2017 IEEE Calcutta Conference (CALCON)

The Impacts of Distributed Generation on Voltage

Stability
Debjani Bhattacharya Soma Biswas
Dept. of Electrical Engineering Dept. of Electrical Engineering
Academy of Technology, Adisaptagram, Hooghly-712121 Academy of Technology, Adisaptagram, Hooghly-712121
e-mail:[email protected] e-mail: [email protected]

Abstract²Power system is forced to operate near to its [2].Minimum singular value of Newton Raphson power flow
stability limit due to the rapid growth of power demand. So, Jacobian matrix [3] is another method to predict voltage
voltage stability has become a primary concern for power stability. In modern power system scenario distributed
systems researchers.The imbalance between reactive power generation (DG) plays a vital role. To meet the increasing
generation and consumption in the system causes voltage drop in
the system which is the main cause of voltage collapse. In order to
power demand DGs are very much useful. DGs can be placed
avoid this, most of the modern power systems are adopting the in proper place with proper size to increase the reliability of
new emerging technologies like Flexible AC Transmission the system by improving the voltage profile which in turn
Systems (FACTS) devices and Distributed Generation (DG) improves the voltage stability margin [4]. DGs are classified
systems etc. Conventional method of repetitive load flow analysis on the basis their capability of injecting real and reactive
for stability study LV KLJKO\ WLPH FRQVXPLQJ $ 7KHYHQLQ¶V power in the system.DG units based on synchronous machine
equivalent network based voltage stability index is proposed in for small hydro, geothermal, and combined cycles are capable
this paper which avoids the continuous load flow analysis. The of generating both active power (P) and reactive power (Q).
index is used to identify the weakest bus in the IEEE 30 bus test
On the other hand fuel cells, Photovoltaic (PV) cells, micro
system. The improvements of the proposed index with the
addition of facts devices and DG technologies are visualized. Also
turbines which are integrated to the main grid with the help of
the optimal location and size are investigated for which the converters/inverters are capable of generating active power (P)
stability improves to a maximum value. only. The DG units equipped with synchronous compensator
are capable of generating reactive power (Q) only. Fixed
Keywords—Distributed Generation; FACTS; Load Flow; speed squirrel cage induction generator (SQIG) used for wind
Radial Distribution System; Thevnin's equvalent network; Voltage turbine generating (WTG) system can generate active power
Stability. (P) but consumes reactive power (Q). Thus, it is quite obvious
that the type of DG technology used will have a significant
I. INTRODUCTION
effect on the performance of power system. Therefore, it is
Load demand in power system is increasing day by day. required to check the effect of different types of DG
Hence the stress on power system which is a huge technologies on the voltage stability to get the maximum
interconnected network also increases. An important task of reliability. The FACTS (Flexible Alternating Current
modern power system control is to keep the voltage level Transmission System) devices are a family of power
within permissible limit. The regular expanding nature of electronics-based devices those are capable of enhancing AC
power system load leads the power distribution system to face system controllability and stability and also increase the power
voltage instability. The consequence of this voltage instability handling capability.
is voltage collapse. So voltage stability problem is a
challenging area of power system researchers. Voltage In this paper area of voltage stability region (AVSR) [5] is
stability analysis is useful in power system operation and used as a voltage stability index to find the weakest bus and
planning, in real time monitoring of power distribution then FACTS devices and different DG technologies are
system. Voltage stability is the ability of power system to applied to improve the voltage stability limit of that weakest
maintain acceptable steady voltages limit at all buses in the load bus. Rest of the section is organized as follows. Section II
system at normal condition and also after being subjected to a describes the methodology used for the proposed work. In
disturbance. Several methods are currently used for measuring section III the proposed method has been implemented and the
voltage stability margin and to determine the weakest bus. results obtained are tabulated. Finally the section IV discuss
Most of these methods are index based. Various stability the conclusions.
indexes are available to measure the degree of weakness of a II. METHODOLOGY
bus or a line. Most widely used technique is load flow
technique for stability analysis [1].Another commonly used The proposed work has described how the voltage stability
method is modal analysis method which shows the sensitivity can be improved with the help of distributed generation. In this
of reactive power with respect to voltage magnitude

