Advertisement

Article

A Low-Cost Smart Web Sensor for Power Quality Monitoring

In recent years, power interference sources were widely spread throughout the main power lines. The PQ disturbances occur in industrial, domestic, and commercial systems. Examples include electronic power control with non-linear components, such as PCs and industrial systems with speed-controlled drives. These systems are also commonly used in domestic appliances such as television sets and inexpensive lamps. PQ disturbances can occur over a large range of frequencies. High-order harmonics, voltage fluctuations, and high-frequency disturbances with slew-rates on the order of microseconds or less are primary interference problems. In some cases the frequency and width of these voltage variations can produce a physiological irritating phenomenon due to lighting fluctuations (flicker effect). Since power disturbances can produce many problems in electrical and electronic systems, power-supply quality monitoring has become an issue of international interest.

Technology advances have facilitated the development of small, low-power devices that combine programmable general-purpose computing with multiple-sensing and wireless-communication capabilities. Composing these sensor nodes into sophisticated ad hoc computational and communication infrastructures to form sensor networks will have significant impact on the identification of electromagnetic disturbance sources. With this aim in mind, we designed a low-cost smart web sensor that allows the implementation of a distributed PQ measurement system in large industrial environments using a standard Internet browser. The system, which complies with the IEC standards, has been characterized and this paper also presents the more significant experimental results regarding use of the sensor.

Sensor Hardware Architecture

We designed the architecture of the smart web sensor to satisfy the requirements of low cost, continuous acquisition and data processing, and wide remote communication capability. A block diagram of the system is shown in Figure 1 .


Figure 1:  Architecture of the proposed smart web sensor

The interface with a low-voltage distribution system requires:

  • Sufficient insulation between the power circuit and the measuring circuit in order to guarantee operator safety and to protect the measurement apparatus
  • Good linearity
  • Bandwidth suitable for monitoring fast transient events
  • Small size and weight.

The voltage input transducer is based on an active attenuator. Transducer features include:

  • Nominal input voltage up to 600 Vrms
  • Overall accuracy of ± 0.3%
  • Thermal drift < 0.1%
  • Linearity < 0.2%
  • Slew rate of 10 V/µs.

The device must be able to measure several PQ parameters, including true voltage rms amplitude, voltage peak amplitude, voltage total harmonic distortion, and amplitudes of the fundamental frequency component and of the most important harmonics. The input voltage acquisition should be performed without triggering, in order to avoid undesired acquisition gaps during fast transient events. Automatic data storage should be performed with a threshold-based technique, using reference values defined via software.

To optimize system performance and increase measurement speed, we assigned the data acquisition, processing, and communication tasks to different blocks—a voltage peak detector; true rms-to-dc converter, input signal acquisition and preprocessing, final data processing, and data transfer to the web-site managing unit. The apparatus has been provided with a communication link to a host PC, for testing operations.

The peak detector is based on a full-bridge rectifier circuit—a software correction procedure evaluates and compensates for the voltage drop across the diodes. The rms voltage measuring circuit is based on the MX536AK true rms-to-dc converter. This device offers the following features:

  • rms measurements for both ac and dc signals
  • Bandwidth of 2 MHz for Vrms > 1 V
  • Low supply current of only 1.2 mA
  • Total error of ±2 mV ±0.2% of reading
  • Input signal range from 0 to 7 Vrms with a ±15V supply
  • Input resistance of 100 MW .

Figure 2 shows the electrical circuits of both the rms-to-dc and peak-to-dc converters.

Data acquisition is based on a microcontroller, the Microchip PIC16F877. This chip performs the following operations:

  • Continuous sampling, without triggering, of the outputs from the voltage transducer, peak detector, and rms-to-dc converter
  • A/D conversion with a 10-bit resolution
  • Pre-processing of acquired raw data
  • Data transfer to the web-site managing unit
  • Communication with the host PC during the apparatus testing.

The processing and communication unit is based on the 80152 microcontroller. This component carries out the final data processing and performs results visualization by means of regeneration of the HTML pages in the web server. The 80152 also controls the TCP/IP communication. The proposed hardware system has a total cost of around 300 Euros.

Control, Measurement, and Testing Software

The proposed smart web sensor requires software for three different tasks—controlling data acquisition, processing raw data, and transferring measurement results via TCP/IP. The data-acquisition control is performed by the PIC16F877 microcontroller, programmed in assembler language. The implemented software controls the sampling rate, input channel scanning, and digital output. An RS-232 standard communication port is used for testing microcontroller performance with a host computer. Both the raw data processing and communication are performed by the 80152 microcontroller. The 80152 has been programmed in Dynamic C, mainly due of the availability of many standard libraries.

