The corporate technical magazine of the ABB Group in South Africa and the sub-Saharan Africa region

2/2005

ABB Technology

SOLUTIONS

4Power transformers are designed to withstand mechanical forces that arise from the shipping or transportation, as well as subsequent in-service events. Possible damage may lead to core and winding movement. At ABB Powertech Trans-formers the Sweep Frequency Response Method is used to detect mechanical damage.

Frequency Response Analy-sis of a Power Transformer

7The on-line planning of pulp & paper production is a precise and sophisticated process. To achieve a product of the desired composition, texture and quality, the right mix of chemicals must be applied under the right conditions. Even slight deviations can adversely affect the output. ABB’s pulp and paper online production optimiser helps meet these objectives. Material fl ows between sub-processes are simulated and forecasts computed from the present state. Mill behaviour can be optimised and operating decisions supported by a powerful tool.

P3: Pulp Production Plan-ning

is published by ABB South Africa.www.abb.com/zainfo@za.abb.comAddress:ABB Park, 3 Eglin Road, SunninghillPostal:Pvt Bag X37, Sunninghill, 2157

Telephone:+27 11 236-7000Facsimile:+ 27 11 236-7001

TECHNOLOGY SOLUTIONS

Managing Editor:Chesney Bradshawchesney.bradshaw@za.abb.com

Editorial Panel:

Richard Beckerrichard.becker@za.abb.comNtaga Mojapelontaga.mojapelo@za.abb.comShivani Chetramshivani.chetram@za.abb.com

Jacqueline Burnjacqueline.burn@za.abb.comMark Chan Yanmark.chanyan@za.abb.comAndrew Williamsonandrew.williamson@za.abb.com

This publication was designed, compiled and produced on behalf of ABB South Africa by M&M Publications (Pty) Ltd.PO Box 1644, Saxonwold 2132Johannesburg, South Africa.

Whilst the compilation and production of ABB Technology Solutions is done with great care and attention and every effort is made to prevent mistakes, neither

ABB in Southern Africa nor its principles or subsidiaries, nor M&M Publications (Pty) Ltd. accept responsibility for any errors or the consequences thereof.

2/20052

IN THIS ISSUE

12ABB was awarded the condition monitoring contract at the Eskom Kendal Power Station in May 2003. ABB is responsible for ensuring avail-ability of plant and reducing maintenance costs through using condition monitoring tools such as vibration and oil analysis on the turbine auxiliary plant, boiler, coal, and ash plants.

To measure is to know

15Interoperability according to IEC 61850 means the capability of two or more intelligent elec-tronic devices (IEDs) from one, or several vendors to exchange information and to use it in the performance of their functions and for correct co-operation.

Protection Modelling

19ABB is the technology leader in two crucial areas of industry: electrical power and automation technologies. In this new section of ABB Tech-nology Solutions we look at some of the latest developments at ABB.

Projects and Products

Delegation from Poland visits mining operations in South Africa.

32/2005

NEWS BRIEFS & EDITORIAL

Jackets are on their way to the fi rst three lucky readers who completed and returned the survey carried in ABB TS 1/2005. They are:

• Chris Hockaday, Principal Engineer, Mintek• Patrick Taylor, Special Projects Manager, Petro SA• Michael Mabija, Electrical Manager, Nampak Corrugated Cape

Readership Survey Winners:

Key Assets – Products, PeopleOne of the key focus areas for ABB in South Africa is to deliver world class engineered solu-tions to its customers, by integrating our products into “systems”. ABB achieves this goal at a remarkably high level of quality by being a technology-driven company with two class-leading assets – its products and its people.

The product portfolio is especially noteworthy for its range. We cover almost every electrical component used by utility and industrial customers, from extra high voltage switchgear and trans-formers, right down to panel lights and wiring terminals; from advanced process optimisation software to industrial robots. This range of products is complimented by comprehensive applica-tion know-how in most areas required by our utility and industrial customers.

To facilitate this delivery of engineered solutions, we have a dedicated department that focuses on providing these “systems” in each of the core divisions of Power Technologies and Automation Technologies.

Over time ABB has built a very versatile group of engineering and project management resources, both in terms of the range of application areas; as well as the aspects of the projects in which we become involved. These ranges are comprehensive:

• The application areas include - the manufacturing of switchgear, machines and transformers; power station control sys-tems; building of substations; substation automation; power system communications; electrical network control; robot-ics; pulp and paper mills; metal processing; mining and minerals processing; cement production and the construction of petrochemical plants.

• Our projects involvement can include: consulting; sales; tendering; design; engineering; programming; manufacturing; testing; installation; commissioning; training and service.