978-1-5386-3745-6/17/$31.00 ©2017 IEEE 105

 2017 IEEE Calcutta Conference (CALCON)

Fig. 1. 7KHYHQLQ¶VHTXLYDOHQWFLUFXLW.

paper voltage stability analysis is performed with the help of

AVSR (area of voltage stability region) method using
7KHYHQLQ¶VHTXLYDOHQWFLUFXLWWRLGHQWLI\WKH ZHDNHVWEXV7KH
proposed work has been tested on IEEE 30 Bus test system.
A. Thevenin’s Equivalent circuit Fig. 2. A typical P-Q curve.

Any power system can be represented with the help of the AVSR is defined as the area under this P-Q curve which
network shown in Fig.1. +HUH ( LV WKH 7KHYQLQ¶V equivalent can be obtained by integrating equation (7) between PL = 0 to
voltage (VTh) and ZTh (=RTh+jXThLVWKH7KHYHQLQ¶VHTXLYDOHQW the maximum possible value i.e. the loadability limit.
impedance.
To find ZTh, ZBUSmatrix is constructed. The diagonal
௉ ௉
element (Zkk) of ZBUS UHSUHVHQWV WKH 7KHYHQLQ¶V HTXLYDOHQW ‫ ܴܸܵܣ‬ൌ  ‫׬‬଴ ೘ೌೣ ܳሺܲሻ݀ܲ ൌ  ‫׬‬଴ ೘ೌೣሺܽܲଶ ൅ ܾܲ ൅ ‫ܥ‬ሻ݀ܲ (8)
impedance of the system corresponding to the kth bus. To find
VTh load flow analysis is run with all the loads ignored. If VL C. Identification of weakest bus
be the voltage at L bus then VThcan be computed using AVSR can be used as a positive indication to determine the
equation (1). degree of weakness of a load bus. If the loading of the bus
VTh = (1+ZTh/ZL)VL (1) changes then the area also changes. If AVSR value is high the
voltage stability margin is high and so the degree of weakness
B. Area of voltage stability region (AVSR) of the bus is low. Using this AVSR index the status of the bus
Load voltage VL and load currents IL from Fig. 1 can be from voltage stability point of view can be obtained.
expressed with equation (2) and (3). D. Modeling of FACTS devices and distributed generation
IL = (PL± jQL)/VL* (2) (DG)units
VL = E ± ZThIL (3) With the application of FACTS devices and distributed
generation the bus real and reactive power injection changes
Using equation (2) and (3) the load voltage VL can be which in turn changes the ThevenLQ¶VYROWDJH9Th. The FACTs
obtained as [5], GHYLFHV DUH PRGHOHG DV WKH UHDFWLYH SRZHU VRXUFHV DQG '*¶V
are modeled as real as well as reactive power sources. If QGi is
ܸ௅ସ ൅ ʹሺ்ܴ௛ ܲ௅ ൅  ்ܺ௛ ܳ௅ ሻܸ௅ଶ െ  ‫ ܧ‬ଶ ܸ௅ଶ ൅ ሺܴଶ்௛ ൅
the reactive power generation, QDi is the reactive power
ଶ ሻሺܲଶ
்ܺ௛ ଶ
௅ ൅ ܳ௅ ሻ ൌ Ͳ     demand and Qfi is the reactive power generated by FACTS
Or,ܸ௅ଶ ൌ devices of ith bus then injected power of this bus will be Qi
మ ା௑ మ ൯൫௉ మ ାொ మ ൯ which is equal to QGi+Qfi± QDi. If PGi is the active power
ିሾሼଶሺோ೅೓ ௉ಽା௑೅೓ ொಽ ሻିா మ ሽേටሼଶሺோ೅೓ ௉ಽ ା௑೅೓ ொಽ ሻିா మ ሽమ ିସ൫ோ೅೓

