Research on the detection method of natural commutation point of three-phase power supply

When using three-phase AC, the natural phase change point of the three-phase AC is often used as the reference point for control. Therefore, it is necessary to detect the natural phase change point of the three-phase AC to ensure the safe and reliable operation of the electrical equipment. At the same time, the frequency, phase sequence, and phase loss of the three-phase AC are monitored in real time, and when the three-phase power supply is abnormal, corresponding alarms are issued and protective measures are taken.

Working principle of phase change point detection

At the natural phase change point of the three-phase power supply, the voltages of the two phases are equal, and it is the starting point of the voltage reversal. This design uses this feature to realize the detection of the natural phase change point, as shown in Figure 1.

Figure 1 Two-phase sinusoidal voltage waveform

In one cycle, there are two intersection points between u1 and u2, namely a and b. Point a is the starting point of u2>u1, and point b is the starting point of u1>u2. This design detects point a, generates a corresponding pulse signal at time a of each cycle through circuit transformation, and sends the signal to the external interrupt of the microcontroller. The microcontroller processes and judges the interrupt, thereby detecting the natural commutation point, and at the same time, the frequency, phase sequence and whether the three-phase power supply is missing are judged by software.

Hardware circuit design

There are many methods for detecting the natural commutation point of the three-phase power supply. Most of them use analog circuits to compare the voltage between phases through comparators, but the accuracy of this method is not high, which will directly affect the accuracy of output voltage control; there are also digital circuits, but most circuits have the disadvantages of more devices, complex circuits, and occupying more microcontroller resources.

Traditional detection circuit design: Figure 2 is a schematic diagram of a traditional digital natural commutation point detection circuit, which occupies 7 I/O ports and also requires ADC, numerical registers and pulse logic combination circuits. The circuit is quite complex, and the program design also needs to be designed before the natural commutation point can be judged. In addition, this method also has the problem of commutation point loss, and the probability of loss increases with the decrease of sampling frequency, which greatly reduces the reliability of control.

Figure 2 Schematic diagram of traditional digital natural commutation point detection circuit

Detection circuit design

The following detection method is designed to overcome the problems of traditional detection methods, as shown in Figure 3.

Figure 3 Schematic diagram of three-phase AC natural commutation point detection

When u 1 >u 2 , the voltage across the Zener diode is 5V, and capacitor C1 is charged because the capacitance of C1 is small. The time of u1>u2 is half a cycle, that is, 0.01s, which is enough to ensure that the charging of capacitor C1 is completed. At this time, the diode connected in parallel at both ends of the base and emitter of transistor Q1 provides a clamping voltage, so that the transistor works in the cut-off area, the control diode of optocoupler U1 is not conducting, the controlled transistor is cut off, and the external INT0 of the microcontroller is pulled to a high level. After passing point a, u2>u1, U2 provides the base voltage for transistor Q1, and capacitor C1 provides the collector-emitter voltage, and transistor Q1 is turned on. During this period of time, C1, the control diode, R3, and Q1 form a loop, the controlled transistor in the optocoupler U1 is turned on, and the external interrupt INT0 of the microcontroller is pulled down to a low level. In this process, the microcontroller captures this falling edge to detect the intersection point a of u1 and u2. The optocoupler U1 realizes the electrical isolation of the input and output ends, and at the same time improves the anti-interference ability of the system.

The other two detection circuits are designed in the same way, with input voltages u 1 and u 3 , u 2 and u 3 , and outputs INT1 and INT2, respectively, to complete the detection of the other two natural switching points in the same cycle.

Pulse width calculation

Ignoring the transistor conduction voltage drop, the loop composed of C1, light-emitting diode, and R3 can be equivalent to a first-order zero input of an RC circuit (Figure 4).

Figure 4: First-order zero-input response of an RC circuit

u0 is the stable voltage of the voltage-stabilizing diode VD, the conduction voltage drop of the light-emitting diode is VF, when t ≥ 0, the energy stored in the capacitor is released through the light-emitting diode and the resistor, and the light-emitting diode emits light during this period. According to KVL, we can get:

Substituting it into formula (1), we get:

According to the initial condition uC(0+)-VF =uC(0-)-VF =u0-VF, and let the general solution of equation (2) be uC-VF = Aem, the solution of the first-order homogeneous differential equation is:

Let the product RC = τ, τ is the time constant of the RC circuit, reflecting the decay rate of the capacitor voltage uC, and equation (3) can be written as:

When uC decays to a value less than VF, the diode is cut off.

Solving equation (5), we get

That is, the diode conduction time is: seconds.

Simulation

The above design scheme was simulated and verified in the Saber environment. The simulation results are shown in Figure 5, where the two sine waves are the voltages of two phases of the three-phase power supply, and the pulse voltage is the signal sent to the external interrupt INT0 of the microcontroller. It can be clearly seen from Figure 5 that near the natural commutation point, a group of pulse signals with an amplitude of 5V, a pulse width of about 2ms, and a frequency of 50Hz are generated, and at the natural commutation point, a falling edge of nearly 90 degrees is generated, which is conducive to the capture of the microcontroller. Changing the time constant τ can change the pulse width of the pulse voltage. As shown in Figure 5, a group of pulse signals with an amplitude of 5V, a pulse width of about 10mS, and a frequency of 50Hz are generated at the natural commutation point. Figure 5 shows the pulse voltage obtained by the detection circuit, and Figure 6 shows the pulse voltage obtained by the detection circuit with different τ values.

Figure 5 Pulse voltage obtained by the detection circuit

Figure 6 Pulse voltage obtained by different τ value detection circuits

Software Design

The pulse voltage signal obtained by the detection circuit is sent to the Atmega16 microcontroller for processing. The microcontroller program is compiled in C language. The main program flow is shown in Figure 7. It is mainly divided into three modules: phase loss judgment, normal frequency judgment and phase sequence judgment. After the system works normally, the three modules are executed cyclically.

Figure 7 Main program flow chart

In order to ensure the safe operation of the system, corresponding protection modules can be added according to different occasions and application environments. For example, in motor control, when the frequency is abnormal, the power supply output of the motor is stopped.

Conclusion

The natural commutation point of the three-phase power supply plays an extremely important role in the use of the three-phase power supply. The accurate detection of the natural commutation point can effectively improve the control ability of the three-phase power supply.