Technical Feasibility Study of an IGBT-based Excitation System

(1)

Juni 2017

Technical Feasibility Study

of an IGBT-based Excitation System

Johan Frisk

(2)

Teknisk- naturvetenskaplig fakultet UTH-enheten

Besöksadress:

Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0

Postadress:

Box 536 751 21 Uppsala

Telefon:

018 – 471 30 03

Telefax:

018 – 471 30 00

Hemsida:

http://www.teknat.uu.se/student

Technical Feasibility Study of an IGBT-based Excitation System

Johan Frisk

This thesis aims to design a cabinet to house some of the required hardware to realize a 1000 A IGBT inverter controlled static excitation system. In the thesis practical design considerations are identified and solved.

The suggested excitation system requires a cabinet to house the inverters. Together with inverter requirements stated by the inverter manufacturer and possible electromagnetic interference from switching of the IGBT:s, practical design considerations arise when realizing the system. Identified design considerations are heat dissipation, EMI, IP-code requirements and mechanical stresses at inverter connections.

In this study, the design considerations are addressed and a cabinet design with required components inside is suggested. The suggested cabinet together with its components could fulfil the suggested system's- and the inverter's requirements. However, the IP-code allowed by the suggested EMC-seals might be lower than the IP54 required by the inverter. The cabinets EMC-properties will probably be lowered if regular rubber gaskets are used.

The study suggests one possible configuration which is possible to realize. It is suggested that further consideration is dedicated to the EMI reducing properties of the cabinet if it is to be installed in an environment sensitive to EMI.

ISSN: 1650-8300, UPTEC ES17 017 Examinator: Petra Jönsson

Ämnesgranskare: Urban Lundin Handledare: Johan Abrahamsson

(3)

Populärvetenskaplig sammanfattning

Magnetiseringssystemet styr strömmen till generatorns fältlindningar och rotorns magnetfält ändras efter strömmen genom fältlindningarna. Magnetfältet från rotorn bestämmer vilken spänning som generatorn ger ifrån sig till elnätet.

För att styra strömmen i rotorn kan tyristorer användas för att likrikta en växelström som sedan leds till generatorns rotor. Storleken på den likriktade spänningen kan ändras genom att kontrollera tyristorerna. Genom att styra spänningen över rotorns fältlindningar styrs indirekt även strömmen, som bestämmer magnetfältet, som bestämmer spänningen som generatorn ger ifrån sig.

Ett annat sätt att styra strömmen till generatorns rotor, är att använda en likriktare tillsammans med en växelriktare, en så kallad H-brygga. Att använda en H-brygga medför att spänningen över rotorns fältlindningar kan anta endast två värden, VDC eller -VDC. Den applicerade spänningen över rotorlindningen kan inte varieras, tillskillnad mot spänningen som appliceras i ett tyristorbaserat system. Detta gör att strömmen styrs direkt i det växelriktarstyrda system, tillskillnad mot i det tyristorbaserade systemet där strömmen styrs indirekt genom att variera spänningen över rotorlindningen. I den förslagna växelriktaren finns transistorer, så kallade IGBT:er (eng. Insulated Gate Bipolar Transistor). Dessa IGBT:er används för att styra strömmen genom rotorns fältlindningar.

Att använda växelriktare är förknippat med värmeutveckling från förluster som uppstår när IGBT:erna slår av och på. Värmeutvecklingen kan, om effekten är hög, bli betydande. Den värme som utvecklas tillsammans med de krav som tillverkaren av den förslagna växelriktaren har på montering av den, skapar design problem.

Det här examensarbetet syftar till att designa ett skåp som kan uppfylla de krav som det föreslagna magnetiseringssystemets ska uppfylla. Även de krav som tillverkaren av den föreslagna växelriktaren har på dess installation ska uppfyllas. Tillvägagångssättet har varit litteraturstudier, 3D-modellering och diskussioner med leverantörer. Att hitta de krav som ställs på växelriktarens installation har varit centralt i litteraturstudien och kraven har legat till grund för de designkrav som ställts på elskåpet.

Under litteraturstudien framkom att växelriktaren måste monteras inuti ett elskåp, vilket gör att värmeförluster från den måste tas om hand. Det framkom också växelriktaren kräver:

 En separering mellan kylflänssida och framsida. Separeringen ska åtminstone vara av kapslingsklass IP54

 Ett luftflöde över sin framsida. Luftflödet ska minimalt vara 2 m/s för att undvika heta områden.

 Att åtgärder vidtas för att minimera mekaniska krafter på anslutningsterminaler för AC och DC.

(4)

I en elektrisk komponents kapslingsklass är den första siffran i IP-koden en indikering på hur beständig komponenten är mot inträngande föremål. I den ena änden av skalan (låg siffra) indikerar siffran hur beständig komponenten är mot inträngande föremål, i den andra änden av skalan står den första siffran för skydd mot damm. Den andra siffran i IP-koden står för skydd mot vatten, där en hög siffra betyder högre motståndsförmåga än en låg siffra.

Växelriktaren kan även orsaka elektromagnetiska störningar (EMI) i sin omgivning på grund av att spänningen bryts och slås på med en frekvens på upp till 15 kHz. För att skapa elektromagnetisk kompatibilitet (EMC) med omgivningen kan elskåp tillverkas i ett särskilt EMC-utförande och skärmade kablar användas. Att jordning utförs på ett korrekt sätt är viktigt för att få EMC.

Genom att skapa 3-D modeller av föreslagna elskåp i SolidWorks och sedan föra diskussioner med skåptillverkare, har en föreslagen lösning tagits fram som uppfyller de krav som växelriktaren ställer på elskåpet. För att minska risken för mekaniska spänningar på växelriktarens kopplingspunkter har diskussioner förts med en tillverkare av strömfördelningskomponenter.

Resultatet visar att det kan vara tekniskt möjligt att bygga ett elskåp som uppfyller de krav som tillverkaren ställer på installation av växelriktaren. Detta kommer att bero på om flänsar tillsammans med tillhörande packningar uppfyller IP54. En leverantör av elskåp hittades som kunde bygga en föreslagen lösning på skåpkonstruktion.