Therefore, to win business and then sustain it across this breadth of activity, requires the South African operation and our engineers who can achieve world-leading performances, on every project we deliver. To help us achieve this goal, we have developed an “engineering career path guide” which succeeds in attracting and retaining top people in the fi elds of automation and power technology systems.

This guide doesn’t only help to recognise good engineers, it focuses on keeping them doing excellent engineering work for as long as possible. It is driven by an objective mechanism of recognition and reward and helps to ensure that we deploy these vital resources in the best possible way in the whole organisation.

The guide provides a complete career path starting with junior and trainee engineers through to project and design engineers and ultimately to senior project engineers, technical specialists and fully-rounded broad spectrum consultants.

It comes back to the start: one cannot lead the market in our high-technology environment without impeccable products and people. ABB South Africa succeeds by constantly ensuring those vital cornerstones.

Johan de Villiers GM: Process and

Manufacturing Automation

You can still win a jacket!Our ABB Technology Solutions Reader Survey 2005 continues – and you can still win a jacket over the next few weeks. Just click your way through the ABB website (www.abb.com/za) to the readership survey, open the document on screen and send it in. Do it now!

Another THREE jackets have been made available, and will go to the three names drawn from all those who complete the survey and send it in. Good Luck!!

2/20054

TRANSFORMERS

To detect these movements the frequency response analysis (FRA) method is used. FRA measures the frequency response of passive elements (RLC) of ap-

paratus. In transformers the FRA measures the impedance of the windings over a wide range of frequencies and compares the results of these measurements with a reference set. The main advantage of the FRA method is its ability to detect faults, especially mechanical damage to the windings that cannot always be detected by other means.

Two different FRA techniques exist: the Sweep Frequency Re-sponse Method (SFRA) and the Low Voltage Impulse method (LVI). The Sweep Frequency Response Method injects a wide range of frequencies by making a frequency sweep using a sinu-soidal signal. The signal is generated using a network analyser, which is also used to measure the voltage and to manipulate the results.

The Low Voltage Impulse (LVI) method injects the wide range of frequencies required as a voltage impulse into one terminal. The voltage at another terminal, or current passing through the winding connected to the terminal, or any of the other wind-ings is measured. It is possible to measure several voltages and currents simultaneously. The signals are fi ltered, sampled and stored in the time domain. They are then transferred to the frequency domain and the transfer function is calculated [1].

Measurement technique

The most commonly used circuit for FRA measurement is shown in fi gure 1. Figure 1a shows the FRA instrument where S is the injected signal; R and M are the reference and the measured signals respectively. For a single-phase transformer, the test will be performed in three different confi gurations as indicated by 1, 2 and 3 at the transformer terminals. The fi rst measurement is performed between the high voltage winding and its neutral, the second between the low voltage winding and its neutral, and the third is the inter-winding (HV-LV). In

a three phase situation, nine tests will be performed i.e. six per phase and three inter-winding.

Figure 1: FRA measurement in a single-phase transformer .

The Sweep Frequency Response Analysis

At ABB Powertech Transformers, the SFRA method is used. The method is applied on new power transformers to make signatures, or fi ngerprints. At a later stage when the transformer has been in service and has been through a number of faults, these signatures are used as a base for comparison. Comparisons could be made with individual phases, or with sister units. SFRA is also done on units that have been in service for some time in order to detect any displacements within the active part.

SFRA measurement was performed on a 500 MVA, 400/132/22kV autotransformer. The connection diagram is shown in fi gure 2. The number of tests to be conducted for an autotransformer with a tertiary winding is 15 (See table 1). Since the HV wind-ing has a tapchanger, two sets of tests are conducted with the tapchanger at different positions for each test set. This is done in order to check whether the problem is on the tapped section of the winding, or not.

Frequency Response

Analysis of a Power

TransformerPower transformers are designed to withstand mechanical

forces that arise from the shipping or transportation of trans-

formers, as well as subsequent in-service events. Possible

damage may lead to core and winding movement.

1

HV

2 3

LV

S

M

R

a b

52/2005

TRANSFORMERS

Test # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Measure A11-

a2B11-b2

C11-c2

a2-YN

b2-YN

c2-YN

3A2-3C2

3B2-3A2

3C2-3B2

A11-3A2

B11-3B2

C11-3C2

a2-3A2

b2-3B2

c2-3C2

Per Phase Per Phase Per Phase Inter-winding Inter-winding

Table 1: Tests required for a three-phase autotransformer with tertiary winding.

