Automatic meter reading system based on two-way power frequency communication

At present, technologies such as spread spectrum, ultra-narrowband communication, multi-carrier modulation, adaptive frequency hopping, and orthogonal frequency division multiplexing (OFDM) have been applied to automatic meter reading systems in China. However, due to the harsh low-voltage power grid environment in my country, the application effect of the above technologies has always been unsatisfactory. Power frequency distortion is the abbreviation of Two Way Automatic Communication (TWAC). Compared with carrier communication, it has the advantages of long effective transmission distance, reliable communication, and signal can pass through transformers.

1 System composition

Automatic meter reading The system is generally composed of a collection terminal, a concentrator, and a master station system for collecting user electricity meter information. The collector and concentrator are located under each transformer to manage the data uploaded by the collector and serve as a bridge between the collector and the master station (the master station is a computer management system located at the top level). The automatic meter reading system based on two-way power frequency automatic communication can transmit distorted signals across transformer stations. There is no need to add a concentrator as a connecting bridge between the collector located near the meter and the master station. The system composition is shown in Figure 1.

The two-way power frequency automatic communication system consists of a central control unit, a modulation and demodulation device at the substation end, and a modulation and demodulation device at the user end. Its principle is to use the zero-crossing modulation method of the power frequency voltage fundamental wave to achieve communication. For outbound signal modulation, at the moment △T/2 (30 degrees before the zero-crossing point) before the voltage crosses the zero point, the thyristor in the modulation circuit Figure 2 (a) is turned on, and the instantaneous current generated is coupled into the current ic of the power frequency voltage, causing a voltage drop emod, and distortion occurs at the zero-crossing point of the 10 kV voltage E, as shown in Figure 2 (b). The encoding of the voltage distortion signal uses two adjacent periodic voltage waveforms to carry one bit of information, and uses the different modulation positions to represent "1" or "0". The modulation method of the inbound signal is similar to that of the outbound signal, except that the inbound signal modulation is to add the distortion signal to the current at the moment when the voltage crosses the zero point.

2 Hardware Circuit Design

The modulation circuit system is mainly composed of a filter circuit, a zero-crossing detection circuit and a modulation circuit. The equivalent circuit of the modulation circuit has been given and will not be discussed again. The hardware circuit shown in Figure 3 is a filter circuit and a zero-crossing detection circuit composed of A1 (MAX-291), A2 (TA7504P) and A3 (OP07). A1 can change the filter cutoff frequency by changing the clock input frequency, and the cutoff frequency is 1/100 of the clock frequency. A square wave signal with a 5 V level is added to the clock input end, and the characteristics of a low-pass filter can be obtained between the input (IN) and output (OUT) ends of A1. A2 is used to smooth the output step waveform of A1 and enhance its effect. The zero-crossing detection circuit is mainly composed of an op amp OP07, 4 diodes and 1 transistor.

The detection circuit of the distortion signal is composed of a pre-filter circuit, a data acquisition circuit, and a data processing circuit. The data acquisition circuit is composed of an AD1674 chip and its peripheral circuits. The core of the entire detection system is a single-chip microcomputer system consisting of a PLC24 series microprocessor chip plus necessary data storage, program storage, and necessary input and output circuits. The single-chip microcomputer technology is very mature, and only the system block diagram shown in Figure 4 is given here.

3 Signal Detection

Signal detection is a problem of determining whether there is distortion at the zero-crossing point. At present, the digital differential technique is generally used in China for detection, that is, the difference between the previous sampling value and the current sampling value is calculated.

If F(t)=A1 sin(ω1t), T is an integer multiple of its period Tper, then d(t1)≡0. From this result, it can be seen that when the digital differential technology described by equation (1) is applied to a periodic signal with a stable period, its differential result is always equal to O. However, due to the complex environment of the power grid channel, which is full of a large number of harmonic components and noise interference, the digital differential technology that is very feasible in theory is not ideal in practical application.