En färdig strömfördelningskomponent som kunde överföra 1000 A och utgöra en övergång mellan kabel och växelriktare hittades inte. En specialtillverkad lösning togs fram genom diskussion med en leverantör av strömfördelningskomponenter. Den framtagna lösningen består av en solid kopparskena till vilken kabelskor för anslutning av kablar samt anslutning av flexibel skena kan ske. Den flexibla skenan (Flexibar) bockas i U-form för att kunna absorbera krafter från termisk expansion av kopparen. Bockningen är avsedd att förhindra mekaniska krafter på växelriktaren.

Luftflödet mot växelriktarens framsida åstadkoms genom att montera en fläkt framför, snett nedanför växelriktaren. Fläktens förväntade livstid samt dess strömförsörjning har varit avgörande vid val av fläkt.

Kablar innehållandes en ledare har valts för överföring av ström till generatorn. Detta val har gjorts på grund av ett högre strömvärde jämfört med kablar innehållande flera ledare (till exempel en trefaskabel).

(5)

Executive summary

In this thesis the technical feasibility of a new concept of static excitation system is studied.

The new concept of excitation system uses an inverter with IGBT:s to control the output voltage of the generator.

The study shows some practical problems that can be encountered when designing a cabinet to house high power inverters (1000 A) and how to meet them. The result shows that it is technically feasible to build the system. Components needed for current distribution and cooling of the inverters are designed and presented.

It is suggested that further consideration is dedicated to the EMI reducing properties of the cabinet if it is to be installed in an environment sensitive to EMI.

(6)

Nomenclature

IGBT - Insulated Gate Bipolar Transistor EMI - Electromagnetic Interference EMC - Electromagnetic Compatibility RFI - Radiofrequency Interference XLPE - Cross-linked Polyethylene

PAC - Programmable Automation Controller

(9)

1

1 Introduction

Excitation of the rotor field winding inside a generator can be achieved in different ways and two examples are the brushless AC- and the static excitation systems [1]. Nowadays, the static excitation system is the preferred system [2] and in the static exciter system, the components used are, as implied by the name, static and non-moving [3]. A static exciter system uses a power rectifier bridge to rectify AC into DC. The power rectifier bridge consists of power electronics, which can be thyristors [4]. The output voltage from the thyristor rectifier can be controlled and used to control the voltage on the output terminals of the generator. [4]

A novel way to obtain a static excitation system is to use a H-bridge inverter with IGBT:s to control the voltage on the output terminals of the generator. If the field winding can withstand the fast and large voltage transient across it, a chopper based control offers better controllability than a thyristor based system [5]. However, the use of inverters are coupled to heat generating switching losses. In this study the heat generated could be significant since the exciter system can attain power levels in the hundreds of kW area. [6]

The heat generated from the inverter will have to be dissipated and this can become troublesome since the chosen inverter manufacturer states that the inverter has to be installed inside a cabinet. The cabinet allows for a certain IP-code separation between the heat sink side and the electronics side of the inverter. Protection against accidental touch of electrically conductive, and accessible parts, with up to 1000 VDC is also obtained by mounting the inverters inside a cabinet. The requirement of cabinet installation from the manufacturer also helps to assure only authorized and qualified personnel can access the inverters [7].

Since the system in the study is dimensioned for currents up to 1000 A the size of the required cables can make them cause mechanical stress to their connection points. The chosen inverters' AC- and DC terminals are sensitive to stresses and measures have to be taken to minimize them [7]. Also, since switching of an inductive load can give rise to EMI [8]

measures will be taken to reduce EMI.

1.1 Purpose

The purpose if this study is to investigate the technical feasibility of an IGBT controlled exciter system.

1.2 Problem formulation

To realize the IGBT controlled static exciter system the necessary hardware needs to be found and made compatible with the inverters.

The inverters must be installed in some sort of casing with the required IP-code to protect the electronics from dust and protect from accidental touching of electrical connections.

The switching losses coupled to the inverters may cause heat dissipation problems as they are mounted inside a casing.

(10)

2

The inverters heat sinks must be cleaned in recommended intervals. Hence the heat sinks must be installed in a way which allow access to them for cleaning.

The switched output voltage from the inverters will cause harmonics and could thereby cause EMI-problems to its surrounding.

The DC-link and cables which connect the inverter to the generator have to be connected to the inverter without causing mechanical stresses at the inverters AC- and DC-terminals. Since the system is dimensioned for 1000 A, the risk of mechanical stress from cables can be

substantial.

1.3 Scope

This study will address problems coupled to housing IGBT based inverters inside a cabinet.

Problems regarding heat dissipation, IP-code separation, electrical connections and EMI will be looked into and technical solutions will be found to meet them.

The hardware required for rectification, control system and software are deemed out of scope and will not be covered in this study.

1.4 Aim & limitations The aims of this study are:

 Perform a literature study to get knowledge of required design criteria

 Create a 3-D drawing of a cabinet containing the required components needed for an IGBT controlled excitation system handling 1000 A

 Create a guide for assembly of the proposed cabinet The study does not include:

 Hardware in control system

 Software

 Electrical connections to the slip rings on the rotor shaft

 Hardware for rectification 1.5 Background

The basic function of any excitation system is to provide direct current to the field winding of the synchronous machine. The excitation system also performs control and protective functions. The protective functions of the exciter system consists of ensuring that the capability limits of the generator are not exceeded. The excitation system can also control the power system by controlling the field voltage, and thereby controlling the field current. These functions are essential to make the power system perform in a satisfactory way. [3]

Excitation of the field windings inside a generator can be obtained in different ways. One of them is the static excitation system. In contrast to DC excitation systems and AC excitation systems, all the components in the static excitation system are still, or static [3]. The maintenance of excitation systems with moving parts to supply the excitation field to the generator is not needed in the static excitation system [9]. Replacement of rotating exciters

(11)

3

and its associated equipment with static excitation systems provides a solution to this problem [4]. However, the static excitation system can introduce problems with harmonics due to thyristor switching. [9]

1.5.1 Static exciter system

The static excitation system consists of three basic components. These components are power control devices such as the power rectifier bridge, a voltage regulator and the power potential transformer. Together these components provide field control to maintain the output voltage of the generator. [4]

Voltage regulator

To obtain an automatic voltage control of the generator, an automatic voltage regulator (AVR) is used [4]. The AVR is used to maintain a steady armature voltage and the voltage to be maintained is set to some predefined limits [10]. To achieve an automatic voltage control, some components have to be used together. The components required are sensing transformers, a firing circuit and an automatic voltage regulator [4].

