Battery life simulation
charging and discharging simulation
Introduction
A goodstrategy is required to reduce the energy used in IOT devised which are controlled using microcontrollers. He re one needs to take care of the scheduling of tasks . The steps are
(i) the dependency of the usable capacity on load current and
(ii) the recovery of the capacity when rest periods are inserted.
Here a State-of-the-art approaches are studied by using simulation using MATLAB/SIMULINK as programing too.
The methods are demonstrated using simulation using models both of the workload and the battery. In the IoT domain, conversely, devices rely on microcontrollers, with varied workloadsand, batteries needs to be sized to guarantee lifetimes say days.
Load current magnitudes in these IoT devices are therefore relatively small compared to other more powerful devices, and they hardly trigger the conditions that emphasize the battery non-idealities.
We need to carry out a battery charging and discharging cycle assessment about whether task scheduling is really relevant to extend lifetime in IoT devices.
How to carry out Battery life simulation
Background, is taken from work ot Tim. Tim carried out a measurement-based assessment about whether task scheduling is really relevant to extend lifetime in IoT devices. We run experiments both on a physical device hosting four sensors, an MCU, and a wireless radio, as well as on a “synthetic” device emulated with a programmable load generator. We used both secondary lithium-ion and primary alkaline batteries to explore the impact of battery chemistries further. Results show that the impact of different schedules is essentially irrelevant, with a maximum difference of only 3.81% in battery lifetime between the best and worst schedules.
The effective capacity of a battery (i.e., its available energy) decreases as the discharge current increases. Hence we need to match power levels of battery and load to reduce converters’ losses and reduce the rated capacity effect.
The magnitude of the recovery effect is strongly dependent on the battery chemistry; recent works have shown that in typical lithium-ion batteries, the recovery effect is virtually absent [15]. Therefore, an IoT device powered by a lithium-ion battery playing with idle times will hardly benefit.
The battery capacity reduces with increase of discharge rates. When the discharge current increases from 25mA to 500mA, the corresponding available capacity decreases from 3,000mAh to 1,700mAh.
For a Li oin battery, of nominal capacity 3,200mAh. The usable capacities are obtained when the battery voltage reaches the cut-off voltage value (2.5V in this case). We can immediately notice how the sensitivity to current is quite modest and shows only a small improvement for a 0.2C current (i.e., 3,200mA*0.2 = 640mA). The rated capacity effect is much more significant in the primary battery than it in the lithium-ion battery.
Recovery effect. Compared to the previous two, this one is potentially beneficial from the battery duration perspective. It states that if the battery stays in the periods of rest (zero or very little discharge) for a sufficient duration, capacity lost during discharge can be replenished to a certain extent. This phenomenon occurs because, when the load is reduced, reactants in the battery diffuse to the reaction location, allowing more of them to be used during the battery lifetime.
Different sensors are characterised as per Current requirements and voltage loading on the power management system as shown in Table 1
Table 1 : Voltage demanded and current drawn by different Units
Battery capacity is defined by parameters given by Battery fabricator as shown in table 2
Table 2 : Battery parameters as defined by battery fabricators
We can simulate the battery power loading as per three schedules
• Full serial schedule:
All sensors are powered in series in non-increasing order of power consumption. Each sensor has a 15-seconds sampling period. After all the sensors finish sampling, MCU is powered for 3 seconds to process the captured data. Then we power the transmitter after 2 seconds who will transfer data and finish data transmission. As soon as the transmitter finishes the operation, the device goes to the idle state until the next cycle starts.
• Full parallel task schedule:
Here sensorsare not activated sequentially, but all the sensors are activated together by a 15-seconds sampling period at the start of each cycle. The rest of the procedure is same like serial policy.
• Intermediate task schedule:
Here six sensors are divided into three groups. The sensors in each group collect data simultaneously. Groups are activated one after another in serial. After all groups finish sampling, the rest procedure is identical to the previous two schedules.
The battery charging and discharging can be programed as shown in following SIMULINK model
The sensor loading as per Table 3 are simulated, and we can see the charging and discharging patterns.
Power consumption of different sensor units is summarised in Table 3
Table 3: Power requirements of sensors
A battery charging and discharging Simulation model is made in MATLAB 2021B, as shown in Figure 1
Figure 1 : Battery charging and discharging model programmed in SIMULINK of MATLAB 2021
The exploded view of Loading on Battery due to sensors is given in figure 2
Figure 2 : The exploded view to show how different units are modelled
The SOC charging and discharging
Charging happens when SOC < 40 and it charges from 40% to 80% SOC.
We can see three charging and discharging pattern here in figure 3
Figure 3 : The charging and discharging pattern
Here the units demands different currents and duration for current drawn can also the changed as per sensor units.
For sensor unit 1 here let us see how it done.
We use logic commands and here sing from 25 sec to 50 sec this unit demands power. As per this the logic is build up.
The sensor current drawn timings can be changed as required. The exploded view of one of sensor sub block is shown here in figure 4
Figure 4 : Exploded view of logic defining timings for power demand of the units
The subblock defined the amount of current to be drawn.
Now let us see the exploded view of the sub block.
The sub block Exploded view is shown in figure 5
Figure 5 : Exploded view of the logic and programing for arriving at the current demand by the units.
The current is wattage/voltage. So we define the wattage and the voltage and using division we arrive at the current in Amp, or milli amp as the wattage definition.
So in the main program model we have defined the sub blocks for power demand for each unit of units as given in table number 3.
The bottom part of the diagram takes care of Battery charging and discharging logic as shown in figure 6
Figure 6 :Battery charging and discharging programing block diagrams
The SOC plot can be extracted by clicking on the plot unit to variable SOC as marked here in figure 6.
Appendix A explains the modleing components used.
Appendix A
Different sensors modules
Adder to add the current demang from the battery
Battey capacity can be set by chosing the battery parameter setting
Battery SOC is stored in the variable SOC, which is then connected to a plot sub program for displaying the charging and discharging
The discharging of the battery is controlled using a ideal switch
The battery charging is done using DC souce like abattery, and sincing charging requires power to be send in puses, suitable puge generator is used.
The pulse generator parameter can be set as required
The battery parameter selection can also be done as required
Here charging is set to start when SOC reached 40 and the charging is stopped whe SOC reached 80.
These limits can be changed as per simulation requirement.