The wavelet detection method used in this paper is a powerful tool for time-frequency analysis. The continuous wavelet transform of the signal x(t) is:

Where: a is the scaling factor; b is the translation factor. The discrete wavelet function ψj, k(t) can be expressed as:

In order to make the wavelet transform have variable time and frequency resolution, it is necessary to change the size of a and b so that the wavelet transform has the function of "variable zoom". In practice, the binary discrete wavelet is widely used, that is, using a binary dynamic sampling grid, a0=2, b0=1, the scale corresponding to each grid point is 2j, and the translation is 2jk. The resulting wavelet ψj,k(t) is called a dyadic wavelet.

The binary wavelet has a zooming effect on the signal analysis. Assume that a magnification of 2-j is selected at the beginning, which corresponds to a certain part of the observed signal. If you want to further observe the smaller details of the signal, you need to increase the magnification, that is, reduce the j value; conversely, reduce the magnification, that is, increase the j value. Any signal can be expressed in the form of formula (5):

The values ​​of j and k are both within ±∞, which means that refinement is performed at all scales to supplement detailed features. When analyzing various signals from the perspective of scale, after exceeding a certain scale (for example, j0), the detailed features no longer work. At this time, equation (5) can be divided into two parts with scale j0 as the boundary. The scales below j0 are used as approximations of refined features; the scales above j0 are used to extract basic features. From the perspective of filtering, the scales below j0 correspond to bandpass filter groups with different center frequencies, and the scales above j0 correspond to low-pass filter groups with different bandwidths. Equation (5) can be expressed as: The first part on the right side of the equation can be regarded as the approximate low-frequency signal of the signal x(t) with a scale of 2j0; the second part can be regarded as the detailed high-frequency signal of x(t). The approximate signal of any scale can be expressed as the sum of the approximate signal of the next scale and the detailed signal.

4 Simulation Experiment

According to the equivalent circuit, a power frequency distortion signal simulation circuit is built in Simulink, and a positive pulse is added to the 1st and 3rd zero-crossing points of two consecutive cycle voltage signals. The generated single-phase voltage distortion waveform (the distortion signal is exaggerated) is shown in Figure 5. It can be seen intuitively from the simulation diagram that the two distortions of the voltage waveform occur near the sampling points 50 and 150.

200 sampling points are taken within two power frequency cycles (0.04 s), and the distorted signal is divided into two sub-signals using wavelet basis db4, as shown in Figure 6, the approximate signal a1 (i.e., low-frequency signal) and the detail signal d1 (i.e., high-frequency signal).

The approximate signal a1 is similar to the original signal (Figure 5); the detail signal d1 has strong changes near the sampling points 50 and 150. The distortion point of the signal can be clearly found from d1, so the detection of the distortion point by the binary wavelet change method is to detect the change in the detail signal and determine the distortion moment by setting the threshold.

5 Communication Protocol

The low-voltage distribution network channel environment is complex and the data transmission distance is limited. In order to ensure the reliability of communication and expand the transmission distance, a relay is needed at the collector end of the meter reading system. Based on DL/T6 45-1997, the frame format supports the control of relay forwarding and requires that the frame cannot be too long. The basic frame format is shown in Table 1. Among them, each byte contains 8 bits of binary code, and a start bit, a check bit and a stop bit are added during transmission, for a total of 11 bits. In the control code C, when D7=O, it is the command frame sent by the master station, D6, D5 control relay forwarding, and D4~DO are used for function coding control; when D7=1, it is the response frame sent by the collector.

6 Conclusion

The automatic meter reading system under industrial frequency two-way communication implemented in this paper has strong stability in the complex channel environment of power lines. The communication distance is significantly improved compared with the traditional spread spectrum carrier meter reading system. The use of low-voltage power grid as the communication medium saves the cost of building the system. It is a meter reading system that is very suitable for my country's power channels and will become a key research direction of automatic meter reading systems in the future.