The sensing transformers provides an insulation and a voltage matching between the automatic voltage regulator and the generator instrument transformer [4].

The firing circuit generates the turn-on pulses to the rectifier bridge if thyristors are used for rectification (see section 1.5.2). By changing the time relationship between the firing pulses, the output from the rectifier bridge is increased or decreased [4].

The automatic voltage regulator rectifies a sample from the output of the generator and compares it to a reference dc voltage. If the voltages deviate to much from one another, a signal is sent to the firing circuit. The firing circuit then alters the firing angle according to the error signal to restore the output voltage of the generator to the set value. [4]

Power transformer

The power potential transformer provides power to the excitation system. The output voltage on the generator terminals is stepped down by the transformer to make it compatible with the requirements of the field windings in the generator [4].

Power rectifier bridge

The power transformer supplies the power rectifier bridge with an AC voltage. The AC voltage is rectified to supply the generator field windings with a DC voltage [4].

(12)

4

A principal sketch of the static exciter system is shown below.

Figure 1.1 A principal sketch of the static exciter system. The power and sensing is taken from the generator output via a transformer and is rectified with thyristors to feed the rotor with a DC current. The DC current is fed to the rotor windings via slip rings on the rotor shaft.

In Figure 1.1 it can be seen how the power needed to excite the rotor is taken from the generator output terminals. It is then sent to the thyristor rectifier bridge, which output voltage is set from the AVR. By comparing a rectified sample from the output voltage of the generator to a reference value, the AVR alters the firing angle to the thyristors. The firing angle of the thyristors determines the output voltage from the rectifier bridge and hence the voltage across the field windings.

1.5.2 Thyristor controlled static exciter system

The thyristor is a device which have the ability to control the start of conduction by delaying it from when it gets forward biased to a desired time. Because of this, the thyristor is named a semi-controlled device [11].

The thyristor is an electrical switch which can be turned on to start conducting when it is forward biased. The turn on is accomplished by sending a current pulse to the gate of the thyristor and in its on-state, the thyristor works as a diode. It is not possible to turn off the thyristor with a current pulse to its gate. Instead, the thyristor turns off when the current through it falls to zero [11].

When using thyristors to rectify an AC-voltage, the DC-voltage output is set by the firing angle, α. The firing angle can be seen as the delay from when the thyristor gets forward biased to when the gate pulse is sent to it. This means a larger firing angle will reduce the output voltage.

(13)

5

How the firing angles affect the output voltage is graphically shown in Figure 1.2.

Figure 1.2 Output voltage from a thyristor rectifier. The average output voltage decreases as the firing angle increases.

The firing angle α, dictates the field voltage, which in turn is proportional to the field current.

As the firing angle is changed, so is the field voltage.

The rotor field current is what determines the excitation of the stator winding, which in turn determines the output voltage of the generator. This means, by applying a certain DC-voltage to the rotor field winding (altering the firing angle α of the thyristors), the rotor field winding current can be controlled and thereby also the output voltage of the generator.

A schematic picture of a thyristor controlled exciter system can be seen in Figure 1.3

Figure 1.3 A principal sketch of the thyristor controlled excitation system. An AC-voltage is rectified by a 6-pulse thyristor bridge. By altering the firing angle of the thyristors, the output DC-voltage can be controlled. The current through the rotor field winding is indirectly controlled by the field voltage. The thyristor firing circuit has been left out in the figure.

The voltage sources in Figure 1.3 is the excitation transformer (see Figure 1.1) used in the thyristor controlled static excitation systems. The voltage of the secondary winding in the

(14)

6

excitation transformer will determine the maximum voltage, the ceiling voltage, which can be applied to the rotor field winding.

1.5.3 IGBT controlled static exciter system

IGBT modules have a history of being widely used in converters for electric-motor drives.

The IGBT modules used in these converters can be seen as reliable components and their reliability are equal to or even greater than that of thyristors. [12]

In the IGBT controlled static excitation system shown in Figure 1.4, the IGBT:s will be supplied with a DC-voltage. This means, the rotor field winding (the RL-load in Figure 1.4) will be subjected to the ceiling voltage as the IGBT:s switches, and not to of a variety of voltages in-between as in the thyristor controlled system.

Figure 1.4 A principal sketch of the IGBT controlled exciter system. An AC voltage is rectified by a 6-pulse diode bridge and the resulting DC-voltage is supplied to the H-bridge constructed with IGBT:s. The rotor field winding is represented by the RL-load in the H-bridge.

The current is controlled with a current sensor placed in one of the legs of the H-bridge. The measured output current from the H-bridge will be compared to a reference value by a Programmable Automation Controller (PAC). The reference current value is determined from the output voltage of the generator. A lower output voltage than required will lead to an increased reference value of the rotor field current in the PAC.

The voltage applied to the rotor field winding will determine how steep the slope of the current ripple will become. A higher voltage will mean a steeper slope and possibly a faster current control. In the IGBT controlled exciter system in the study, the same voltage, the ceiling voltage, will always be the voltage applied to the field winding. A sketch of the current wave form in the rotor field winding can be seen below in Figure 1.5.

(15)

7

Figure 1.5 Sketch of the field current wave form in the rotor winding of the IGBT controlled exciter system. The slopes of the wave form will approximately attain the value or .

The graph in Figure 1.5 appears as it does due to the altering ceiling voltages. As the ceiling voltage attains a positive value, the current is increasing. As the current reaches some tolerance value, the two conducting switches in the H-bridge will open and the other two switches will close (see Figure 1.4). This causes a negative voltage to be applied across the field winding. The applied voltage will be the ceiling voltage but with an opposite sign as before the opening and closing of switches. The now negative voltage will start to limit the current in the field winding until it reaches a tolerance value below the reference value, the switches will then change state again, and the procedure continues.

If the field winding can withstand the fast and large voltage transient across it, a chopper based control offers better controllability than a thyristor based system. This way of field control opens up possibilities. Possibilities as e.g. a segmented field winding where different poles are excited by different choppers. [5]

Because of this possibility, it is interesting to look into if the suggested IGBT controlled static exciter system can be realized. The realization should be made with components, preferably on-shelf than tailor-made, which can be bought and made fit together to allow implementation of the system.

(16)

8

2 Pre-study

Except from laws and regulations, information regarding EMI and EMC was also gathered since the switching together with an inductive load gives rise to transient interference [8].

The pre-study also aimed to acquire knowledge about the inverter specific needs regarding its housing. Specific requirements in mind were the inverters IP-code, guidelines regarding electrical connections and conductor sizing.

2.1 Laws, regulations and recommendations

The technical feasibility study is at this state, a research project and does not intend to provide a ready to sell product on the market. Hence, the rules of electrical equipment from the National Electrical Safety Board will not apply. Because of this, there are no demands on how electrical connections are designed nor any demands on providing user manuals and storing documentation of the product in a certain amount of years. [13]

In "starkströmsföreskrifterna" [14] advice is given about dimensioning conductor cross sectional area. It is written that the smallest cross sectional area of the conductor should be chosen with consideration to:

 Highest allowed temperature of the conductors.

 Acceptable voltage drop.

 Electromagnetic stresses which can arise due to a short circuit.

 Other mechanical stresses which the conductors may be subjected to.

2.2 EMI & radiated emissions

Electromagnetic interference is an increasing form of pollution. It's effect can vary from a disturbing noise on a radio receiver to more severe effects such as potential fatalities due to corruption of safety-critical control systems. As electrical and electronic equipment penetrates more into society, the risk of EMI increases and so does the possible damage. [15]

To dampen the electric- and magnetic fields originating from an electrical circuit, shielding can be used. Shielding means an electrically conducting surface (a barrier) is placed around or around a part of the electrical component. The barrier can be made completely out of metal if protection against low frequency EMI is desired. If protection against higher frequency EMI (30 MHz or more) is desired, a thin conductive layer placed on e.g. plastic is sufficient to achieve a shielding effect. [15]

2.2.1 Cabinet to reduce EMI

To shield against EMI, a cabinet constructed of a conductive material can be used. Principles of absorption and reflection of electric fields are presented in Figure 2.1

(17)

9

Figure 2.1 Description of how an impinging electric field is absorbed and reflected from a barrier surface.[15]

The shielding effectiveness of a barrier is a result of both reflection and absorption. As an AC electric field impinges on a conductive surface, a current in that surface will be induced. The induced current flow (J in Figure 2.1) is attenuated a certain amount each skin depth of the barrier. Skin depth depends on material properties such as conductivity, permeability and frequency. The skin depth is defined as (1). [15]

(1)

As an example, steel can offer higher absorption than copper at low frequencies due to its high relative permeability. The absorption loss is the same whether the field is electric or magnetic. [15]

The reflection loss depends on the wave impedance of the field. In the near field, when the distance from the source to barrier is less than the E-field impedance will be high and it causes the reflection loss to be high. For magnetic fields, the wave impedance in the near field is low, and correspondingly, the reflection loss is low. Barrier material also affects the reflection loss. For materials as copper and aluminium which are high conductive, the reflection loss is higher than for lower conductive materials such as steel. [15]

Chassis radiation

A common source for radiation is the seams of the chassis. Higher frequency emissions, typically greater than 200 MHz, are more common to originate from the chassis of the equipment. [16]

An electromagnetic shield commonly consists of more than one part which is joined together in seams. When two parts are joined together, the joint will not be perfect, which will cause the electrical conductivity in the joint to be non-perfect. The reason for this may be more than one and distortion, painting and corrosion are a few examples of things which can create an

(18)

10

insulating layer which reduces electrical conductivity. As a consequence, problems will arise from an EMC-perspective when removable panels and doors are used. [15]

Circuit boards placed inside a cabinet can generate high-frequency currents on the inside of the chassis and the high-frequency currents can leak out from gaps or seams. These leaked currents will then flow around the outside of the chassis. The chassis then becomes an antenna which radiates. [16]

Seams

When a high frequency current flowing on the inside of the chassis come to a seam it must be able to pass it easily. A small impedance (a few milliohms) will create a voltage drop, which means an electric field, that radiates. [16]

Holes

To keep the EMC-performance of the cabinet if a hole has to be made for e.g. a display, a sub shield can be used. The sub shield have to be made sure to surround the back side of the display and have a good electrical contact with the door/wall in which is placed. [15]

2.2.2 Shielded cables to reduce EMI

Shielding of cables can be obtained in several ways. Foil-shield, braided shield and Helical/spiral shield are just a few examples. The shields differ in design and material properties, and hence also in shielding properties. [8]

Braided shields are effective in both reducing emissions and increasing immunity against EMI. The braided shield is effective when possible sources of interference are e.g. switches which are operating an inductive load or when the source is e.g. motor control circuits. [8]

Braided shields offer protection against magnetically induced interference in the frequency range 30 to 100 MHz. Generally, EMI protection is increased with higher grade of braid coverage and typical values are 80-95 percent coverage. 100 percent shield coverage is unattainable, but if the braid covers 85 percent or more of the conductor, the braided shield can offer significant lowering of the radio-frequency interference (RFI) [8]. Radio-frequencies have traditionally been defined as frequencies ranging from a few kHz to roughly 1 GHz. [17]

A spiral shield can be used to shield against inductive coupling and capacitive coupling when possible sources of interference are for example power lines. [8]

Overall shields are effective against power line frequencies. The overall shield is also

effective against high frequency electromagnetic and electrostatic interference if the shield is grounded at both of its ends. A counter current that cancels out the interfering current

interacting with the protected current, is induced when the shield is grounded at both ends. [8]

To reduce radiated emissions, it is important to assure that all shielded cables have a low impedance bond at both ends. The shields should be terminated with direct contact to the connector or the chassis and the use of pigtails should be avoided unless absolutely necessary.

If pigtails must be used, assure they are made as short as possible. [16]

(19)

11

A metallic shield is a barrier to high-frequency fields on one side of the barrier and the other.

Since it is important that all parts of an enclosing barrier are well bonded together, problems arise when a cable needs to be penetrated through a barrier. To prevent noise current from leaking outside the barrier, following the outside of the cable shield or the cable wires, the cable needs to be bonded to the enclosure. The bonding should be made with a low- impedance connection to the chassis, and ideally made with a 360° connection. [16]

2.3 Inverter requirements

The manufacturer of the proposed inverter provides some requirements for the operation of the inverter. There are certain requirements on the cabinet which are to house the inverters and there are also requirements regarding temperatures and electrical connections on the inverter.

2.3.1 Inverter cabinet

When the inverters are mounted, there must be a separation between the inverters front and the heat sinks on its back. This separation must fulfil at least IP54. [7]

The National Electrical Safety Board explains the first- and second figure in the IP-code as can be seen in Table 2.1.

Table 2.1 Explanation of the figures in the IP-code system. The explanations are made by the Swedish National Electrical Safety Board. Values in the table are adopted from [18].

IP-code

First figure Second figure

0 No protection. No protection

1 Protection against penetration of solid objects larger than 50 mm.

Protection against dripping water.

2 Protection against penetration of solid objects larger than 12 mm.

Protection against dripping water. The apparatus is not allowed to lean more than 15°.

3 Protection against penetration of solid objects larger than 2.5 mm.

Protection against sprinkling water. Maximal angle 60°.

4 Protection against penetration of solid objects larger than 1.25 mm.

Protection against sprinkling water. All angles.

5 Dust protection. Protection against rinsing water from a nozzle.

6 Dust tight. Protection against heavy rinsing with water.

7 Can be temporarily submerged in water without

taking damage.

8 Suitable for long term submersion in water.

According to manufacturer directions.

The IP-code specifies the protection degree of the electrical equipment and the higher figure in the IP-code the higher protection is offered.

2.3.2 Inverter temperatures

The inverter manufacturer specifies maximum temperatures for different components on the inverter. They also specifies a minimum airflow across the snubbers on the front of the inverter. An air flow ≥1 m/s is required, but if the air flow is <2 m/s the maximal operating temperature of the snubbers risks to be exceeded. [7]

(20)

12

A snubber is an electric circuit typically placed in parallel with each semiconductor element.

The snubber circuit is used to provide a protection against voltage- and current transients and often consists of a resistor and a capacitor. [19]

The specified temperatures are shown in Table 2.2.

Table 2.2 Temperature limits at certain components on the inverter specified by Semikron. The table is adopted from [7].

Temperature limits

Component Maximal surface temperature

which is not to be exceeded

Recommended maximal surface temperature during operation

Thin film capacitors 85 °C 80 °C ⁽¹⁾

Snubbers 95 °C 90 °C

AC- and DC-busbars 105 °C 100 °C

(1)Lifetime of the capacitors is reduced if operating between 85 °C and 90 °C for extended periods of time

To verify that the internal air cooling is sufficient, measurements have to be made to ensure temperatures does not exceed the values in Table 2.2. [7]

The placement of the components mentioned in Table 2.2 is shown in Figure 2.2 below.

Figure 2.2 A sketch of Semikron's SlimLine 150. The red and green rectangles show the DC- and AC-terminals respectively.

The yellow rectangle shows the snubbers and the blue rectangle shows where the thin film capacitors are placed. The thin film capacitors cannot be seen in the sketch.

(21)

13 2.3.3 Inverter electrical connections

The electrical connections on the AC-terminals of the inverter are made out of aluminium or tin plated copper. To ensure electro-galvanic compatibility of the materials in the electrical connection to the inverter, tin plated copper should be used. Tin plated copper is compatible with bare copper, bare aluminium, tin plated copper and nickel plated copper. [7]

The manufacturer of the inverter issues precautions to be taken during its installation. The precautions regards the mechanical stresses which the electrical terminals of the inverter might be subjected to. Mechanical stresses may come into existence during connection of cables or busbars to the inverter. [7]

The manufacturer of the inverter also specifies the maximum allowed temperatures for the AC- and DC-terminal connections. The temperature of the AC- and DC-terminal on the inverter is not allowed to exceed a temperature of 100 °C (Table 2.2) [7]. This has to be taken into account when choosing the conductors which are to be connected to the inverter. The chosen conductors ampacities have to be high enough to allow for a continuous load current without exceeding 100 °C.

2.3.4 Inverter heat sink cleaning

If dust accumulation on the heat sink fins starts, the thermal efficiency of the heat sink may be reduced. Recurring stops due to overtemperature may be an indicator of dust accumulation on the heat sink fins and it is advised to add an air filter if dust accumulation becomes a problem.

Maintenance intervals of the inverters in the SlimLine series are shown in Table 2.3.

Table 2.3 Cleaning interval recommendation and replacement interval of some components of the inverters in the SlimLine series. The table have been adopted from the SlimeLine series user manual [7].

Year 0 1 2 3 4 5 6 7 8 9 10 11 1

2

next Cleaning

Heat sink fan X X X X X X X X X X X X

Het sink fins and capacitor bank

According to the environmental conditions the inverter is installed in

Replace Capacitor bank and heat sink fins

X Every 100 000 h

Heat sink fans X Every 70 000 h

To clean the heat sink fans, the fans have to be dismounted from the heat sink. The fans are then to be blown down in a separate area. [7]

If the inverter is mounted with a IP54 separation between its heat sink side and its front side, the inverter does not have to be dismounted to clean the heat sink and capacitors. If it is not mounted with the stated IP-code separation, the inverter have to be removed and transported to a separate area for cleaning. [7]

(22)

14

2.4 Ampacity & principles of cable- and busbar dimensioning

Ampacity is a measure of conductors ability to carry electrical current. All metals will carry electrical current, but copper and aluminium are the most commonly used materials for conductors. Copper is the most widely used because it is a better conductor and is physically more strong than aluminium. The weight of copper is about three times that of aluminium and due it's low density, aluminium is the most common choice for over head power lines.

However, the resistance of aluminium is more than 150 percent higher than that of copper.

[20]

Copper's higher conductivity compared with aluminium reduces the heat losses for a copper conductor compared to an equally sized aluminium conductor. The lower heat losses for a copper conductor infers a higher ampacity than that of an equally sized aluminium conductor.

[21]

The ampacity of a conductor will depend on several factors. The material of the conductor, the area in which it is installed and it's cross sectional area are some of the factors affecting the ampacity [20]. When dimensioning the required cross sectional area needed to provide a certain amount of current to a load, ampacity tables and reduction factors are used.

The required ampacity tables are found in IEC standard 60364-5-52 Low-voltage Electrical Installations Part 5-52: Selection and erection of electrical equipment - Wiring systems [22].

Also the necessary reduction factors, which derates the ampacity of conductors depending on their way of installation, ambient temperature and number of circuits can be found there.

(23)

15 2.4.1 Principles of cable dimensioning

The dimensioning of cables is done with respect to their ampacity, which is their maximum current carrying capacity. The ampacity will differ between two same sized conductors depending on if they are single core cable or multi core cable. A sketch of a single core and a multi core cable is shown in Figure 2.3.