Figure 2: Winding connection diagram of an autotransformer with a tertiary winding.

Results interpretation

Results are displayed as plots of magnitude and phase of the transfer function. The results can be displayed either in linear, or log scale. Figure 3 shows the magnitude response (in log scale) for the SFRA test conducted on the transformer as specifi ed above. Figure 3 compares traces of the individual phases, i.e. LV to Star neutral for each phase.

The Per-phase measurement

The per-phase measurement targets the individual phase of a given winding. At low frequencies, the infl uence of capacitance is negligible and the winding behaves as an inductor. The mag-netic circuit of the core determines the inductive characteristics. The phase angle is around –90 degrees, indicating the inductive nature of the total impedance.

The magnetic fl ux coupled with the outer phases, a2-YN (white trace) and c2-YN (blue) in Figure 3 shows that outer windings face a different reluctance than the fl ux coupled with the middle phase, b2-YN (red trace). According to [3], the middle phase trace should differ from the traces of the outer phases, which correlate very closely with each other.

As the frequency increases, the capacitive effects start to domi-nate and the phase angle quickly approaches +90 degrees in the

region above 3 kHz. Now the attenuation and the phase shift of the high frequency sine wave passing through the winding are determined by the inductive and capacitive nature of the network. However, the inductive characteristics are determined by the leakage fl ux coupling. In this region the winding response becomes less dependent on the magnetic circuit; hence the traces of the three phases converge and become similar.

At very high frequencies (over 100 kHz), the sinusoidal sig-nal mostly travels outside the winding and refl ect the other elements in a transformer (e.g. leads, support insulation etc). The magnitude and the phase of the transfer function in that frequency region are infl uenced by the inductive/capacitive/resistive nature of these elements.

Figure 3: Per-phase measurement, magnitude comparison of the three phases.

Inter-winding measurements

The Inter-winding measurements target the space between two windings in a given phase. Since the other transformer terminals are left fl oating during this measurement, the effects of the magnetic circuits are excluded. The response is more capacitive from the low frequencies up to the higher frequencies.

Comparing with sister units

At the moment SFRA is designed in such a way that the data is based only on a subjective comparison of traces. For the ini-tial measurement, the traces are analysed for changes between responses of the three phases of the same transformer and changes between responses of transformers of the same design [3]. The response of this transformer was compared with the response of similar transformers.

A11 B11 C11 YN

a2 b2 c2

HV

LV

TV

3A2 3B2 3C2

The second transformer was also a 500MVA, 400/132/22 kV of the same design as the previous one. Two traces were compared, A11-a2 T1, the fi rst 500 MVA transformer and A11-a2 T2 is the second transformer. It can be seen from fi gure 4 that the two traces correlate very closely with each other. In some areas one trace is on top of the other. This then gives some kind of surety to the customer, that the transformers were designed the same way. It also proves that the manufacturing processes and the assembly were done almost exactly the same way.

Figure 4: Comparison of two 500 MVA sister units.

Conclusion

• FRA can be used to detect winding and core displacement in a transformer.

• It is important to have a fi ngerprint of the FRA response of a new transformer. This serves as a base for comparison in future when the transformer has seen a number of faults.

• Transformers of the same design (sister units) should have the same response.

References

[1] Tenbohlen S, Ryder S.A, Making Frequency Response Analysis Measurement: A comparison of the Swept Frequency and Low Voltage Impulse Methods, XIII International Sympo-sium on High Voltage Engineering, Netherlands 2003.

[2] Rahman M, Hashim H, Ghosh P.S, Frequency Response Analysis of power transformer, Electrical Eng. Department, Univeriti Tenaga Nasional, http://www.itee.uq.edu.au/~aupec/aupec03/papers/057%20Hashim%20full%20paper.pdf.

[3]. Doble Engineering, SFRA user guide, Sweep Frequency Response Analysis, Document 72A-1849-01, Rev. B, 06/02.

Contact

Max ChaukeABB Powertech TransformersTel: +27 12 318 9899max.chauke@za.abb.com

2/20056

TRANSFORMERS

P3: PULP PRODUCTION PLANNING

Pulp and paper making is a precise and sophisticated process.

To achieve a product of the desired composition, texture and quality, the right mix of chemicals must be applied under the right conditions. Even slight deviations can adversely affect the output.

The plant’s advanced control system must maintain the desired quality under normal operating conditions. Additionally, it must react to dis-turbances. It must also permit individual sub-systems to be throttled to save energy at times when it is expensive and allow them to be closed down for maintenance – all while keeping output as close to the desired level as possible.

