20 I?

5th International Conference on Information Science and Technology (IClST)

Apn1 24--26, 2015, Changsha, Hunan, China

Applications of FFT Algorithm to Energy Management System of

Electric Equipment for Underground Coal Mine

Xiangzhong Meng ,Wenchao Wang ,Zisheng Hou and Yonggang Chen

Abstract-The essay probes into the present problems of the II. FFT ALGORITHM
measurement errors caused by harmonic interference in the
FFT algorithm has already developed for over 50 years,
electric monitoring equipment for underground coal mine. We
especially in the rapid development in recent years, the
applied three-phase watt-hour meter in view of FFT (Fast
traditional fast Fourier transform algorithm is a recursive
Fourier Transformation) algorithm based on Industrial
decomposition algorithm, which comprising four different
Ethernet, which can effectively analyze the harmonic
forms: radix-2, radix-4, split-radix and the Bruun FFT [2].
component of the voltage and the current from the field. This
can accurately calculate the electrical parameters of electrical
The FFT indexing algorithm of fused multiply/add
equipment, improving the quality of electricity and achieving operations proposed by Linzer-Feig is a new development of
.
optimal energy management of coal mine. the traditional FFT algorithm(scaled algorithm)[3][4], it uses
advanced workstation and signal processing chip to
I. INTRODUCTION complete the integration operation of multiplication and
addition effectively[5], its basic operation is ±a±bc where a
. � � h �
n recent years, with the int grated applicati n of power
Ielectrolllcs
.
and commUlllcatlOn
, technology m electrical
equipment for underground coal mine, the energy
b, c are real numbers, it only needs one single mac ine cycl
whether plus(minus) after multiplication or single
mUltiplication or single plus, so all of these arithmetic
management style of coal mine increasingly tends to be
operations are caller multiply/add operations (mla-op). Some
�
auto atic and networking. The real-time and accuracy of
articles [6] have used mUltiply/add operations to deal with
electnc energy acquisition, the accuracy and rapidity of
FFT.
algorithm processing are the essential requirements of online
FFT algorithm is to simplify the complex sequence
monitoring in the grid of coal mine. The number of electrical
through optimizing the process based on the discrete Fourier
equipment in coal mine is very large with widely distributed,
transform (DFT), which can reduce the computational
complex wiring, which are easy to produce harmonic
complexity and easy to implement the rapidity of power
interference, it can affect the accuracy of watt-hour meters
energy monitoring.
and result in confused metering and shorten the service life
of equipment in coal mine, which can't meet the accurate For N-point fmite sequence X(n), its DFT formula used in
monitoring of the electrical equipment for underground coal the plural is expressed as:
N-J
mine and analyze the consumption of energy power , it isn't X (k) = L x(n)v;/ , k = 0,1,,,, N -1
(1 )
conducive to achieve the integrated requirements of security, n-O

advanced, efficient and energy-saving . I N-J

For the present situation that the harmonic interference x{n)= N � X{k)w;nk,n=O,l,... N-l
(2)
will cause measuring errors in the energy monitoring of
electric equipment for underground coal mine[l], the Where,
questions of traditional algorithms(such as DFT algorithm W;:k = cos ( 211'n �) + jsin ( 211'n �) (3)
and integration algorithms) are lack of real-time and We can know from formulas (1) and (2) that calculating
accuracy when computing. This project is based on the
the N-point X\n) needs N' times complex multiplication
industrial Ethernet to realize network control, and combining
and N(N-I) times complex additions, the amount of the
the three-phase watt - hour meters based on FFT algorithm,
which can effectively analyze the harmonic component of computation of X(n) will be large. When the value of N is
electrical parameters from the field, and ensure the reliability big, it's difficult to meet the requirements of real-time and
of the analysis of power consumption , which is conducive accuracy, besides, a lot of trigonometric functions may cause
to realize the optimal control of energy saving of electric errors. The FFT algorithm uses the inherent symmetry and
equipment for underground coal mine. the periodicity of butterfly factor [7] W thus:
N
'

Xiangzhong Meng, Wenchao Wang and Zisheng Hou are with the X(k)=� (k)+�X2(k) (k=O,I, , � ) ... -1 (4)
Department of Automation and Electronic Engineering,Qingdao University
of Science and Technology, Qingdao, China (Phone: 13105325633; email:
[email protected] )
X(k+ �)=Xl (k+ �)+Wik+�)X2(k+ �) (5)
Yonggang Chen is with the Daizhuang Coal Mine, Shandong energy
N
Zibo mining group Co., LTDJining, China (email: [email protected]) ( k=O " 1 ... --1)
' 2