Figure 2.3 Single core and multi core conductors. The single core conductor to the left offers higher ampacity for a given conductor size. The conductors in the picture uses one single copper strand, to increase the cable's flexibility, more strands can be used.

Also, the choice of insulation material of the conductor will affect it's ampacity. A PVC- insulated conductor will have a lower ampacity than a XLPE insulated conductor, since the highest temperature it can withstand is lower. [22]

In order to take into account the ambient temperature and the method of installation of the conductors, reduction factors are used. The ampacities in [22] are calculated at an ambient temperature of 30 °C. This means the ampacity for a given conductor will increase if the ambient temperature is lower, and decrease if it is higher. If conductors are placed together on e.g. a cable ladder with no spacing in-between, their ampacity is reduced. [22]

To calculate the required ampacity of a conductor at a given installation (2) is used [23]:

(2)

where is the rated load current and is the correction factor which changes the required conductor ampacity given the way of installation.

In Table 2.4, the ampacities of two loaded, horizontally placed single-core, PVC-insulated and XLPE insulated copper conductors can be seen. The ampacity of an equal sized multi- core, PVC-insulated cable is also shown for comparison. The ampacities of the conductors are

(24)

16

at 30 °C ambient temperature for a copper conductor with a maximal conductor temperature of 70 °C for a PVC insulated conductor and 90 °C for a XLPE insulated conductor. [22]

Table 2.4 Ampacities of a few chosen conductor sizes. The values in the tables are adopted from table B.52.10 and B.52.12 in IEC 60364-5-52. The ampacities are valid for conductors installed horizontally on a cable ladder.

Single-core cable Multi-core cable

Cross sectional area [mm²]

2 loaded conductors 70 °C

185 463 A 575 A 364 A

240 546 A 679 A 430 A

300 629 A 783 A 497 A

Correction factors for the ambient temperature are shown in Table 2.5.

Table 2.5 Correction factors for different ambient temperatures which are to be used if the ambient temperature at the installation differs from 30 °C. The values in the table are adopted from table B.52.14 in IEC 60364-5-52.

Ambient temperature [°C] Correction factor for PVC-insulated conductor

20 1.12

25 1.06

30 1.0

35 0.94

Correction factors to be used when more than one conductor is used in parallel can be seen in Table 2.6.

Table 2.6 Correction factors which takes number of circuits into consideration. If a group of cables consists of n cables, it may be considered as n/2 circuits of two loaded conductors or n/3 for a multi-core cable with three loaded conductors. The values in the table are adopted from table B.52.17 in IEC 60364-5-52.

Arrangement: cable ladder and cables touching

Number of circuits or multi-core cables Single layer on cable ladder system

7 0.79

8 0.78

9 0.78

The correction factors in Table 2.6 are used to take into account for more than one circuit of two loaded conductors placed on the same cable ladder. The highest correction factor in the table are for a group of 9 cables placed tightly, touching each other on a cable ladder. [22]

Finally, the corrected ampacity can be calculated with (2).

2.4.2 Principles of busbar dimensioning

A busbar is a conductive bar, usually made of copper or aluminium. The busbar enables connection between two or more electrical circuits and it can be used in e.g. substations. [24]

Busbars can be made more flexible by using sheets of copper to create a laminated busbar and the lamination allows for more flexibility during installation. [25]

The dimensioning of these insulated busbars, Flexibars, from Mericon is made from a table in one of their product catalogues. The dimensioning is made from the maximum allowed temperature of the conductor, the ambient temperature, the load current in the conductor and the width of the connection point [26]. The connection points are the AC- and DC-terminals of the inverter.

(25)

17

In Table 2.7, the maximal current which do not cause the conductor to exceed 50 K (ΔT = 50K) of the ambient temperature is shown. The choice of ΔT = 50 K is a result from discussions with Mericon.

Table 2.7 Maximal allowed current to not exceed a conductor temperature which is higher than 50K than that of the ambient.

N in the leftmost column shows the number of copper sheets in the Flexibar and W is the width of the Flexibar. The thickness of the sheets are 1 mm for all the conductors shown in the table. In the rightmost column, the recommended overlap at the point of connection is shown, the recomended overlap is 5xW mm. The values in the table are adopted from the table

"Dimensionering av Flexibar" at page 6 in [26].

Dimension (NxW) [mm] Allowed current [A] (ΔT = 50 K) Overlap [mm]

8x40 1040 A 40

8x50 1175 A 40

6x63 1215 A 30

4x80 1015 A 25

5x80 1175 A 25

The Flexibars can be delivered with a conductor material consisting of bare copper or as tin plated copper [26]. A sketch of the Flexibar with its laminated tin plated copper sheets is shown in Figure 2.4.

Figure 2.4 Sketch of the insulated Flexibar with laminated conductor material of tin plated copper.

2.5 Final cabinet design criterions

The pre-study resulted in a list of design criteria. The design criteria are mostly specified by the inverter manufacturer. The EMC design criteria was added since switching together with an inductive load can give rise to transient interference [8].

The list of design criteria becomes as follows:

 IP54 separation between front and heat sink side of the inverter

 Mechanical stress of AC- and DC-terminals must be lowered as much as possible

 Airflow over the front (≥2 m/s)

 Allow for cleaning of fans and heat sinks

 A cabinet to reduce EMI from the inverters

(26)

18

3 Method

When the design criteria had been determined, 3-dimensional models of cabinets were created in SolidWorks. The models were created in an iterative process where a manufacturer of cabinets was contacted to discuss designs and their ability to fabricate the proposed designs. A reseller of busbars and current distributing components was also contacted for discussions.

Selection of components could be made when the pre-study was finished and it followed the procedure presented in section 3.2.

3.1 Dassault Systemes' SolidWorks

SolidWorks is a tool for computer aided design and it has been used to create 3-D models of cabinets. The modelled cabinets were housing the inverters together with the components needed to power and cool the inverters. An example of such modelled component is the fan mounting structure (Figure 4.9) and fan used to provide forced air cooling on the inverter's front side.

The 3-D modelling could show distances between parts and cabinet walls and how much room was left for e.g. the installation of a transition from cable to busbar for connection to the inverter. Room for bending radii of cables and routing of cables are examples of other limiting parameters which became visible when creating the models.

One of the main issues during the design was to keep an IP54 separation between the heat sinks and the front side of the inverter. The 3-D modelling helped when trying to create paths for the air to flow towards the heat sinks when at the same time keeping the required IP-code.

Different topologies and the possibility of them fulfilling the inverter criteria were evaluated using the 3-D models.

The modelling in SolidWorks was an iterative process which followed the procedure in Figure 3.1.

Figure 3.1 The iterative procedure used when designing the cabinets in SolidWorks.

The flow chart in Figure 3.1 describes the designing procedure. A sketch of an idea of a design was created and contact was taken with the cabinet manufacturer for discussion and to see if it was possible for them to fabricate the design. In the discussions, pictures of the cabinets were used as a foundation for better understanding and reduce risk of misunderstanding when details about the designs were discussed.

(27)

19 3.2 Component selection

The selection of components to realize the IGBT controlled exciter system followed the procedure seen in Figure 3.2.

Figure 3.2 The applied procedure in choosing components to the IGBT controlled static exciter system.

The adopted procedure consisted of the steps seen above in Figure 3.2. Firstly, the system specifications were studied to get knowledge of what components were need. The second step consisted of finding the required component. In a third step, the components specifications were compared to the system requirements, if it fulfilled them, the component selection moved on to a fourth step. In the fourth step, the recently found component were sketched in SolidWorks. The sketch was made to get an in-scale model of the component which could be fitted together with the other components to see if and how they fitted together. If the recently sketched component could be made to fit together with the other components and it had an acceptable price, the component could be bought and used in the technical feasibility study.

To start the study, the inverters were used as a starting-point. Inverters from the SlimLine series from Semikron (SL150) have been used to design a cabinet and the electrical connections needed to connect conductors to the inverter.

3.2.1 Cabinet

The cabinet design started with contacting two companies which manufactures cabinets to house electrical equipment. One of the companies responded and the design work were continued with them. The cabinet needed to fulfil the design criterions in section 2.5 was not among the chosen manufacturer's product range. Hence, a tailor made design had to be produced.

The cabinet design had to fulfil the criteria mentioned in the Pre-study, section 2.5 and different designs of cabinets were produced in an iterative process. A cabinet which could fulfil the requirements were drawn in SolidWorks and then discussed with the manufacturer.

During the discussions it became clear how the production machines at the manufacturer and the materials they used limited how the cabinets could be designed.

Problems to produce suggested designs resulted in a final design which originates from the manufacturers own cabinet design with slight changes to it.

(28)

20 3.2.2 Forced air cooling at inverter front

To keep the cabinet integrity as much as possible, a decision was made to not use an inlet- outlet air cooling system.

The forced air cooling is attained from a fan mounted in front of the inverters. The fan will cause an airflow over the snubber circuits, which is specified by the inverter manufacturer.

Since the excitation system is crucial for power production and the inverters risk to shut down if their maximum temperature is exceeded at some part, a fan with long lifetime is desirable.

The power supply to the fans were decided to be taken from the control cabinet next to the cabinet housing the inverters. In the control cabinet, there is an available power supply of 24 VDC which provides power to the fans pushing air through the inverters heat sinks.

Together with the design criteria in section 2.5 this infers that the criterions for fan selection become:

 24 VDC

 As long lifetime as possible

 Produces an airflow of ≥ 2 m/s

 Mounted in such a way that it forces air over the front side of the inverter 3.2.3 Electrical connections

When the design criterions of the system had been determined, it became obvious some components were needed in order to connect cables connecting the DC-link and generator to the inverter. This was deemed necessary to reduce risk of mechanical tension on connection terminals on the inverter.

To solve the problem, a company selling current distributing products were contacted.

Discussions containing inverter specifications, required currents and desired possibility to connect more than one cable to each output terminal led to a solution.

An electrical connection forming a bridge between cable and inverter is used to prevent mechanical tension from cables and from thermal expansion of conductor material. The bridge is used to fulfil the design criterion regarding mechanical stress in section 2.5.

The same solution can be used to connect to the output terminals (AC-terminal) and the input terminals (DC-terminal).

(29)

21 3.3 Assembly

To reduce the risk of metallic splinters finding their way into the inverters, the suggested order of installation is:

 Place the cabinet on the supportive leg structure, fasten the cabinet to the structure and fasten the supportive leg structure to the floor if needed.

 Install flanges, gaskets cable glands and cable routing inside the cabinet. Make the necessary holes for fastening of the heat sinks protective cage to the supportive leg structure, to each other and to the cabinet.

 Make the required FL21 sized holes for cables to pass between the control- and inverter cabinet.

 Fasten the two cabinets into each other.

 Install the inverters.

 Attach the inverters heat sink fans.

 Install the electrical connections at the AC and DC-terminals of the inverters.

 Install the fans which provides airflow over the inverters fronts inside the cabinet.

 Install the heat sinks protective cage on the cabinets back side when the system is up and running.

More details (e.g. torques) regarding assembly can be found in Fel! Hittar inte

referenskälla.. However, it is advised to read the section regarding installation in [7] before installing the inverters.

(30)

22

4 Result & Analysis

The result from the technical feasibility study of an IGBT controlled excitation system is presented in this section. Firstly, the results regarding cabinets are presented. Results regarding the component selection are presented in section 4.2 and onwards. A bill of materials can be found in the Fel! Hittar inte referenskälla..

4.1 Cabinets

The system is suggested to make use of two cabinets. One cabinet is to contain the inverters with the components required to make the system function and fulfil the design criterions in section 2.5. A second cabinet is used to house the control equipment used in the IGBT controlled exciter system.

4.1.1 Inverter cabinet

The proposed cabinet is a result from using the cabinet manufacturer's original design, without any significant changes to it. In the manufacturer's design, there is a metal sheet mounted close to the back of the cabinet. The function of the metal sheet is to be a surface onto which electrical equipment can be mounted. In the proposed design, this sheet has been removed to give room for the inverter heat sinks to be placed outside the cabinet. The placement of the heat sinks outside of the cabinet is a result of the limitations in fabrication and required IP-code separation between the heat sink side and the front side of the inverter.

The proposed cabinet is shown below in Figure 4.1.

Figure 4.1 The proposed design of the inverter cabinet. The cabinet has pre-fabricated holes for flanges and holes in its back to make room for the inverters' heat sinks. The back of the cabinet is made in two parts due to fabrication limitations.