ABB’s Pulp and Paper online production optimiser helps meet these objectives. Material fl ows between sub-processes are simulated and forecasts computed from the present state. Mill behaviour can be optimised and operating decisions supported by a powerful tool.

On-line planning of pulp & paper production

Background

Integrated pulp and paper mills are complex systems. The workfl ow consists of many interdependent sub-processes. Disturbances at one of the stages can quickly propagate to other steps and lead to considerable production losses, quality deviations and wastage of material.

It often takes more than eight hours before the result of a control action has taken full effect in the production process – well into the next crew’s shift. Furthermore, the operation and supervision of the different process sections are divided between different control rooms and hence different operators.

Production changes and disturbances directly affect income. It is diffi cult for operating crews to consider all variables involved in this complex system simultaneously. Predicting the effects of change and taking the necessary actions require appropriate support.

The challenge

The increased recirculation of process streams due to environ-mental constraints in the pulp mill processes makes the recovery cycle increasingly sensitive to disturbances. These lead to slow oscillations in the chemical composition at different points in the production cycle, in turn causing quality deviations, reduced yield and increasing consumption of chemicals.

Pulp and paper mills consume considerable heat and electric power and therefore energy management presents considerable potential for savings. Depending on the location of the produc-tion bottleneck, energy costs can be optimised by better use of buffer tank volumes: sections with great heat demand can be operated at reduced load during periods with high electricity prices and buffers refi lled when prices are low.

Previous improvement schemes have attempted to minimise losses and quality deviations by identifying ideal operation conditions. However, the cost of reaching these conditions is not considered. To be truly useful, an optimisation must consider processes dynamically. Only in this way can it be used to evaluate remedial actions against disturbances, or for operation strategy changes.

Such a tool must optimise mill operations by balancing supply and demand between sub-systems. Every sub-system has to be fed with material, and suffi cient supply must be available for the sub-system to produce as required. The system must be as robust as possible towards disturbances, or partial closures for maintenance (for example through buffering) and provide a common operating strategy for the whole mill.

The system minimises:

• Deviation from the production plan. • Cost of make-up chemicals.• Deviation from preferred liquor composition.• Production loss through disturbances.• Energy costs.

It supports decision-making in real time through:

• Identifi cation of production bottlenecks.• Identifi cation of faulty measurements.

Generic process diagram of pulp and papermaking process.

Pulp and paper mills consume considerable heat and electric power and hence energy man-agement presents consider-

able potential for savings.

2/20058

PULP & PAPER TECHNOLOGY

• Preparation of the mill for maintenance stops. • Better decision support for mill management.• On-line production optimisation.

Pilot project

Pulp & Paper On-Line Production Optimiser is an IndustrialIT

application for real time optimisation. A pilot installation is in progress at Billerud’s Gruvön Mill 2, a large integrated pulp and paper plant in Sweden.

This mill typically produces 640 000 metric tons of paper an-nually. It consists of six paper machines, three fi bre lines and a coating machine. The pulp mill is supervised from fi ve control rooms. It produces sack and kraft paper, containerboard and market pulp.

The mill has recently invested in a new recovery boiler and evaporation plant to raise capacity and eliminate production constraints. To ensure that the anticipated increase in produc-tion capacity is achieved, the company is implementing an on-line production optimiser and expects it to boost produc-tion and reduce consumption of chemicals to a value of several millions of US dollars annually.

The goals were to minimise the following:

• Variation in active production chemicals composition.• Cost of make-up chemicals.• Production losses due to disturbances. • Variation in amount of active production chemicals in stor-age. • Variation of the distribution of active production chemicals in the system.

These goals were met by estimating the state of process vari-ables based on mass balance modelling, and searching for the set of control variables to meet the objectives at minimal operational cost.

Implementation

Modelling – The dynamic mass balance of the complete mill is modelled. The modeling is based on descriptions of all components in the system. These components are categorised into object type groups such as: streams, production units, measurements and calculated properties.

Production unit examples are digester, pulp tank, black liquor tank, lime kiln and paper machine. Measurement examples are fl ow, tank level, sulfi dity, effective alkali, reduction ef-

ficiency, black liquor density and alkali charge.

Streams are connectors between the process units and described by the total fl ow and by the concentrations of the different components that are relevant for each stream.

Measurements are connected to either production units or to streams. The pilot mill system with production units, streams and measurements has the following sizes:

• 25 Production units. • 38 Buffer tanks. • Approx. 250 streams or pipes.• Approx. 250 measurements or ob-servations .• Approx. 2500 variables.