Measurement of the acquired voltage's harmonic content was done according to the EN 61000 4-7 standard, using the FFT-based approach. In detail, the data acquisition system has been configured with:

  • A sampling rate of 12.8 kHz
  • An acquired window width of 320 ms, equivalent to 16 cycles of the 50 Hz fundamental frequency
  • A rectangular window
  • No gap and overlapping between successive windows
  • A frequency resolution of 3.125 Hz.

These features make possible the simultaneous measurement of quasi-stationary and fluctuating harmonics, inter-harmonics, and spurious components.

We also evaluate the voltage total-harmonic-distortion index (VTHD%). The actual version of the software measures the VTHD% along with the amplitudes of the fundamental frequency and the 3rd, 5th, and 7th order harmonics. The sensor acquires both the peak and rms values at the described sampling rate, with a D t of 78 µs. You can use the peak value as a first index for the evaluation of transient disturbances. The rms value is helpful in the evaluation of different kinds of disturbances: voltage short dips, with a width of less than one period of the fundamental frequency and a voltage shape distortion; voltage long dips, with a width of more than one period, without a voltage shape distortion; and voltage flicker. You can use the relative voltage change for initial flicker evaluation, that is, the difference of any successive values of the rms voltage, according to the EN 61000-3-3 standard.

Testing the data acquisition and processing requires performing three different testing programs, for measuring peak and rms voltages, VTHD%, and harmonic amplitudes.

Figures 3, 4, and 5 show the front panels related to the testing programs. A host PC performs these tests, using the National Instruments LabVIEW tool. Both the PIC16F877 and 80152 microcontrollers are linked to the PC via an RS-232 standard communication port.


Figure 3:  Front panel for peak voltage
measurement testing


Figure 4:  Front panel for rms voltage
measurement testing


Figure 5:  Front panel for VTHD% and
harmonics measurement testing

Concerning the web interface, our most important aim has been the possibility of loading the page using widely adopted web browsers, such as Internet Explorer and Netscape.

In the HTML page the measured quantities are stored as HTML variables. When the web browser receives a refresh command, a new HTML variable set is generated by the 80152 microcontroller. To guarantee a reasonable working of the browser, we limited the access to a maximum of six users. As an example, Figure 6 shows the measurement of a steady-state distorted voltage waveform. Figure 7 shows a voltage waveform affected by voltage dips.


Figure 6:  Example of measurement of a steady
-state distorted voltage


Figure 7:  Example of a voltage
waveform affected by dips

System Characterization and Experimental Results

We have done system characterization for the rms voltage measurement, the peak voltage measurement, and the VTHD% evaluation.

For instrumentation, we used the Keithley 2001 mutimeter for measuring both peak and rms reference voltages. To check system performance related to the measurement of both total harmonic distortion and harmonic components we generated the reference voltage signal by means of the Tektronix AWG 2005 arbitrary waveform generator and the Kepco BOP 1000 power amplifier. To verify performance with an input voltage affected by transient disturbances, we generated the reference signal with the EMC Partner Transient 2000. Tables 1 through 3 show some experimental results.

Reference Vrms
Measured Vrms
Relative Error (%)
179.2
176.8
1.34
189.6
188.0
0.84
199.2
198.4
0.40
209.6
208.0
0.76
219.7
218.4
0.59
229.6
228.0
0.69
239.2
238.4
0.33
249.6
250.4
0.32
260.1
258.4
0.65

Table 1:  Voltage rms measurement error for
different values of input voltage

Reference Peak (V)
Measured Peak (V)
Relative Error (%)
254.4
256.0
0.63
269.6
271.2
0.59
282.4
284.0
0.56
297.6
297.9
0.10
311.2
312.0
0.26
325.6
324.8
0.24
339.7
339.2
0.14
354.4
352.8
0.45
368.81
367.2
0.43

Table 2:  Peak value measurement error for
different values of input voltage

Reference VTHD (%)
Measured VTHD (%)
Relative Error (%)
1.09
1.07
1.83
2.08
2.05
1.44
3.11
3.07
1.29
4.12
4.07
1.21
5.10
5.05
0.98
6.09
6.04
0.82
7.06
7.00
0.85
8.10
8.04
0.74
9.15
9.09
0.65
10.19
10.13
0.59

Table 3:  VTHD% measurement error for
different values of distortion

Conclusions

We proposed a smart web sensor for power quality monitoring. The sensor evaluates some important power-quality indices in the low-voltage main power lines, and does data storage and communication via TCP as a web server. Main advantages of this sensor are a simple hardware architecture, low cost, and wide connection capability. Both hardware and software have been implemented according to the IEC (EU) standards related to harmonics measurements. We performed system characterization, achieving satisfactory results.

0 comments on “A Low-Cost Smart Web Sensor for Power Quality Monitoring

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.