೅೓ ಽ ಽ
 generation, PDi is the active power demand and PDGi is the
ଶ active power generated by DG units of ithbus then injected
At the verge of voltage collapse its discriminant becomes power of this bus will be Pi which is equal to PGi+PDGi± PDi.
zero and hence the following equation can be obtained: The solar photo voltaic cell may be considered as the source of
‫ ܧ‬ସ ൅ Ͷሾʹܲ௅ ܳ௅ ்ܴ௛ ்ܺ௛ െ  ‫ ܧ‬ଶ ሺ்ܴ௛ ܲ௅ ൅  ்ܺ௛ ܳ௅ ሻ െ real power only and diesel engine, wind generators are
ଶ ଶ
்ܴ௛ ܳ௅ଶ െ  ்ܺ௛ ܲ௅ଶ ൌ Ͳ (5) assumed to be capable of generating both real as well as
reactive power.
Or, ݂ሺܲ௅ ǡ ܳ௅ ǡ ‫ܧ‬ǡ ்ܴ௛ ǡ ்ܺ௛ ሻ ൌ Ͳ (6)
III. IMPLEMENTATION AND RESULT
By varying PL from zero to a maximum value with some
steps, QL¶V DUH REWDLQHG XVLQJ  NHHSLQJ ( DQG =Thconstant. A. Identification of the weakest bus and Improvement of
The variation of QL with respect to PL can be plotted on a voltage stability
graph. A typical such P-Q curve is shown in Fig. 2. The proposed technique is implemented on an IEEE 30 bus
Applying the basic curve fitting technique a second order test system. The voltage stability analysis is performed for four
polynomial is obtained as given in (7). cases: Case1: the base case without any external device; Case
2: with FACTS devices; Case 3: with DG (generating only real
ܳሺܲሻ ൌ ܽܲଶ ൅ ܾܲ ൅ ܿ (7) power) Case4: with DG (generating real and reactive power).

106
 2017 IEEE Calcutta Conference (CALCON)

