedule. This results in a total of 6 hours and 40 minutes of daily engine run-time (including warm-up and breaks)

The testing was performed during two weeks in February 2022, Feb 8-10 and Feb 15-17. The second week was a repetition of the first test week.

Test schedule:

Feb 8 – Full engine RPM, maximum vibration frequency
Feb 9 – Eco-mode, maximum vibration frequency
Feb 10 – Eco-Mode, Seismic
Feb 15 – Full engine RPM, maximum vibration frequency
Feb 16 – Eco-mode, maximum vibration frequency
Feb 17 – Eco-Mode, Seismic

Timing was determined by a timed presentation with a countdown clock indicating to the operator what the current running mode should be. One minute before changing mode of operation the presentation tells the operator about the machine setting to be changed to.
One driver ran all three days the first week. During the second week, we utilized several drives and changed drivers during the lunch break.

Diagram einfügen

In general, the different data points in the same direction when it comes to the relative fuel savings between the different operation modes. There is a discrepancy on Feb 15, however as commented below we were forced to move the refueling position of the roller which affected the amount of fuel that the tank could accept. As the physical fuel amounts for Feb 16 and 17 are similar to the corresponding days in the first week (Feb 9 and 10), we attribute this discrepancy to the unfortunate change of refueling position. The consumption data collected by the Dynalink system matches the logged CAN data from the engine control system (total_fuel_used) in a very good way. This fact further supports the theory that the actual amount of fuel filled at the end of Feb 15 was incorrect and should have been somewhat higher if the original refueling position had been maintained. This assumption is supported by the fact that the actual fuel filled matches the logged Can and telematics data for Feb 16 and 17 (where the second refueling position was still used)
The realized savings when using Eco mode and full vibration frequency, compared to full engine RPM and full vibration frequency, is in the range of 21-22%, in this specific case resulting in 18-20 liters of fuel saving per day. Going from full engine RPM and full vibration frequency to Eco Mode and Seismic results in 36-38% savings. In this case corresponding to 30-32 liters of fuel saving per workday. As one liter of diesel produces 2,64 kg/CO2 when used the fuel saved will also significantly reduce the CO2 emissions. In the best case; CO2 emissions can be reduced by 79-84 kg per day compared to a traditional roller running at full engine speed and with full vibration frequency.


Between the first and the second test week we were forced to change the refueling position. This potentially affected the amount of fuel filled on Tuesday Feb 15 and potentially Feb 16. These values deviates from the measured amount from the engine management system. In all other aspects the readings correspond well. The machine had to be taken off site on Thursday Feb 10 for minor repair. This additional driving meant that fuel was consumed between week 6 and 7. Hence, the machine had to be refueled before starting compaction on Feb 15. The amount of fuel filled (6 liters) was greater than the amount used from Feb 10-14 (3 liters) and this indicates that the tank contained somewhat more fuel on the morning of Feb 15 than after refueling in the evening of Feb 10. The reason for this can be attributed to the design of the tank where the filler neck is also the breather tube, placed on the right side of the tank and the fact that the longitudinal and transversal angle was different in the two refueling positions. This will have an effect on the amount of fuel that can go in the tank before the fuel filler pump shuts off. Total fuel used as measured by the engine ECU on Feb 15 was 81 liters. By adding the additional three liters that was the result for the new refueling position to the 77,3 l refueled at the end of Feb 15 we would get 80,3. This is very close to the 81 liters measured by the engine management system

Fuel tank with filler neck
Refueling position w6
Refueling position w7

It is worth noting that during high amplitude operation at full frequency (29 Hz) the ABC (Active Bouncing Control) very frequently turned vibration off due to the high stiffness in the compacted material. Analysis shows that during one working day, Feb 16, the ABC switched off vibrations some 150 times during high amplitude compaction in Non-seismic mode. When operating in Seismic mode there was not one single ABC intervention in high amplitude. This means that with SEISMIC, the material absorbs all the compaction energy. Compaction is improved and we reach higher stiffness compared to the traditional, full frequency setting.

Static compaction

Low amplitude

High amplitude

The above graph shows the vibration frequency for about a half hour operation on Wednesday Feb 16, non-seismic, fixed vibration frequency. 9 minutes, static operation, 14 minutes of low amplitude and 7 minutes of high amplitude vibration, the high amplitude frequently interrupted as the ABC shut the vibration off due to high levels of double jumping.

Static compaction

Low amplitude

High amplitude

This graph shows the vibration frequency for a half-hour operation on Thursday Feb 17 in Seismic mode. On the same material and same conditions as Feb 16 but in Seismic mode there is not one single interruption in high amplitude by the ABC. In low amplitude, the frequency adjustments made by the Seismic system can be seen as a slight waviness in the graph as the frequency was adjusted.

A couple of driver errors has some effect on the recorded fuel consumption. Both of these are easily identified in the Dynalink fuel consumption data. In the afternoon of Feb 15, one 9-minute session that was planned for static operation was driven with low amplitude. Average fuel consumption for this period was around 13 l/h (instead of around 10l/h that was measured for the other static operation on full engine RPM). In the afternoon of Feb 16, the engine was shut off for 15 minutes during the afternoon break. It should have been left running in idle (fuel consumption in idle is around 2,3 l/h so the effect on the total fuel used is not very big).

In addition to this, the driver’s comfort in terms of vibration and noise levels in the cab is significantly improved when the machine is operated in Eco and Seismic mode. The lower noise and vibration levels
are confirmed by laboratory measurements. Eco and Seismic modes also reduce the noise from the machine to its surroundings.

The fuel consumption data in Dynalink is presented in deciliters per hour; however, the legend says l/h. The change to deciliters per hour was made to improve the resolution of the data. Please see Appendix B for a presentation of the fuel consumption graphs recorded in Dynalink during the tests.

Conclusion / summary

As expected, the fuel consumption decreases with ECO mode in operation and it is reduced even further with Seismic activated. In this specific scenario, the savings significantly exceed the theoretical calculations. Actual, absolute savings will vary between jobsites depending on how heavy the work is and how much the roller is utilized. In this case we ran the machine for six hours per eight hour working day giving a utilization of 75%. However, the relative improvements between the different operation modes should be consistent. Regardless of the utilization, the relative saving (in percent) from full engine RPM and full vibration frequency to ECO and Seismic is expected to be quite stable.

Fredrik Akesson – Dynapac Sweden

Share this content: