747

9. Dynamic stability of industrial networks

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

748

9. DYNAMIC STABILITY OF INDUSTRIAL NETWORKS

Because electrical energy is difficult to store, there must be a continuous balance between production and consumption. Indeed, the only stored energies in an electrical network are the kinetic energy of rotating machines and the energy of capacitors: these have storage capacities of several seconds. The generators, loads and electrical networks that connect them up have mechanical and electrical inertias which make it difficult to maintain a balance guaranteeing relatively constant frequency and voltage. Normally, when confronted with a power variation, the electrical system recovers its stability after several oscillations. In some cases, the oscillating operating conditions may diverge. It is necessary to carry out studies to determine the means to be implemented in order to avoid this phenomenon and guarantee stability of the electrical network. Studies are especially necessary in the case of industrial networks which have one or more generator sets and motors. This chapter will allow us to understand why instability may occur, what are the most frequent causes and what are the induced effects. It gives the precautions to be taken, shows how a study is carried out and provides an example.

n network stability This is characterised by variations in the active and reactive powers which flow through the network and is measured by variations in time of voltages, currents and frequencies associated with these powers. Several definitions characterise network stability:
o steady-state stability

The network has a stable behaviour; this means that when it is subject to small disturbances, it returns to its initial operating point with possible damped oscillations up to the return to a balanced state.
o transient-state stability

When there is a change from a static stable state to a different state, following a lengthy disturbance, whether intentional or not, this change in balance is accompanied by a damped oscillating variable state considered as being acceptable with respect to pre-defined voltage, frequency and duration intervals ( U , f , duration below a maximum value).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

749

o transient-state instability

This can be seen when, following a considerable disturbance, the oscillating state is divergent. It induces a loss of power supply.

o dynamic-state stability

The network is capable of avoiding any divergent oscillating state and returning to an acceptable stable state. This includes possible intervention of the protections and automatic devices. Dynamic stability studies consist in: - imagining the main critical scenarios such as a short circuit, loss of generator sets, loss of utility power supply, considerable load variation, etc. - predicting the behaviour of the network when faced with these disturbances - deciding on the means to be implemented such as types of protection and settings, load shedding/restoration, forbidden configurations, etc., which allow instability to be avoided These studies thus allow us to understand the behaviour of the network and determine the means of reducing the risks of loss of power supply.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

750

9.1.

Behaviour of an industrial electrical network

The behaviour of an electrical network on occurrence of transient phenomena depends on the behaviour of each of its elements. Starting from a stable state, these elements will influence the transient behaviour of the entire network. At the end of disturbance, they will either be in the same stable state as before, or in a different stable state, or in an unstable state, which generally leads to the loss of one or several elements through the activation of protections. It is important, therefore, to know the behaviour of each element in order to determine the behaviour of the entire electrical network.

n passive loads These are loads such as lighting, heating, capacitors, etc. having the following electrical variation laws: V P = Pn Vn V Q = Qn Vn
2 2

(heating, lighting, etc.)

(capacitors)

n power electronic assemblies More and more loads of this type are to be found in industrial networks (see 3.4.7). They are generally highly sensitive to voltage variations. For example, a varying speed motor can be de-energized by a protection for a voltage variation of roughly - 15 %.

n transformers and cables Transformers and cables which ensure the flow of electrical energy between sources and loads are characterised by their impedances. These create voltage drops and Joule losses which depend on the current flowing through them. The values of these impedances are very influential in transient operating conditions: - high inrush currents cause voltage drops which may be critical - high reactances between sources may induce long duration oscillations.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

751

n asynchronous machines Owing to their behaviour and their majority presence in industrial networks (up to 80 % of consumed power in some installations), asynchronous motors play a preponderant role in stability phenomena.

o influence of voltage dips

Figure 9-1 shows torque curves in relation to the speed of a double cage asynchronous motor feeding a pump (see 3.3.1).
T Tn

1 torque at 0.8

O Vn

B
torque at 0.7

Vn
A

0.2 load torque

0.1

0.5 0.6

0.8

N Ns

N Ns T Tn

: speed in relation to motor synchronism speed : torque in relation to motor nominal torque

Figure 9-1: torque curves in relation to the speed of a double cage asynchronous motor feeding a pump

The operating point ( O ) is located at the intersection of the motor torque and load torque curves. The motor torque is proportional to the square of the voltage (see 3.3.1 and fig. 3-15). Motor stability depends on the relative positions of the motor torque and load torque curves. If the motor suffers an interruption or a high voltage dip, it will slow down and operate at a reduced speed, e.g. 60% of the synchronism speed. The reduction in speed is all the greater the smaller the inertia. Its ability to reaccelerate and return to its original stable state depends on the voltage value when it is recovered.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

752

Let us assume that, owing to the current inrushes in the network, the voltage is equal to 0.7 Vn at that moment. The motor torque is only just above the load torque (point B in relation to point A, see fig. 9-1). The motor will "creep" (accelerate very slowly), and be disconnected from the network through the activation of excessive start-up time, rotor locking or undervoltage protections (see 7.9 and 7.12 of the Protection guide). Figure 3-14 shows that once the motor has slowed down a little, it absorbs a high current. This current causes voltage drops which makes reacceleration all the more difficult. If all the motors of an industrial installation slow down (e.g. after a high voltage dip in the public distribution network), the current absorbed by all the motors on reacceleration creates voltage drops which may make reacceleration impossible. The solution is often to use an automatic fast-acting load-shedding and progressive load restoration device. Stability may thus be improved by reducing current inrushes and hence voltage drops. To sum up, asynchronous motors are important elements in dynamic stability; they may encounter operating difficulties following the sudden passage of a reduced voltage supply.

o influence of undervoltages

The absence of supply voltage does not immediately eliminate voltage at the motor terminals. Indeed, the flux stored in the motor cannot be extinguished instantaneously. The rotating field created by the rotor then induces a "remanent" voltage in the stator the amplitude of which decreases exponentially (time constant equal to several tenths of a second). The frequency of this voltage decreases with the rotation speed. If, when voltage reappears on the network, it is in phase opposition with a remanent voltage having an amplitude which has decreased very little, a high overcurrent is then generated which may reach twice the motor starting peak, i.e. 12 to 15 times its nominal current. This may have considerable consequences on the motor: - additional heating and electrodynamic stress in the windings possibly causing insulation breakdowns - torque jolt possibly leading to unacceptable mechanical stress (especially on couplings). The means of providing against this risk is to install a remanent undervoltage protection which forbids supply restoration if the remanent voltage is above a threshold generally set at a value close to 20% of Vn (see 7.13. of the Protection guide).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

753

n synchronous machines These play a predominant role in network stability phenomena. Paragraph 4.2 explains the operation of synchronous machines and their steady-state stability. The following are the main equations which govern their operation: E = V + Xd I P= 3EV sin Xd

: internal electromotive force (e.m.f.); its value varies proportionally to the direct current that flows through the rotor (without saturation)

V : voltage at the machine terminals X d : synchronous reactance I P

: current circulating at the machine terminals : machine active power : angular variation equal to the phase angle of E in relation to V or the mechanical angle between the rotor shaft E and the synchronous reference made up of the voltage V at the machine terminals

The fundamental rule of generator steady-state stability, i.e. its ability to meet a slow load variation, should be pointed out here: operation is only stable if the angular variation remains below 90 .

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

754

o dynamic stability of synchronous machines connected to an infinite power network

Dynamic stability problems come from the machine changing from a stable state to a different state. Let us consider the case of a machine connected to a public distribution network which is subject to a surge of mechanical power supplied by the turbine. The turbine suddenly goes from a supplied power P to a supplied power P2 (see fig. 9-2). 1 P E P2 P1 B C D

3 90

Figure 9-2: displacement of the generator operating point following an increase in mechanical power

The slow increase in power from P to P2 would slowly cause a displacement from point A 1 to point C while remaining on the curve. But the sudden application of this power increase does not directly lead to point C. It is indeed impossible, taking into account the mechanical inertias, to suddenly go from an angular variation 1 to an angular variation of 2 . Thus the machine instantaneously goes from point A to point B, then the angle increases from 1 to 2 . But on arriving at point C stabilisation is not immediate. The mechanical inertia of the machine brings its operating point up to D . From here, deceleration down to point C is finally stabilised, possibly after several oscillations. The calculations carried out using the mechanical and electrical energies of the machine show that the position of point E is defined by the law of areas; areas ABC and CDE are equal. Indeed, we can show that area ABC is proportional to the mechanical energy stored when the machine moves from point B to point C and area CDE when it moves from point D to point C . Consequently, the maximum angular variation max may be above 90 for a transient period. The dynamic stability limit is thus higher than the static stability.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

755

Nevertheless, it is possible that the difference between P and P2 is so great that there is no 1 point D allowing the law of areas to be complied with (see fig. 9-3). The generator accelerates from point B to point C , then up to point X . At this point, it continues to accelerate remaining on the curve and the power transmitted to the network decreases becoming negative (motor operation). The angle continuously increases and there is a loss of synchronism through overspeed.
P P2 B C X

P 1

90

180

Figure 9-3: instability following a sudden increase in mechanical power

Following a similar logic, we can arrive at the following two conclusions: - the risks of loss of dynamic stability are related to sudden and considerable changes in network state, mechanical power or exciting current supplied to the generators - the risks of loss of dynamic stability are all the greater the closer the power supplied by the machine is to the static stability limit. The notion of steady-state stability is evaluated through the expression of the synchronizing torque: Ts = Tel P 3 EV = cos Xd (see 4.2.1.1 - Static stability)

Tel : electromagnetic load torque Indeed, when is low, cos is high, as well as Tel ; in this case, a small variation of

causes a large torque variation and thus an energetic return to the balanced state. Reciprocally, when is close to , a variation of causes a small torque variation and thus a much less 2 energetic return to the balanced state. Thus, a high value of Ts (low ) encourages stability and thus maintenance of synchronism. On the other hand, a low value of Ts ( close to ) renders 2 the machine unstable when varies which causes loss of synchronism.
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

756

In practice, industrial generators are generally made up so that the angular variation is almost equal to 70 for the nominal active power and zero reactive power, so that a margin of stability is kept in the event of a transient disturbance. However, if the generator absorbs reactive power, comes close to 90 and the risk of instability is greater on occurrence of a disturbance. Furthermore, generator speed and voltage regulators play an essential role in the improvement of network stability when disturbances occur.

n regulations The purpose of regulations is to allow correct operation during variations of load or driving power: - voltage and frequency stability for an independent network - active and reactive power stability of generators when they are connected to a public distribution network Let us take the simple case of a generator, the only voltage supply source of an independent network, which is fitted with a speed regulator. The network frequency, proportional to the generator rotation speed, is set by the primary speed regulation of the mechanical drive; the mechanical power is adapted to the power to be supplied so that the required frequency is maintained. Thus, automatic regulation is defined by its droop which expresses the frequency variation in relation to the power (see fig. 9-4).
f
action of secondary regulation droop line at P Pn (secondary regulation) primary regulation droop line

f0

2%

f0

P f

Pn f0

f0

2%
droop line at P = 0 (secondary regulation)

P 0

Pn 2

Pn

Figure 9-4: generator droop line (primary regulation) and action of secondary regulation
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

757

Any increase in the active power supplied causes a drop in frequency and vice versa. Thus, for example, a droop of 4 % guarantees a frequency varying from 49 to 51 Hz (50 Hz x 4% = 2 Hz) when the power supplied varies from 0 to Pn . To overcome this variation, it is possible to introduce compensation that shifts the droop line prallel fashion in relation to the speed through a secondary regulation (see fig. 9-4). Thus, the frequency is maintained at f 0 whatever the power supplied. For example, points A ( P = Pn )

and B ( P = 0) in figure 9-4.

In dynamic operating conditions, the time constants of the system (machine + regulators) are several hundred ms to several seconds. A PID type regulator enables the unavoidable consequences of this relative slowness to be partially overcome. When two generators are connected up, the assembly operating point depends on their respective droop and power (see fig. 9-5). Any power variation is accompanied by a frequency variation and the power sharing between the generators depends on their respective droop. It is thus possible to imagine multiple operating configurations.
f

f0

2%

generator 1 droop line generator 2 droop line

f0 f f0 2%

B' B M' M C' C

P2n 2

P2 P1n 2

P2n P1

P1n

The characteristic points are:

( f 0 + 2 % , P = 0 , P2 = 0) 1

M , M ' ( f , P , P2 ) 1 C , C ' ( f 0 2 % , P n , P2n ) 1

P P B , B' f 0 , 1n , 2 n 2 2

Figure 9-5: operating points of two connected-up generators in relation to their droop line

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

758

The case of a generator connected to a public distribution network continues on from the previous case for which the network has a practically zero droop. The network frequency is applied to the generator and its regulation is thus carried out by the secondary regulation which modifies the power in relation to the set-point (see fig. 9-6). f
action of secondary regulation (regulation of P)

public distribution network droop line

f net

Pn

P=0

Pn

fnet : public distribution network frequency

The characteristic points are:

( fnet , P = 0) ( fnet , Pn )

P L fnet , n 2

Figure 9-6: power regulation of a generator connected to a public distribution network

To summarise, the action of the generator drive machine regulator allows the network frequency (independent operation) or the active power (operation connected to the network) to be modified.

If we use the same logic with the generator exciting current regulator, we obtain the following results: - when the generator is alone (independent operation), the regulator modifies the voltage magnitude - when the generator is connected to the public distribution network, the regulator modifies the reactive power of the generator.
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

759

n public distribution network The frequency and voltage magnitude characteristics in normal operating conditions are contractually guaranteed by the utility (see EN 50 160 table 4-1). The utility gives the short-circuit power at the take-over point (generally three values: high, low and average, which depend on the configuration of its network, see 4.1). Public distribution network faults are reflected on the customer site; their characteristics and occurrence depend on the specific characteristics of the network (overhead or underground, earthing system, harsh environment, etc.). The utility protection system gives standard power supply interruption times.

n the protection, control and monitoring system During transient phenomena, the protections act when one or several electrical variables exceed their pre-defined thresholds for a duration exceeding their time delays (see Protection guide). Automatic load-shedding/restoration or source changeover systems act to save the power supply to priority loads or to help the network to return to a stable state. The protections and automatic systems thus have a considerable influence on network stability.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

760

9.2.

Industrial network dynamic stability study

The purpose of this paragraph is to supply a general description of the objectives of studies and what they involve and to give information about the causes, effects and remedies of dynamic instability. A study carried out by the Schneider "Electrical Engineering and Systems department" is given as an example. 9.2.1. Study objectives

Dynamic stability studies consist in determining the variations in time of electrical variables at different points of a network and the development of the mechanical parameters of rotating machines, following heavy disturbances. The aim of these studies is to find: - the network operating conditions ensuring good continuity of supply to loads - the maximum power that can be backed up during a disturbance - optimum protection setting values which encourage stability and avoid spurious tripping during non-critical disturbances - load-shedding/restoration programmes during a disturbance which allow continuity of supply to priority loads to be ensured (see 12.2.3.3.) - the best machine regulator settings.

Each study is a specific case depending on: the type of sources the types of loads the network structure the network operating mode the causes of instability taken into account.

There are various reasons for carrying out dynamic stability studies: - preventive study during the design of the network - addition of generators and/or high power loads on an existing network - curative study following an incident. When the study is carried out before the installation is built, different factors determining stability can be changed. Thus, once the network behaviour during transient operating conditions is fully understood, its operation in relation to pre-defined constraints can be optimised.
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

761

Studies can be global or limited to a precise problem. For example, in the case of a generator connected to the public distribution network, it is useful to determine the maximum power that the utility can supply during normal operation so that in the case of disconnection the generator correctly feeds the priority loads, 9.2.2. Causes of instability

n electrical phenomena The disturbance phenomena which affects a network's stability are those which cause variations in active and/or reactive power.

o source disturbances

- voltage dips and drops - short and long supply interruptions - frequency variation (independent network).

o network load variation

For example: - at nominal load, the generators have a low synchronizing torque as the value of the angular variation comes close to 90 - at no-load a network can become capacitive. The generators then absorb reactive power, which means that the value of the angular variation tends to come close to 90.

o electrical faults

The most problematic is the three-phase short circuit as during the presence of the fault: - the power supplied to loads is zero since U = 0 - generator acceleration is maximum.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

762

n structure and network operating mode Numerous parameters influence the stability: - the way machines are connected up and connected to the public distribution network - the location and power of priority and non priority busbars - the generator operating point which determines the synchronizing torque Ts - the synchronous machine regulation mode; regulations are carried out on the speed or active power, on the voltage or reactive power - the impedances of cables or lines (e.g. of parallel transformers) - the types of protection and their settings, the connection-disconnection logic, the loadshedding/restoration programmes - the relative characteristics of the motor torque and load torque curves (see fig. 3-17 and table 3-3) - the inertia of the rotating machines.

n industrial process operation In back pressure turbo-generator operation, part of the steam is used for the industrial process and the rest is used to produce electrical energy. Thus, variations in the steam requirement of the industrial process cause variations in mechanical power supplied by the turbine. These variations can cause unstable electrical network operating states due to the resulting power fluctuations and oscillations. Fluctuating loads, such as an arc furnace, welding machine, electron torches, etc., cause power variations which can make the network unstable. 9.2.3. The effects of instability

n on rotating machines During transient periods, the exchanges of power between machines and between machines and network give torque jolts; the resulting mechanical stress can cause mechanical damage (shafts snapped). Used to supply power beyond their capacity, the generators are subjected to frequency and voltage drops. Their voltage and speed regulations may be in resonance with a disturbance and amplify the effects of instability.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

763

The motors slow down as they undergo frequency oscillations and voltage drops. The instant the disturbance is cleared, the current absorbed and induced voltage drops are high and motor reacceleration is thus difficult. Some motors creep or even stall with abnormal overheating and the network finds it more difficult to return to stable operation unless large loads are rapidly shed.

n on the network The circulation of high currents causes damage due to equipment overheating (transformers, cables, etc.). Voltage drops cause some sensitive devices (contactors, electronics, etc.) to operate abnormally. Putting one or several generators out of service destroys the consumption-production balance and may cause the total collapse of the network. 9.2.4. Handling instability

There are different provisions to prevent the stability limit from being overstepped. They are applied to generators, the network and loads and their purpose is either to avoid instability or efficiently deal with it as soon as it starts . n applied to generators The use of high mechanical inertia generator sets reduces the influence of load variations. But this solution is more and more difficult to implement as mechanical inertia is being increasingly reduced. The choice of regulators and regulator settings determine machine response times. They must be set with respect to possible disturbances. The choice of generator operating point is important; a small angular variation and thus a high synchronizing torque improves stability. To meet this requirement, the generator must: - supply reactive power; it is thus preferable to put the compensation capacitors out of service when the generators are operating independent of the network - have a mechanical power and exciting current reserve margin. Installing protections against active power reversals on the generators (see 7.19 of the Protection guide) allows motor operation on drive machine breakdown to be avoided. Network stability is thus improved and the drive machine is not destroyed.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

764

n applied to the network Reducing cable and line impedances encourages the return to a stable state after an incident. Source redundancy and the possibility of shedding non-priority loads reduces the duration and depth of voltage dips once the cause of disturbance has been removed. Load-shedding/restoration per increment of power avoids large disturbances. The fast and selective clearance of a short circuit limits its consequences on the network. The protection system must be designed with the various instability scenarios in mind. The use of logic or differential selectivity (see 9.3 and 7.6 of the Protection guide) instead of time-graded selectivity reduces the short-circuit clearance time and thus prevents large disturbances. Separate phase tripping to clear single-phase faults in transmission networks and the use of shunt circuit-breakers for MV distribution networks have beneficial effects on industrial network stability. Installing a protection against active power reversals at the utility take-over point allows the industrial network to be disconnected when a disturbance occurs on the public distribution network and thus avoids a possible cause of instability.

n applied to loads The use of electronic motor power supply systems (see 3.3.4) reduces the transient currents of the motors during voltage or load torque variations. Installing protections against active power reversals (see 7.19 of the Protection guide) and undervoltage protections on large motors (see 7.12 of the Protection guide) improves stability.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

765

9.3.

Stability studies

The dynamic stability of a network is the ability of the network to return to normal operation following a heavy disturbance. A stability study thus consists in analysing the electrical and mechanical behaviour of the machines between the instant the disturbance occurs and the moment when, once the disturbance has been cleared, the network returns or does not return to normal operating conditions. There are three aspects to the problem:

n electrical Conventional network equations (Kirchoff's law) are used for this. The machines are represented by Park equations which allow transient states to be studied. During the transient period, the subtransient reactances

(X )
" d

' and transient reactances X d

( )

of the machines must be taken in order to carry out the calculations (see 4.1.2. of the Protection guide); these are used for any dynamic phenomena.

n the dynamics of variations around a balanced state This uses the speed and exciting regulator transfer functions.

n mechanical For this aspect, it is necessary to know whether the speed of the machines is maintained or not using the mechanical equations of each machine: J d = Tm Tl dt

J : moment of inertia of the generator and its drive machine or of the motor and its load Tm : motor torque (mechanical for a generator, electrical for a motor) Tl : load torque (electrical for a generator, mechanical for a motor)

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

766

9.3.1.

Calculating methods used

n analytical method (manual) In the case of simple networks, i.e. for networks which only have one machine (possibly two) and passive loads, the analytical description of the machine's parameters in case of disturbances can be carried out without too much difficulty. This analysis is possible in the case where the speed can be considered as being constant. The equations of the machines allow their behaviour to be described with sufficient accuracy even if some parameters are neglected. The different analysis methods (Behn-Eschenburg, Potier's diagram, Blondel's diagram) can be used to determine the efficiency, exciting current and voltage drops of generators and motors. Applied to machines, the Park transform can be used to analyse the steady state as well as transient states.

n simulation on a small-scale network In more complex cases, designers for a long time used small-scale network simulation methods; this enables the behaviour of the machines to be reproduced on a smaller scale (laws of similitude). Although it produces good results, this method requires considerable resources and especially a lot of preparation time to build the small-scale network representing the installation to be studied. Furthermore, this method, which has practically been abandoned today, only applied to networks having a stable structure.

n numerical simulation This method is currently the one which is universally employed. A computer is used to numerically resolve the systems of equations which describe the behaviour of the network and machines. The increasing capacity of today's micro-computers allows large networks to be simulated in a reasonable period of time and detailed analyses of the behaviour of machines and network elements to be carried out. All loads and all generators contribute to the operation of the entire assembly and act on each other; the scale of the problem is thus very broad and, to remain within an area compatible with the capacity of a micro-computer, it is advisable to simplify data so as only to represent several dozen machines: - by grouping together passive loads - by grouping together motors in the form of "equivalent motors" having identical behaviour - by grouping together generators in the form of "equivalent generators" having identical behaviour - by assimilating a very powerful source with a perfect source in series with an impedance (case of utility power supply).
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

767

These calculation preliminaries are obviously very important since they define hypotheses which must be reasonably complex and representative of reality. The resolution method adopted is a step by step method taking into account: - slow varying variables such as the motor torque, rotor speed, field magnet flux and exciting voltage (time constants > 100 ms) - fast varying variables such as current and voltage in the different network branches and different machine circuits, the voltage at the machine terminals and the power supplied (time constants < 50 ms). Software enabling all industrial network cases to be dealt with, such as the Schneider MGSTAB calculation program, is used to implement this method. 9.3.2. Development of a study

A stability study is carried out according to a certain logic and is broken down into several stages which have been summarised below.

n calculation preliminaries The accuracy of results is directly related to the exactitude of the network data; the study begins by the data being collected taking the exact numerical values of the network element characteristics. If they are not available, standard values are used. Next, the mathematical model consists in quantitatively describing the physical laws which govern the operation of the network elements and their cables, in the form of a data file. The calculation of the initial load flow is determined by the specific program which processes the data file: - voltages at node points, currents and powers in branches, sources and loads - operating point of machines.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

768

n simulations The structure and elements of a network vary from one study to another; there are numerous types of disturbance and their point of application is variable. After looking at the arrangement studied, the specialist will select disturbances and their point of application depending on the critical level of the problem. In general, the main disturbances studied are public distribution network supply interruptions, short circuits, loss of network elements (cables, transformers, generators, etc.), starting of large motors and variations in generator mechanical power following the occurrence of phenomena related to the operation of the industrial process (see 9.2.2 - back pressure turbo-generator operation). The calculations of the dynamic state with respect to possible disturbances reproduce the network behaviour and allow the actions to be taken to be determined. The different scenarios are studied so that all the chosen cases can be dealt with and sensitivity to variations on data values be determined.

n results These are mainly shown by time/development curves: - voltages on different busbars, currents in cables and powers carried - machine data (speed, electrical and mechanical torques, excitation) - excitation and mechanical drive regulations.

To conclude, they determine operation of the electrical network on occurrence of a disturbance and allow: - the stability to be checked - the back-up capacity after a disturbance to be known - the protection system to be validated - the regulations to be set.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

769

9.4.

example of a study

The case presented below is taken from an actual study of a typical industrial network in heavy industry; the impact of a three-phase short circuit on the secondary of a 63/20 kV transformer has been studied.

n network description The network comprises (see fig. 9-7): - a 63 kV power supply from the public distribution network. - two 63/20 kV transformers feeding the factory's 20 kV busbar. - an internal production source made up of two generators, which can be connected up, feeding the 20 kV busbar via two 3.2/20 kV step-up transformers. - asynchronous motors with a 5.5 kV supply via 20/5.5 kV transformers connected to the 20 kV busbar. Some of these motors are equivalent machines. - an equivalent passive load representing all the other factory loads. The transformer incoming feeders are fitted with directional overcurrent and earth fault protection (see 7.3 and 7.4 of the Protection guide).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

770

Figure 9-7: network description

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

771

n objective of the study

o disturbance studied

A solid three-phase short circuit on the secondary of one of the 63/20 kV transformers.

o event to be prevented

The fault must not lead to the loss of the 5.5 MVA motors.

o question to be solved

What is the maximum allowable fault clearance time avoiding dynamic instability?

n qualitative description of phenomena

o on occurrence of the fault

The voltage at the short-circuit point is zero as well as on the 20 kV busbar (the line/cable impedances are negligible). The electrical power supplied by the generator sets goes from the initial value to a very low value due to the losses in the step-up transformers; this considerable drop in active power supplied results in acceleration of the generators as they are still driven by the turbines, the mechanical regulations of which do not react instantaneously. At the same time, the voltage regulator controls the increase in exciting current up to its maximum value to try to correct the reduction in voltage. The motors supply the short circuit until their flux is extinguished and then the absence of motor torque, due to the very low voltage, leads to their slowing down. The public distribution network supplies a current which depends on its short-circuit power and the equivalent impedance of the two parallel transformers.

o fault clearance

The directional overcurrent protection is activated and causes the circuit-breakers upstream and downstream of the faulty transformer to be tripped.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

772

o following fault clearance

The voltage reappears on the 20 kV busbar; its value depends on the combined action of the public distribution network, the generators being excited to their maximum limit and the load current inrush. The generators are no longer in phase with each other, or with the network (indeed each source has changed independently of the others since the voltage at their terminals was practically zero) and their speeds are different. They supply a low power since the mechanical power of the turbines has been decreased by the regulators. Indeed, these have seen a drop in active power demand but they have a delayed reaction; the generators will thus slow down. The motors have slowed down, the rotating field of the rotor is phase shifted in relation to the rotating field of the stator produced by the network, and their speeds are different. This induces considerable voltage drops in the cables since all the motors try to reaccelerate at the same time. Oscillating exchanges of energy then occur between the various machines through the network cables and transformers. If the generator speed variations which are at the origin of these transient phenomena decrease, the network returns to its normal operating state. If this is not the case, the generators do not stabilise and pull out of synchronism and the asynchronous motors stall or creep. We can thus see that the study of the behaviour of this network requires a complex calculation to determine whether it is able or unable to return to a stable operating state and for the electrical and mechanical variables to be known.

n quantitative study After having calculated the steady state for a period of 0.1 seconds (we can thus be sure that the model is valid), the short circuit on the 63/20 kV transformer secondary is simulated then cleared by the simultaneous opening of the circuit-breakers upstream and downstream of the transformer; the calculation is then continued for a period of 5 seconds, which is long enough to analyse the network behaviour. The network behaviour is determined for two hypothetical protection activation times: 300 and 350 ms. These two values are close to the acceptable limit aimed at.
Note: in order to simplify the example, we only take into account the protections upstream and downstream of the two 63/20 kV transformers.

We shall now examine the result of the simulation concerning one of the 12.5 MVA generators (they are both identical), and one of the 5.5 MVA motors.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

773

o the generator

active power analysis (see fig. 9-8) The moment the fault occurs, the active power supplied by the generator greatly decreases and remains at a very low value throughout the duration of the fault. When the fault is cleared, an active power oscillation occurs which corresponds to the exchanges between this generator, the other generator and the public distribution network. This exchange of power corresponds to the power required for synchronism to be recovered between the generator voltage and the distribution network voltage. If the protections are activated after 300 ms (fault cleared by the circuit-breaker 40 ms later), the power oscillations (alternately positive and negative values) rapidly decrease becoming stabilised at the initial value (see fig. 9-8-a). On the other hand, in the case of activation after 350 ms, the oscillations continue without presenting a significant decrease; the generator is not capable of recovering synchronism (see fig. 9-8-b).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

774

P (MW) 25 20 15 10 5 0 -5 -10 -15 -20 -25 1 2 3

a) time delay = 300 ms
fault clearance occurrence of fault

t (seconds)

P (MW) 25 20 15 10 5 0 -5 -10 -15 -20 -25 1 2 3

b) time delay = 350 ms
fault clearance occurrence of fault

t (seconds)

Figure 9-8: active power of one of the 12.5 MVA generators

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

775

reactive power analysis (see fig. 9-9) When the fault occurs, the reactive power greatly increases and remains at a high value throughout the duration of the fault, roughly 2.7 times the value before the fault occurred. It increases again when the fault is cleared owing to the return of the voltage to a normal value. Indeed, the loads then have a great need of reactive power in order to recover their normal magnetizing flux. If the protections are activated after 350 ms, the loss of synchronism prevents the reactive power from stabilising (see fig. 9-9-b).
Q (Mvar) 20 16 12 8 4
fault clearance

0 -4 -8

occurrence of fault

t (seconds)

a) time delay = 300 ms

Q (Mvar) 20 16 12 8 4 0 -4 -8 1 2 3 4 5 t (seconds)
b) time delay = 350 ms
occurrence of fault fault clearance

Figure 9-9: reactive power of one of the 12.5 MVA generators

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

776

speed analysis (see fig. 9-10) When the short circuit occurs, the speed increases as the network active power demand is very low while the turbines continue to supply mechanical power. Clearance of the fault leads to the generator slowing down and its speed begins to oscillate. If the protections are activated after 350 ms (see fig. 9-10-b), the generator does not recover a stable operating state.

N Ns
1.039
fault clearance

1.017 1 0.984

occurrence of fault

0.962

0.940 1 2 3
a) time delay = 300 ms

t (seconds)

N Ns
1.039
fault clearance

1.017 1 0.984

occurrence of fault

0.962

0.940 1 2 3
b) time delay = 350 ms

t (seconds)

Figure 9-10: speed of one of the 12.5 MVA generators

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

777

voltage analysis (see fig. 9-11) If the protections are activated after 300 ms (see fig. 9-11-a), the voltage quickly recovers its nominal value after clearance of the fault. On the other hand, the voltage does not recover its nominal value and even tends to decrease if the protections are activated after 350 ms (see fig. 9-11-b).
U (V) 3500

2800

2100

1400

700

occurrence of fault

fault clearance

0 1 2 3
a) time delay = 300 ms

t (seconds)

U (V) 3500

2800
fault clearance

2100

1400

700

occurrence of fault

0 1 2 3
b) time delay = 350 ms

t (seconds)

Figure 9-11: voltage of one of the 12.5 MVA generators

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

778

o current development (see fig. 9-12)

If the protections are activated after 300 ms, the current, just like the voltage, recovers its initial value (see fig. 9-12-a); however, it remains at a high average value if the protections are activated after 350 ms (see fig. 9-12-b). In the case of activation after 350 ms, the protections ensuring the safeguard of the generator must cause it to be put out of service. This does not, however, ensure correct operation of the installation.
I (A) 9000

7200
occurrence of fault

fault clearance

5400

3600

1800

0 1 2 3
a) time delay = 300 ms

t (seconds)

I (A) 9000

7200

5400
occurrence of fault

3600
fault clearance

1800

0 1 2 3
b) time delay = 350 ms

t (seconds)

Figure 9-12: current of one of the 12.5 MVA generators

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

779

o behaviour of a motor

The behaviour of the motors is also representative of the instability observed when the protection time delay is too long. When it is equal to 350 ms, in spite of the fault being cleared, the speed of the motor continues to decrease (see fig. 9-13-b), and the current absorbed is maintained at an average value close to 2 I n (see fig. 9-14-b). This operation is critical for the motor (overheating of the windings) and may be dangerous for the machine driven. It is essential that the protections put the motor out of service.

N Ns 1

0.998
occurrence of fault fault clearance

0.976

0.964

0.952

0.940 1 2 3
a) time delay = 300 ms

t (seconds)

N Ns 1

0.96
occurrence of fault

0.92

0.88

fault clearance

0.84

0.8 1 2 3
b) time delay = 350 ms

t (seconds)

Figure 9-13: speed of one of the 5.5 MVA motors

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

780

I (A) 2500
occurrence of fault

2000

1500

fault clearance

1000

500

0 1 2 3
a) time delay = 300 ms

t (seconds)

I (A) 2500

occurrence of fault

2000

1500
fault clearance

1000

500

0 1 2 3
b) time delay = 350 ms

t (seconds)

Figure 9-14: current of one of the 5.5 MVA motors

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

781

n conclusions of the study The study of the impact of a three-phase short circuit on the 63/20 kV transformer secondary shows that: - setting the transformer protection time delay at 350 ms is unacceptable - 300 ms is the maximum limit - 250 ms allows a security margin.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE

782

CHAPTER 9 BIBLIOGRAPHY

n standards
o

EN 50160 (05.1995): voltage characteristics of electricity supplied by public distribution systems

n Schneider cahiers techniques

o

MV public distribution networks throughout the world, Cahier Technique n 155,

Ch. Puret

Automatic changeover switching on L.V. and H.V. network supplies,

Cahier Technique n 75, G. Thomasset

o o o

Control, monitoring and protection of HV motors, Cahier Technique n 165, JY. Blanc HV industrial network design, Cahier Technique n 169, G. Thomasset Protection of industrial and commercial MV networks, Cahier Technique n 174, A. Sastr

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation crite de nos services. The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide

T&D

6 883 427/AE