x Case 1 (Base Case): For all the load bus the VTh and ZThare 0.9889± 0.4007 +
±1.3686 P2±
obtained and the coefficient of 2nd order polynomial as 27 0.2416P + 0.204673
0.2753i 0.3620i
0.5762
using curve fitting technique and the AVSR for different ±1.2364 P2±
load buses have been shown in Table I. From this table it is 28
0.9897 ± 0.3369 +
0.3185P + 0.401765
observed that bus 26 has lowest AVSR which mean bus 26 0.2058i 0.1651i
0.9290
is the weakest bus. 0.9854 ± 0.5642 +
±1.9677P2±
29 0.1834P + 0.0826219
0.2863i 0.6410i
TABLE I. VTH, ZTH, 2 ORDER POLYNOMIAL AND AVSR FOR THE BASE
ND 0.3470
CASE ±1.9105P2±
1.0422 ± 0.5983 +
30 0.1726P + 0.0801544
Bus 0.2615i 0.7420i
VTh(p.u.) ZTh(p.u.) Q(P) AVSR 0.3357
No.
1.0232 ± 0.3406 + ±1.2749P2±
3
0.1356i 0.1676i 0.276P + 0.9535
0.457036 x Case 2(with FACTS): It is well known that use of facts
±1.3409P2± devices improves the bus voltages. As bus 26 is the
1.0283 ± 0.3352 +
4
0.1625i 0.1458i
0.2462P + 0.52014 weakest bus so this bus is chosen as the location of
1.0353 FACT devices. Here FACTS device is modeled as a
±1.4251 P2± source of reactive power. Using FACTS the 2nd order
0.9932 ± 0.3245 +
6 0.2386P+ 0.519196
0.1946i 0.1278i
1.0482
polynomial and AVSR have been shown in Table II.
±1.2551 P2 ± x Case 3(with DG generating real power only): Solar PV
1.0780 ± 0.3720 +
7 0.2821P + 0.536052
0.2497i 0.1636i
1.0518
cell etc can generate real power only. So we consider a
±1.2293 P2± PV cell to be placed at bus 26. Here also Q(P) and
1.0189 ± 0.3421 + AVSR are computed and shown in Table II.
9 0.2335P + 0.424402
0.2566i 0.2099i
0.8843
±1.3007 P2± TABLE II. 2ND ORDER POLYNOMIAL AND AVSR FOR DIFFERENT
1.0304 ± 0.3574 +
10 0.2322P+ 0.457368 CASES
0.2862i 0.1893i
0.9414
±1.2527 P2± Size of
1.0828 ± 0.3935 + Type of applied
12 0.2422P+ 0.437952 applied Q(P) AVSR
0.2938i 0.2242i unit
0.9101 unit
±1.4471 P2± ±2.4760P2±0.1668P
1.0408 ± 0.4598 + Base Case - 0.0597325
14 0.2627P+ 0.234961 + 0.2978
0.2812i 0.3242i ±2.2076P2±
0.6444 FACTS devices 5 MVAR 0.0687343
±1.2967 P2± 0.1883P+0.3193
1.0394 ± 0.3932 + DG generating
15 0.2656P + 0.361577 ±2.3270P2±
0.2861i 0.2354i active power (P) 5 MW 0.0644691
0.8219 0.1791P+0.3094
±1.3295 P2± only
1.0264 ± 0.4049 + DG generating
16 0.2630P + 0.288894 5 MW, 5 ±2.1275P2±0.1873P+
0.2828i 0.2829i both active and 0.0737075
0.7173 MVAR 0.3303
±1.2785 P2± reactive power
1.0478 ± 0.3893 +
17 0.2436P + 0.374723
0.2990i 0.2426i
0.8294
±1.3780 P2±
1.0037 ± 0.4272 +
18 0.2902P + 0.23027
0.2915i 0.3157i
0.6365
±1.3233 P2±
1.0381± 0.4351 +
19 0.2761P+ 0.250044
0.2946i 0.3268i
0.6588
±1.2529 P2±
0.9986 ± 0.4119 +
20 0.3263P+ 0.248229
0.2926i 0.2969i
0.6656
1.0806 ± 0.3998 + ±1.2872 P2±
21 0.435805
0.3208i 0.2246i 0.2314P+0.9095
±1.2750 P2±
0.9948 ± 0.3968 +
22 0.3154P+ 0.27548
0.2866i 0.2683i
0.7079
±1.3434 P2±
1.0035 ± 0.3731 +
23 0.2595P+ 0.362839
0.2915i 0.2148i Fig. 3. Comparison of AVSR for different cases.
0.8295
±1.3897 P2±
1.0356 ± 0.4237 +
24
0.3095i 0.2835i
0.2393P+ 0.291121 x Case 4 (with DG generatingbothrealand reactive
0.7218 power): Wind generator, diesel plant etc can be used
-1.4629 P2±
25
0.9804 ± 0.4400 +
0.2645P+ 0.175933
here. Here also the AVSR has been computed and
0.2821i 0.3739i
0.5399 tabulated in table II.
±2.4760P2± The variation of AVSR with for different cases is shown
1.0050 ± 0.7040 +
26 0.1668P + 0.0597325
0.2860i 0.7642i
0.2978 graphically in Fig. 3.

107
 2017 IEEE Calcutta Conference (CALCON)