978-1-4799-7489-4/15/$31.00 ©2015 IEEE

 For 8-point DFT, it can be decomposed into four 2-point least 10 times in one sampling period. The FFT algorithm
DFT, its flow chart of butterfly shows as Figure l. are sampled by the FFT simulations under 8,16,32 sampling
X(o)
points through the way of 2' sampling points[7], it shows:
the error of 8 points sampling is too large, it can't meet the
� X(O) requirements of monitoring accuracy of the equipment for
�i�l X(l) underground coal mine; the error of 16 sampling is small,
� X(2)
and the accuracy of measurement meets the basic
X(6) requirement; the accuracy of points more than 32 or 64
X(3) compared to 16 sampling points has increased, but the
X(4)
X(l) increment is a lesser extent[9]. Taking the factors of the
X(S) processing time of FFT algorithm and conditions of
WO
X(S) N X(6) hardware in energy metering devices into account, the
X(3) choice is 16 sampling points one period.
X(7)
�
X(7) IV. SYSTEM IMPLEMENTATION OF FFT ALGORITHM
Fig. I. Butterfly flow chart of FFT when N=8 Using the capabilities of data processing and computing
According to Figure 1, 8-point DFT using FFT algorithm of DSP[10] in intelligent data collection system terminal of
firstly makes input sequence in reverse order, and then electric parameter EDC3300, then analyzing the harmonic
calculates three-stage butterfly. In the same way, whenN=2"', of electrical parameter of the voltage and the current of
electrical equipment underground coal mine through the FFT
there are m-stage butterfly, each stage is composed of
algorithm, and then transforming and calculating the results
N /2 butterfly computation, so there are mXN/ 2 butterfly
after analysis and processing, ultimately, it can realize the
computation. In terms of multiplication [7], the number of
precise measurement of electric parameters of electrical
complex multiplication for DFT iSN2, while the number of
equipment for underground coal mine.
N
FFT iS2"log,N. The realization process of precise measurement of electric
Therefore, compared with the DFT algorithm, the parameters of electrical equipment for underground coal
computation of FFT algorithm is less andthe computing mine using the FFT algorithm as follows:
speed is faster with less errors. It improves the accuracy and
real-time of electric energy acquisition, the speed of
algorithms processing will be faster, which can meet the
real-time for monitoring online.

III. SELECTION OF SAMPLING POINTS

The accuracy of the measurement results of electrical

parameter of electrical equipment underground coal mine
depends on the correct choice of signal sampling frequency.
If the sampling frequency is too high, the sampling points in
one period will be too many. Besides, it will take up lots of
memory and the data processing time is too long, the
requirement of hardware conditions for energy metering Take the FFT operation of data read
fronl RAM directly
device is too high, it can't meet the requirement of the
rapidity and real-time of energy management of
underground coal mines. On the contrary, if the sampling f the calculation is
cOlllpleted
frequency is too low, the number of sampling points is too
y
small, there may be inaccurate waveform, resulting spectral
distortion and reducing the reliability of the system , which Calculation of electrical parameters

can't provide accurate basis of optimal control for energy

Display and systenl COllllllunication
saving.
The Nyquist sampling theorem provides the actual Fig.2.The flow chart of the measurement of electric parameter based
sampling frequency at least twice of the highest frequency of on FFT algorithm

