Lead acid battery

[elecLear78]

I wanted to know your comments regarding the formula which i worked out after quite an effort. The idea is to calculate the SOC of a 12v lead acid battery during run time or vehicle moving condition. First thing is i made simplifications to the algorithm in order to reduce the number of calculations to implement in micro controller.
Inputs: Battery voltage , Battery current.
Output: SOC of battery (0-100%) on CAN.
Step1: As soon as the Ignition is ON the Micro reads the open circuit battery voltage (assuming all loads are OFF). Based on the open circuit voltage i measure the initial soc. (The relation is linear between soc and ocv).
Step2: Assuming the alternator is OFF
Then based on the relation that BatteryCapacity = Ampere * Time.
And an assumption that under loaded conditions if the battery voltage reaches a voltage of 10.5 then the soc is 0%. I derived the following
equation.
Suppose if the initial voltage is v1 and after applying load i1 the let the new voltage be v2. And Initial soc is soc1 then
socdec = (v1 - v2)*soc1/(v1-10.5);
soc2 = soc1 - socdec.
and the calculations will go every second with
soc1 = soc2;
v1=v2;
v2 newly read voltage.
Step3: If the alternator is ON the algorithm will change and it will be the current flowing into the battery. This is because i cannot read the battery voltage.
Now my main questions are regarding step2 how far the formula is valid?
There is case where if i remove the load then v2 is greater than v1 and the soc rises. This somehow is creating some kind of confusion since i do not want soc to rise.
How practical is this for actual implementation? My idea is i do not want a perfect soc meter but at least an approximate one is fine.
Any modifications to the algorithm i can do to improve or totally i have to discard. I am not able to take a decision.
Please suggest.

 

[MikeMl]

Do you understand that in normal automotive service, a lead acid battery is never (and never should be) discharged below about 90% of full charge! The battery is always almost fully charged, provided that the car is driven a few km since the last start.

Normally, the only load in a car that the alternator cannot keep up with is the starter motor. The charge removed during cranking is replaced by the time you drive few km. The charge removed from the battery in the event that you pull up to a stop light with the headlights on, the radio playing, and the AC running will pull the battery down a few % because the alternator might be maxed out, but that charge is replaced by the time you drive to the next stoplight.

What are you trying to do?

 

[Tony Stewart]

Starter batteries have short life when drained fully and recharged and even shorter if high current and elevated temp. But cheap 50Ah.

But 100~0% SoC for flooded lead acid is 12.5 to 11.5V with a light load for equalization. otherwise float V+ noload can be erroneous.

 

[elecLear78]

But 100~0% SoC for flooded lead acid is 12.5 to 11.5V with a light load for equalization. otherwise float V+ noload can be erroneous.

That is true but after some time i thought it will come to the correct SOC based on the loaded voltage.

 

[Tony Stewart]

That is true but after some time i thought it will come to the correct SOC based on the loaded voltage.

Actually the float voltage will equalize to yield the correct SoC vs Vbat.

AS long as the load does not create IR drop from battery ESR yet drain the float charge after the alternator is removed, it can be sensed within a minute using 1A or so to equalize charge voltage to float voltage.

The bulk storage of lead acid car battery is around 50kFarads and the equalization capacitance is a few percent of this in parallel with a much higher series resistance than the bulk cap circuit of milliohms. ( As far as I recall)

------||-----
--||--/\/\/---

 

[Mosaic]

SoC of lead acid batteries based on voltage is merely a guide for NEW condition batteries. It is also different for wet cell , and AGM and gel cell battery chemistry.
It can't even be considered a reliable reference if the battery has had some use and shed active plate material or developed any sulphation.

A lot of OTC chargers intepret the SoC from the battery Voltage and often undercharge batteries resulting in creeping conversion of active plate material into sulphated plates and loss of capacity.

I have processed almost 200 defunct lead acid auto batteries and data logged their behaviour including building a custom data logging Ah capacity load tester now on sale at Clubjameco as a DIY kit.
Here it is open sourced.
https://hackaday.io/project/1375-lead-acid-ah-capacity-testerlogger

Identical battery types (used) can have vastly different discharge curves and knee inflection points. Battery chemistries (Pb-Pb; Pb/Ca;Ca/Ca; Pb-Se/Ca; Pb Sn, Ca etc.) affect their loaded performance due to sulphate nucleation sites 'spawning' or 'enabling' faster discharge sulphation which shows up as a bump in the voltage discharge curves shortly after discharge has begun. Then there is the issue of the % of tetrabasic vs tribasic plate material for batteries that offer both starting & deeper cycle performance. This mix of plate material also charges voltage behavior.

If you wish to develop some sort of reliable SoC monitor it will need to be one that logs charge in/out and every few months you do a relatively deep cycle discharge to characterize your system based on battery aging. SLI batteries can't tolerate more than a 50% deep discharge w/o losing plate material and degrading longevity. I believe products like this exist for the marine industry. With so many battery variants available, you must characterize each battery to profile its SoC before the monitor can be useful.

There is a lot of information on these matters online but it takes significant research to develop a handle on how these batteries actually work given all the improvements made in the last few decades which affect apparent SoC.