B. Optimal location and size of DG(generating real power

only):
To obtain the optimal location of DG, the DG having same
size is placed at other buses also and the AVSR found is
tabulated and shown in table III. From this table it is clear that
when the DG is placed at bus26 then the AVSR of bus 26
(weakest bus) is maximum. So bus 26 is the optimum location
of the DG.
To find the optimal size, DG size is varied from a minimum
value to a maximum value with some steps of step size 10MW.
The AVSR obtained in this case have been shown in table IV.
A graph is also drawn in Fig. 4 which shows the variation of
AVSR with DG size. It is observed that as DG size increases
the AVSR increases first and it becomes maximum (0.104648)
when DG size is 80 MW. Further increase in size of DG the
Fig. 4. Variation of AVSR with DG size.
AVSR decreases. Hence it can be concluded that the optimum
size of DG is 80MW when the system becomes stable to its
IV. CONCLUSION
highest point.
In this work, the AVSR has been highlighted as a voltage
TABLE III. VARIATION OF AVSR WITH LOCATION OF DG stability index. This AVSR index has been used to identify the
Bus no. where
weakest bus in the IEEE 30 bus test system. The AVSR of the
DG is placed
AVSR ORDGEXVHV DUH REWDLQHG XVLQJ7KHYHQLQ¶V HTXLYDOHQW QHWZRUN
3 0.0597708 This approach avoides repetitive load flow for determining the
4 0.0597708 P-Q curves. Impact of FACTS devices and DG technologies on
6 0.0598463 AVSR index are tabulated and it has been seen that in all these
7 0.0598066 cases the AVSR value increases from the base case where no
9 0.0598463 FACTS device and DG are present. Finally, the optimum size
10 0.0598463
and optimal location of DG unit has been obtained where the
12 0.0596454
14 0.0597673
stability is maximum.
15 0.0598066
16 0.0597708 REFERENCES
17 0.0598463
18 0.0598463 [1] 0 )R]GDU ³,PSURYHPHQW RI 9ROWDJH VWDELOLW\ DQG (IIHFWV RI
19 0.0598498 /RDGDELOLW\´,(((8QLYHUVLWLHV3RZHU(QJLQHHULQJ&RQIHUHQFH(UPEC-
20 0.0598498 06), vol. 2, (2006), pp. 486 - 490.
21 0.0599218 [2] Lof PA, Smecl T, Anderson G, Hill DJ. Fast calculation of a voltage
22 0.0599998 stabilityindex. IEEE Trans Power Syst 1992;7(February):54±64.
23 0.0599218 [3] Gao B, Morison GK, Kundur P. Voltage stability evaluation using
24 0.0603211 modalanalysis. IEEE Trans Power Syst 1992;7(4):1529±42.
25 0.0612366
[4] F. B. Araujo and R. B. Prada, "Distributed generation: Voltage stability
26 0.0644691 analysis," 2013 IEEE Grenoble Conference, Grenoble, 2013, pp. 1-4.doi:
27 0.0603984 10.1109/PTC.2013.6652097
28 0.0599218
[5] Priti Sharma, Ashwani Kumar. 7KHYHQLQ¶V HTXLYDOHQW based P±Q±V
29 0.0604855 voltage stability region visualization and enhancement with FACTS and
30 0.0605984 . International Journal of Electrical Power & Energy Systems, Volume
80, September 2016, Pages 119-127.
TABLE IV. VARIATION OF AVSR WITH SIZE OF DG [6] Kundur P, Paserba J, Ajjarapu V, Anderson G, Bose A, Canizares C, et
al.Definition and classification of power system stability. IEEE Trans
PDGI(in
Q(P) AVSR Power Syst2004;19:1387±401.
MW)
10 2
±2.1927P ± 0.1905P+ 0.3202 0.0690991 [7] Nagao T. Voltage collapse at load ends of power systems. Proc IEE
20 ±2.1441P2± 0.1762P+ 0.3382 0.0769833 Jpn1975;95(4):62±70.
30 ±1.9854P2± 0.1907P+ 0.3542 0.0844894 [8] Kundur P. Power system stability and control. New York: McGraw-Hill;
40 ±1.8936P2± 0.1932P+ 0.3672 0.091033 1994.
50 ±1.8386P2± 0.1912P+ 0.3774 0.0964003
60 ±1.7698P2± 0.1989P+ 0.3858 0.100652
70 ±1.7849P2± 0.1855P+ 0.3903 0.103426
80 ±1.7584P2± 0.1904P+ 0.3928 0.104648
90 ±1.7731P2± 0.1877P+ 0.3914 0.103963
100 ±1.7668P2± 0.1995P+ 0.3860 0.100748
110 ±1.8510P2± 0.1953P+ 0.3732 0.094159
120 ±2.1546P2± 0.1912P+ 0.3245 0.0710505

108