Trip computer and long term Fuel Economy Calculation

All Kia Niro related discussions
Post Reply
jerrytaff
Posts: 98
Joined: Wed Oct 03, 2018 8:28 pm

Trip computer and long term Fuel Economy Calculation

Post by jerrytaff » Tue Jan 29, 2019 3:30 pm

On previous cars, where the trip computer has computed 2 trips, I have used one for current, and the other for the very long term. The long term calculation of mpg has been stable and reliable. On the PHEV the most recent trips are stored in the trip history, so I disabled the mpg reset on fill-up and have let it accumulate since I received the car. The mpg calculation from the accumulated mileage is now wildly inconsistent. :? see below

Before a trip yesterday, the computer displayed 25479 miles at 103.0 mpg.
Yesterday I did a trip of 117 miles at 70.7mpg.

Some quick calculations would indicate that, if the figures are true, prior to yesterday, the car would have used 24.74 Gallons. Yesterday's trip would have used 1.65 Gallons. My new mileage is 2664.9 miles and total petrol consumption 26.4 Gallons, so I would expect it to bring the computed overall fuel efficiency down to 101.0mpg.

However. although the car now displays the expected mileage, it computes it at an average of a mere 84.3mpg. For such a short journey (4.4% of miles driven) to have such a big effect the car would have had to average 17mpg yesterday. I prefer to believe the approximation of 70.7.

So, it is apparent that for the most recent journey to have had such a big effect on the calculated economy, it cannot be counting all the miles with equal weight. I suspect it is ignoring the older journeys altogether; Perhaps as they fall off the displayed history.

So, just a word of warning - even if the car can compute reasonably representative fuel economy figures short term, over a longer interval its figures are wildly inaccurate and cannot be believed; Especially on the PHEV when the variation between economy on short and long journeys can be orders of magnitude.
2018 PHEV 3 in Gravity Blue :D
Previous DS (formerly Citroen) DS5 Prestige BlueHDi S/S Auto 2.0, Jaguar X type 2.0 S Diesel


niroal
Posts: 177
Joined: Thu Jun 29, 2017 11:22 am
Location: North Kent

Re: Trip computer and long term Fuel Economy Calculation

Post by niroal » Mon Feb 11, 2019 1:52 pm

Agreed.

I let my full one to reset on fillups and it can be 5-10mpg out either way. I do a Fuelly comparison and have been running that for just under 17000 miles in 13 months & my overall average is 87.7.

Currently going for my 2nd at least 1000 miles on a tank on the trot which I will live with in this weather!!!!
Niro 3 PHEV Graphite


RMT
Posts: 2
Joined: Wed May 01, 2019 4:10 pm

Re: Trip computer and long term Fuel Economy Calculation

Post by RMT » Wed May 01, 2019 4:28 pm

I have done 19,000 miles in the Niro phev. Over each tank I average well over 100 mpg as I do quite a few short journeys. I am quite happy with the mpg. However the computer is unable to calculate correctly when mixing hybrid driving and then some pure EV driving. When you drive in pure EV the mpg increases far quicker than it should (about three times the rate). I have spent a year querying this with the dealer and Kia. I have sent many examples and the dealer has confirmed my figures. Kia UK are not interested, refused to give me contact details for the appropriate director (I found it anyway) and have closed my complaint and refuse to acknowledge that there is an issue , they refuse to communicate on the matter any more. The reverse of this issue also occurs, ie, the mpg falls too quickly after many miles on EV.

Sometimes my fill to fill calculation agrees closely with the car computer, this depends on the mix of driving, however when I have done a lot in pure EV towards the end of a tank full it is miles out. The last tank I did 1050 miles on a tank, the computer said I had done 157 mpg, (it increased quite rapidly towards the end) I had actually done 115. This error is replicated many times.

It is quite clear there is a fault in the calculation that is programed into the car. I guess the cumulative mileage is not properly taken into account when in EV.

The car is generally very good, especially after binning the awful tyres and putting Michelin Cross Climate XLs on, it now grips in poor weather and doesn't understeer. It's a shame that Kia Customer Service are so poor.


jerrytaff
Posts: 98
Joined: Wed Oct 03, 2018 8:28 pm

Re: Trip computer and long term Fuel Economy Calculation

Post by jerrytaff » Thu May 02, 2019 12:48 am

RMT, your findings back up my original post. As we have observed, if your last few trips are predominantly EV, the mpg goes through the roof, but if they are on the ICE, then it plummets. I've been getting the cumulative mpg to fluctuate between 90 and 230 over a few weeks without ever resetting the trip computer for the last 5000 miles. Both figures are impressive, but neither is anything like accurate, and such a variation is impossible over a few hundred miles out of 5000 total.

I don't think its a bug, just a very misleading algorithm, which uses a digital low pass filter of some kind. Either FIR (weighted average of last few trips / miles / days) or IIR (weight of older trips decays exponentially, but never quite gets to zero). The idea being that the engineer responsible does not believe that you are particularly interested in the economy you used to get, but are more interested in the recent trends. Yes , I know, you would just reset the trip computer if you weren't interested in the older stuff.

So, following those models, I haven't had the patience/time to collect the data that I'd need to model it, but if I had a sequence, starting with a reset, and the mileage (and date) of every journey from then, together with the indicated mpg for that journey, and the cumulative since reset, then I could probably deduce what the algorithm is that they have decided to implement.
2018 PHEV 3 in Gravity Blue :D
Previous DS (formerly Citroen) DS5 Prestige BlueHDi S/S Auto 2.0, Jaguar X type 2.0 S Diesel


RMT
Posts: 2
Joined: Wed May 01, 2019 4:10 pm

Re: Trip computer and long term Fuel Economy Calculation

Post by RMT » Thu May 02, 2019 7:49 am

Jerrytaff, I think you are probably right, and I often thought that the algorithm is not picking up the milage from a reset point, but some more recent point, but I would say that that was a fault as you are given the option of resetting manually, resetting on each trip or on refueling etc. I have mine set to show the current trip. and use option 2 to reset on refueling, so I can see what I have done on the current tank of fuel, and what I have done on the current trip. If that option (and others) are available why would the engineer program it to ignore your preference? I could perhaps understand that on a long term reading, someone may have thought it is more relevant to weight it to the later trips (but I would disagree). One observation that has made me doubt that is the intention is that even after a reset having done just a few miles of mixed motoring, it goes up at an incorrect rate per mile driven that is similar to what it does with a greater cumulative trip. I was just very annoyed that Kia UK will neither admit it is wrong, or explain why they do it that way.


Post Reply

Return to “Main Kia Niro chat”