The challenge
The EMTEC Corporation is a UK based start-up founded in July 2017, which aims to use data analytics to reduce transport-based emissions and pollution. They have developed a cloud-based platform that receives data from connected vehicles via smart devices, which can then be used to provide real-time analysis of vehicle performance and driver behaviour.
EMTEC asked NPL to provide an independent assessment of their carbon emissions algorithm for road vehicles, both petrol and diesel. The EMTEC model uses datasets from vehicle on-board diagnostics (OBD), and an analytical model to reflect different regimes of driving. The results of the model can be compared with direct measurements of carbon emissions from the Millbrook proving ground, a dedicated vehicle test environment.
The solution
NPL carried out a critical assessment of the principles and assumptions used to derive the EMTEC model and reviewed the parameters used within the calculations. The NPL approach used a ‘ground truth’ test, where data from the EMTEC algorithm was compared with actual field test measurements from the Millbrook proving ground, which would in turn provide useful information and confidence in the algorithm.
A comparison was carried out between measured data via on-board diagnostics and predicted values via the EMTEC model. NPL assessed the validity of the EMTEC model based on data analysis and error assessment, as well as the physical principles of the engine flow rate, and compared modelled and test datasets (petrol and diesel tests in the Millbrook facilities) from the on-board diagnostics.
NPL was able to assess the similarity between direct measurements of carbon emissions performed at Millbrook with the EMTEC quantification of emissions derived from the OBD mass flow rate.
NPL have assessed the EMTEC petrol CO2 algorithm and believe it to be a justifiable approach to the calculation of CO2 emissions from vehicle onboard data. The current values of selected parameters and the specific type of engines and fuels give useful information. It was found that the car should be driven for a sufficient amount of time (at least 30 minutes) to produce driving data with an error in estimation of total carbon emissions of below 5%. Specific testing profiles and driving regimes were considered, and the cumulative error of the EMTEC petrol CO2 algorithm was 0.82% after a 2.5-hour test, which is much less than the 5% instrumental uncertainty stated in the specification of the sensors.
The impact
The collaboration led to a better understanding of the testing sensors and equipment, and EMTEC received independent assessment of their technology. EMTEC have used the information from NPL to endorse their methods, as shown on their website. EMTEC have increased confidence that their algorithm and associated parameters are suitable for their overall goal. The subsequent comparison between test results and calculated values demonstrated that the differences between the model predictions and reality were acceptably low. This means that they can offer a practical solution to assess real-world fleet vehicle CO2 emissions with real time and cumulative reporting and within acceptable uncertainty limits. This will enable them to support companies with over 500 employees which are mandated to report vehicle emissions.