the signal [8] fm, we often take the frequency as (4-10) 1. Considering the harmonic interference of the electrical
fm when practical application. As the electrical equipment Equipment for underground coal mine, and select the
for underground coal mine, the 3rd, 5th, 7th, 9th harmonic appropriate sampling frequency, now we select 16 sampling
are relatively more serious, now we take the analysis of the points per cycle in this system;
5th harmonic as an example, the maximum sampling 2. Taking the software program for FFT algorithm into
frequency of the signal should be 5 times of the power DSP processors, the processor take sample of the
frequency, namely 250Hz. To prevent aliasing, the actual three-phase power according to the set of the main program
sampling frequency should be higher than 500Hz, namely at and interrupt routine.
 3. Perfonning an FFT on the sampled signal, then The current signals are switched to voltage signals in
calculating the magnitude of each corresponding point. current transformer, and put the digital conversed into the
4. Extracting the amplitude of the fundamental frequency, memory directly through the direct memory access channel
fmd the rms voltage and current in accordance with the after A I D conversion in the operational amplifier waiting
fundamental amplitude, then calculating the active power for FFT module processing, and then get electrical
and reactive power. parameters and displayed by arc-tangent after the FFT
5. Sequentially extracting the amplitude of other operation. The entire measurement process is handled by the
frequencies, then find out the active power and reactive module, so it doesn't take up the CPU and improves the
power, finally got its active power and reactive power of the reliability of system operation.
total in accordance with the above process.
The design of the corresponding measurement module VI. SIMULATION ANALYSIS OF FFT ALGORITHM
based on the principle of above, it can accurately collect the In an ideal power supply system of electrical equipment
parameters of electrical equipment for underground coal for underground coal mine, the waveform each phase
mine, and the parameters is transmitted to the monitoring voltage is standard sinusoidal, but due to a large number of
station through the RS-485 bus, the monitoring station can electrical components which have characteristic of nonlinear
display the real-time and historical data in the form of used in coal mines [13], such as the inverters and the
reports, curves and so on. at the same time, the data will be rectifiers, which inevitably generate hannonics on the
transmitted to the main control room ground through the practical application wave interference, resulting in the grid
switches, the staff can analyze the power dissipation voltage deviates from sine wave. Now we use the conveyor
according to the collected data and make corresponding belt as an example to verify the validity and accuracy of the
decision, so as to adjust the running status of equipment FFT algorithm for harmonic analysis of electric equipment
under abnormal state timely[ll], which will ensure the for underground coal mine, using the Matlab/Simulink
normal operation of equipment, and optimizing the operation module to establish a three-phase rectifier circuit models
of the normal equipment, make sure that the with inductive load, in order to analyze the validity and
equipment work in the normal state as long as accuracy of the model compared the electrical parameters
possible[12], which will improve the power management calculated by theoretical formula with the output of the
level in coal mine. electrical parameters from Matlab model. The simulation
model is shown in Figure 4.
V. THE DESIGN OF MEASUREMENT MODULE BASED ON

FFT ALGORITHM
FFT algorithm can effectively analyze and calculate the
hannonic component of each electric equipment for
underground coal mine, It can analyze the harmonic
component of the voltage and current of electric equipment,
then take advantage of its fundamental voltage and current
of as the real and imaginary part of the function, the real and
imaginary parts are stored to two array, use the formula in
the part I to calculate the effective value of voltage and
current signals, active power P and reactive power Q, then
using the average power for the time integral to calculate
electrical degrees.
Taking the FFT module into the DSP control chip of the
Fig. 4. Matlab I Simulink simulation model of power system
three-phase watt-hour meter EDC3300, and using the
operational amplifier and current sensor, then applying the In the DSP operation, conversion factor of butterfly
function that direct memory access process and modulus operation are indicated used the plural so as to reduce the
sampling of micro-controller to achieve accurate FFT computation, and save the trigonometric function involved
computation. The measurement module of the hardware in solving parameters in an array, and calling the parameters
structure is shown in Figure 3. through the table, which can effectively shorten the running
time of the program and meet the requirement of real-time.
The FFT module is responsible to complete computation
of the input voltage, current signal in reverse order, butterfly
shape using FFT algorithm, so as to obtain the harmonic
content of the input signal, and then calculates the electric
parameters through various parameter calculation module,
thus completing the software analysis process.
The harmonic of conveyor belt is mainly produced in the
motor starting and frequency control process, the voltage is
660V while the frequency is 50Hz during the process of the
Fig. 3. The measurement module hardware structure motor starting, braking and speeding, giving a corresponding
three-phase power signal to the model and set the simulation values and measured values of the model is within 5%, it can
time as 0.1 s. We can get the signal waveform by the power more accurately calculate the actual electrical parameters of
system simulation model shown as Figure 5: each electrical equipment and efficiently analyze the
600
harmonic of power system. For high harmonic larger
is
400 equipment for underground coal mine, it should be
considered that adding a low-pass filter to filter out the
200
high-order harmonic interference so as to prevent the
waveform of the current and voltage from distortion, which
·200
can cause serious harmonic interference and result in
decreasing the power quality and harming the power grid of
-400 coal mine.
-600

