Design of a 24V DC uninterruptible power supply for distribution network

3.3 Battery parameter detection

₁ ) Charging current Ic and discharging current _Id When the battery is in charging state, due to the clamping effect of D4 _, the load current _{Io is} completely _provided by _{the charger} . At this time, _{the terminal} voltage of R7 U _R7 = Io _R7 , Id ₌ 0 , I _R8 = Io + Ic , _and R7 ₌ R8

I _c = ( UR ₈ - U _{R 7} ) _/ R ₇ ;

When powered by a battery alone, the clamping effect of D4 _disappears . At this time, _{UR8 = 0, Ic = 0, therefore, Id = Io = UR7 / R7. Therefore, as long as UR7 and UR8} are _used to obtain a 0-5V current _signal through a differential amplifier and _sent to the _{two A / D} conversion _{channels of PIC16C73 ,} Ic and Id at any moment can be detected through processing by the _{microprocessor} .

2) Battery voltage Since two 12V batteries are connected in series, the battery terminal voltage should be detected separately, and the voltage output is obtained by resistor voltage division to obtain the voltage sampling value. When the battery voltage U _B1 ( U _B2 ) is detected to be | ( U _B1 -12) |/12 ≥ _{δ 0} (δ ₀ is the balance rate, 4% here), it means that the battery voltage is not charged evenly, and corresponding measures should be taken.

3) Battery temperature: The AD590 temperature sensor is used to send the temperature sampling value to the single-chip microcomputer. When it is detected that the battery temperature exceeds 80℃ for more than 10 minutes, the discharge control signal is immediately canceled, and the high level of R5 is changed to a low level to disconnect the relay.

4) Battery capacity Battery capacity detection has always been a difficult problem in battery management. Common methods include: capacity prediction based on electromotive force, capacity prediction based on battery internal resistance, capacity prediction based on both battery internal resistance and electromotive force, capacity prediction based on current discharge rate, capacity prediction based on current time integral, etc. In this system, since the load change follows _Io = 0.2n ( n is the number of parallel loads), the capacity detection uses the capacity prediction of current time integral, which makes the detection simple and feasible. The battery discharge capacity CΔ = Iddt _. Due to the switching of the load , the current changes according _{to a fixed rule, so CΔ = 0.2n1t1 + 0.2n2t2 (n1t1, n2t2} is the time _{of different} loads ) . _If the initial _capacity C o of the battery before discharge _is known _{, the} battery _{capacity after the} change _is Cx = C _{o -} CΔ . This method is relatively simple and easy to implement, and can use the current sampling circuit of the system itself without the need for additional special equipment.

3.4 Prediction of remaining time

The purpose of battery capacity prediction is to obtain information about the working time that the battery system can provide. Therefore, in fact, we only need to know the working time that the battery system can provide under the current conditions (voltage, current, temperature). At a certain moment, the voltage, current, and temperature values ​​can be measured, so we can predict the duration of the battery's constant current discharge at this current, that is, there is such a table in the system, which divides the voltage into several levels and the current into several levels, as listed in Table 1.

Table 1 Battery capacity prediction table

What the system needs to do is to fill the table and calculate the time the battery can run at the current according to the terminal voltage and current at a certain moment. The size of the voltage and current grading interval determines the accuracy of the battery remaining capacity prediction. The following takes a 12A·h/12V lead-acid battery as an example to illustrate the working process of the system.

1) Initialization of the table Initialization can be done in two ways: first, by obtaining data from the battery discharge curve provided by the battery manufacturer; second, by obtaining data from operation. The initialization data does not need to fill the table completely, but the amount of initialization data determines the accuracy of the remaining capacity prediction at the initial stage of system operation. We divide the current into 4 levels: 0.05 C / 0.1 C / 0.15 C / 0.2 C , and the voltage is divided into 0.1V levels.

2) Correction voltage At different discharge currents, the battery internal resistance and polarization voltage are different. Therefore, the correction voltage at different discharge currents must be obtained first. Using 0.05 C as the benchmark, the battery is discharged and the correction voltage at different voltage points is obtained.

_{3) Predicting the remaining time Based on the initialization result, obtain} part of the data in the prediction table. If t ( V1 , I2 ) is known from the prediction table, predict the remaining time _when I1 _discharges to V1 , _and use the conversion formula (2) to predict.

t ( V ₁ , I ₁ )= t ( V ₁ − V _{x 2} + V _{x 1} , I ₂ )× I ₂ / I ₁ (2)

Where: t ( V ₁ , I ₁ ) is the remaining time when the battery voltage reaches V ₁ after discharging at I _{1 ;}

Vxn is _the correction voltage corresponding to each current.

The selection of I ₂ takes into account the principle of proximity to ensure the accuracy of the prediction. The most recent discharge result is selected for prediction. For example, if the last discharge was 0.1 C and 0.2 C is to be predicted this time , I ₂ is 0.1 C. This is because the physical and chemical state of the battery is changing at any time. The closer the time is, the more accurate the result should be. The error of this method in predicting the remaining time is less than 15 minutes. In practical applications, it can improve the measurement accuracy of time and voltage, thereby improving the accuracy of prediction.

4) Correction of the prediction table The correction of the prediction table is carried out when the battery is discharged to the cut-off voltage and is divided into several cases:

(1) Before the end of discharge, the battery is discharged at a constant current, and the discharge prediction table under the constant current value is corrected;

(2) Before the discharge is completed, the battery is discharged by several currents, and the corresponding voltage prediction table of these currents is corrected by the time conversion formula (2);

(3) At different temperatures, correction is made according to -4mV/℃.

FIG4 is a flow chart of the remaining time prediction program.

Figure 4 Flowchart of the remaining time prediction program

The advantages of this forecasting method are:

——No need for a large number of preset curves;

——No need to add extra measuring equipment, making full use of the voltage and current measuring system of the original system;

——As the running time increases, the prediction accuracy increases;

——This prediction method can adjust the memory capacity as needed to improve accuracy.

4 Conclusion

The DC uninterruptible power supply system introduced in this article has high reliability, good EMC performance, simple battery detection scheme, convenient control operation, simple and practical remaining time prediction method, simple and flexible software and hardware design. It can be seen from the use of the substation site that the effect is very good.