(31)

23

The cabinet is fabricated in IP-code 55 prior to making the holes in the back to fit the inverters' heat sinks. Also, the cabinet is fabricated in the manufacturers EMC-design to reduce the risk of EMI from the inverters.

The proposed cabinets dimensions and approximate weight are presented in Table 4.1

Table 4.1 The proposed cabinets dimensions and approximate weight.

Cabinet data

Width 160 cm

Height 160 cm

Depth 40 cm

Weight 200 kg⁽¹⁾

(1)The cabinet is tailor made and have not yet been built. Because of this, the weight is approximated with a similar sized cabinet found in the manufacturers range of products.

Pre-fabricated holes for flanges have been tailor fitted to accommodate the outgoing cables from the inverter to the flanges. One flange is used for each inverter. The pre-fabricated holes for flanges in the top of the cabinet are placed to accommodate for electrical connection of the DC-link to the inverter. An extra set of flanges have been fitted to allow for the incoming DC- link to be connected to either side of the row of inverters.

The cabinet is placed on supportive legs to give room for bending the radii of the outgoing cables to the generator. The distance between the floor and the bottom of the cabinet is 50 cm.

Since the heat sinks of the inverters add significant weight to the back of the cabinet, the supportive legs have been designed to compensate for this. The supportive leg structure continues further back to prevent a bending force from the cabinets centre of mass to tip the cabinet. The supportive legs have feet which could be fastened in the floor to even further prevent the cabinet from tipping. The supportive leg structure is shown in Figure 4.2.

Figure 4.2 The supportive leg structure. In the right picture the holes in the structure feet is shown. The holes permits fastening to the floor. Fastening prevents the cabinet from tipping.

Since the supportive leg structure is constructed with a larger depth than the cabinet itself, a metal bar is added to the place where the back edge of the cabinet will be placed. The metal bar permits the cabinets back lower edge to rest and not be suspended hanging in the air. The

(32)

24

supportive leg structure is fabricated in 35x35 mm square pipe steel. The choice of using nine legs is discussed in section 5.

The heat sinks sticking out of the back of the cabinet is a crucial part of the system. Without proper cooling, the switching losses will cause the inverter to get overheated. The protective cages seen in Figure 4.3 were added to prevent damage to the heat sinks and also to prevent accidental touching of the warm heat sinks.

Figure 4.3 The inverter cabinet with the protective cage. The protective cage gives some protection against mechanical damage and also offers protection against accidental contact with the heat sinks.

The protective cage is made out of 2 mm thick cold-rolled steel sheets and there are two mirrored halves creating one cage. The two halves is a result from limitations in fabrication.

The two parts of the cage are fastened to the supportive leg structure with screws which allows for easy dismounting. The easy dismounting helps fulfilling the design criterion regarding cleaning of fans and heat sinks in section 2.5.

The perforation of the cage seen in Figure 4.3 is mainly used to reduce weight.

4.1.2 Control equipment cabinet

The cabinet housing the control equipment is placed next to the inverter cabinet. This placement of the cabinets close to each other ease the connecting of cables for measurements on the inverter. The placement also eases the connection of power supply to the fans in front of the inverters as well as connecting the power supply to the heat sink fans. The control equipment cabinet placed to the right of the inverter cabinet is shown in Figure 4.4

(33)

25

Figure 4.4 The control equipment cabinet and the inverter cabinet. The control equipment cabinet placed to the right of the inverter cabinet.

The suggested control cabinet's dimensions and weight are presented in Table 4.2

Table 4.2 The proposed control cabinet's dimensions and approximate weight.

Control cabinet data

Width 50 cm

Height 200 cm

Depth 50 cm

Weight 82 kg⁽¹⁾

(1)The cabinet proposed cabinet is fabricated in an EMC-design and the weight found is valid for a non-EMC cabinet. Slight difference in weight may occur.

The control cabinet is fitted with holes for flanges to allow for cables passing through the control cabinet into the inverter cabinet. However, the inverter cabinet is not pre-fitted with these holes for flanges. This is due to the precise alignment of the cabinets which is needed if

(34)

26

the holes are to be pre-fitted cannot be guaranteed on site. Hence the holes in the inverter cabinet have to be made on site.

The flanges on top of the control cabinet is a dividable type of flange. This eases the passing of measurements cables into the cabinet and reduces the need of cable glands.

The control cabinet is placed on a socket which panels can be removed to allow cables to enter from the bottom of the cabinet. The removable panels also permits cables from the inverter to be laid underneath the control cabinet on their way to the generator if needed.

In Figure 4.4 a hole can be seen in the door of the control cabinet. The hole will hold a screen to show system parameters of interest.

4.2 Components in inverter cabinet

To make the IGBT based excitation system work, more components than the inverters are needed. Components to allow for electrical connections, flanges and cable glands as well as fans to cool the inverters' fronts are some of the components required.

4.2.1 Electrical connections

The electrical connections in the system can be divided into three parts. The first part is the flexible busbar connecting the inverter to a solid busbar mounted on insulators some distance from the AC- and DC connection point of the inverter. The solid busbar forms a bridge between the inverter and the cables. The second part is the bridge itself and the third part is the connection of cables to the bridge.

The connection points are summarized below:

 The connection of cables to the bridge with cable shoes

 The bridge

 The flexible busbar connecting the inverter to the bridge 4.2.1.1 Cables

The recommendations regarding voltage drop, electromechanical stresses and other

mechanical stresses in section 2.1 have been disregarded. The voltage drop will not cause any issues since the system uses current control to obtain its desired function. Furthermore, the recommendations regarding stresses have not been considered since the inverter will measure and cut the current fast if a fault occurs. No cause of other mechanical stresses have been foreseen.

In the process of choosing cables, two main types of cables have been considered. The two types of cables are the single core cable and the multi core cable. The single core cable consist of one single conductor (could have more than one strand) and the multi core cable which is made with three or more conductors, e.g. three phases and a neutral.

The chosen cable is of a single core type. The single core cable have a higher ampacity for a given conductor size and a comparison between the ampacities of single- and multi core cables can be seen in Table 2.4.

Technical Feasibility Study of an IGBT-based Excitation System

Juni 2017

Technical Feasibility Study

of an IGBT-based Excitation System

Johan Frisk

Technical Feasibility Study of an IGBT-based Excitation System

Populärvetenskaplig sammanfattning

Executive summary

Table of Contents

Nomenclature

1 Introduction

2 Pre-study

3 Method

4 Result & Analysis