VII. CONCLUSION
·800 '------:-':-:--:-':-::-
: --:-':-=---:-':c:----:--':-:---:c':-::-
:- --:-':-:::----:-":-=--�'_:_---:"
o 0.0 1 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 vsO.1
This paper apply the three-phase watt-hour meter based
Fig. 5. Voltage waveform
on the FFT algorithm to energy management system of coal
From Figure 5, we can find that the harmonic interference mine, which can effectively measure the harmonics of each
wave do exists, and then analyze the spectrum of the voltage electrical equipment and achieve accurate monitoring of the
signal waveform by SimulinkiFFT Analysis function in electrical equipment for underground coal mine, which can
powergui module; it's shown as Figure 6: improve the quality of energy power and achieve the target
of optimal management of electrical equipment for
underground coal mine.

REFERENCES
[1] ZHAO Wen-chun; MA Wei-mingo FFT Algorithm with High
Accuracy for Harmonic Analysis in motor testing [J]. China
Electrical Engineering, 200 I.
[2] CHEN Jian-ping; LU Jian-xin; LI Yun-hua.Implementation of Fast
Fourier Transforms on Digital Signal Processors [1]. Nantong
institute of Technology,20 I I
[3] E.Linzer and E.Feig. Modified FFT for Fused Multiply/Add
Architectures . IBM Research, Yorktown Heights, NYTech Rap.
RC-I74 I4, I99 I

Fig.6. Voltage spectral analysis [4] E.Linzer and E.Feig. Implementation of Efficient FFT Algorithms on
Fused Multiply/Add Architectures.IEEE Signal Processing,
We can see from Figure 6 that it can accurately measure I993,41(1),pp.93-107
the harmonic components of the voltage and current by FFT [5] R.Meyer and K.Schwarz. FFT Implementation on DSP-Chip-Theory

Analysis function aList (relative to specified base), and Practice . IEEE Proc.lCASSP, I990
[6] C.Liu, J. W.Cooley and R.Tolimieri. Variants of the Winograd
compared the theoretical electric parameters with the Multiplicative FFT Algorithms and Their Implementation on IBM
measured values from the model, as shown in Table I and II. RS/600. IEEE Proc.lCASSP, 1991.
[7] HUANG Tian-shu; TAN Zhi-jiang.Applications of FFT algorithm to
TAB I intelligent motor protection system[1]. COMPUTER ENGINEERING
THE THEORETICAL ELECTRIC PARAMETERS DESIGN,2003.
[8] WANG Feng-wen; SHU Dong-mei; ZHAO Hong-cai. Digital signal
theoretical values processing [M]. Beijing University of Posts and Telecommunications
harmonic press, 2006
Umax lmax P Q [9] Siva Kumar Palaniappan and Tun Zainal Azni Zulkifli. Design of

fin:Iatrtntal 526.65 6.72 1565.()S &9531 I6-point Radix-4 Fast Fourier Transform in 0. I8J.1m CMOS
Technology . American Journal of Applied Sciences, 2007.
3rd hanrrnic 2.19 0.24 �J.043 .Q.041
[10] LlU Mei-rong. DSP Implementation of FFT Algorithm[J].MICR
51h hanrrnic 1.75 0..19 0..m2 .Q.028
-OELECTRONICS &COMPUTER,20I5
71h hanrrnic 152 0..14 .Q.025 .Q.Ol2
[ I I] FU Rong. Design and Physical implementation of Radix-4 FFT
91h hanrrnic 1.19 0.00 .Q.0l2 .Q.OOO Processor[D].Xi'an University of Electronic Science and
Technology,2014
TAB H [12] PANG Hao; LT Dong-xia. An Improved Algorithm for Harmonic
THE MEASURED VALUES OF THE MODEL Analysis of Power System Using FFT Technique [J].Proceedings of
the CSEE, 2003.
measured values of the model
harmonic [13] LI Hai-tao;DENG Ying. MATLAB programming tutorial[M].Higher
Education Press,2010
Umax Imax p Q
fin:Iatrtntal 530.68 6.68 1571.14 90328
3rd hanrrnic 228 021 -0.102 -0.1(1)
5th hanrrnic 1.83 0.13 0.151 0.074
71h hanrrnic 1.0) 0.08 0.085 0.046
91h hanrrnic 122 0.03 -0.016 -0.015
Table I and II show that the error between theoretical