Each production unit consists of equations describing the time de-pendent relationships between all incoming and outgoing streams and manipulated variables connected to the unit. These relationships may be

Billerud’s Gruvön Mill.

Multiple mill functions are considered in the optimisation.

92/2005

PULP & PAPER TECHNOLOGY

considered as a gigantic dynamic mass balance for the com-plete mill and are represented by a system of differential and algebraic non-linear equations. When these are discretised in time, the corresponding number of variables to be optimised can approach 50 000.

To be able to couple the process model to the measurements, a model of each sensor or lab measurement is required. The mea-surements are related to the state by using methods that consider both uncertainty of the models and the observations.

State estimation: To determine the current state, the sum of all measurement noises (uncertain-ties) is minimised. The noise levels at the optimum provide excellent information for sensor, or pro-cess diagnostics. Noise values are tracked in the database and are used to identify faulty sensors or model refi nements.

Optimisation: The production-planning problem is formulated similarly to a model predictive control problem. The optimisation criteria consist of different terms, quadratic and linear, and include terms to enable the following functions:

• Minimisation of deviation from set-point trajectories of any vari-able. One example is the paper machine production plan with its production rates and recipes. These set-points are automatically imported from the scheduling sys-tem. The paper machine is often re-scheduled at two-hour intervals. Other common examples are tank level set-points and preferred con-centrations of active production chemicals.

• Limiting the number of changes of the manipulated variables. Too many adjustments of manipulated control variables, as for example production rate in lime kiln, are to be avoided.

• Minimisation of costs (such as make-up chemical cost).

• Maximising revenues from main products and byproducts. One ex-ample of byproduct may be power production supplied to the grid.

The optimisation criterion is mi-nimised subject to the following constraints:

• Process model.

• Sensor model.

• Upper and lower bounds on variables.A maintenance stop is introduced in the optimisation by adjusting the upper bound of the production rate during the stop time for a production object, eg, digester.

Customised trend describing issue in focus.

Web-based graphical display with some common properties published on the customer intranet.

2/200510

PULP & PAPER TECHNOLOGY

• Upper and lower bounds on change of manipulated variables. Changes in the manipulated variables are sometimes bounded by hard constraints to avoid causing overly aggressive trajecto-ries. For example, a digester should not change its production rate at more than a given rate.

• The production planning should start from the current value of the state estimation.

The optimisation results are used as a basis for production planning.

Simulation

Following the optimisation, a simulation is performed. This provides an indication of the effectiveness of the strategy and predicts how long the process can run before tanks over- or under fi ll, production goals are missed etc. This simulated trajectory is compared to the original plan.

On-line Optimisation

The automatically scheduled calculation 3 includes a state estimation, future optimisation and a simulation. This cycle is repeated approximately four times an hour.

What-if analysis: Several what-if analyses can be started at any time on historical or future scenarios to evaluate alternative strategies and paths.

Solution Output: In the engineering user interface, many as-pects of the model can be changed and all property values from measurement, estimation, optimisation or simulation viewed. A customised trend view with many history and future data of given properties is generated in a couple of minutes and the trend confi guration can be saved for later use. As soon as a trend view is created, any user can access this on the intranet web.

From the web based user interface the most important computa-tion result may be viewed from any PC connected to the mill intranet or internet. It is also possible to perform the most im-portant interaction with the system through this interface, e.g. information of maintenance stop duration and stop time etc.

Client expectations

Benefi ts:

• Increased paper production.

• Savings in make-up chemicals.

• More stable quality of active production chemicals.

• On-line identifi cation of production bottlenecks can be easily viewed by the web based user interface indicating current and future bottlenecks.

• Improved decision support for production management. When operation decisions are made, one must consider many secondary conditions in the mill, the on-line optimiser provides invaluable alternative scenarios for production.

• Common operation strategy for all shifts and control rooms. By using the on-line optimiser, operation management can systematically prioritise between different objectives and targets. When the preferred prioritisation is confi gured in the database and the optimiser’s recommendation adopted, it is easier to motivate the operation strategy.

• More effi cient preparation of the mill prior to maintenance stops is desirable. When a maintenance stop is introduced in the prediction horizon, the optimiser automatically considers this constraint to minimise production loss. It is also possible to evaluate a change in maintenance stop time and quantify the consequence of a longer than expected stop.

• Identifi cation of faulty measurements by using state estima-tion diagnostic results.

This is a brand new product in ABB’s portfolio. The on-line production optimiser employs new and advanced methods, capitalising on the impressive evolution in computing power. Large mill systems including the complete chemical balance and energy management coupled with dynamic production rates can now be optimised in real time using a large number of on-line measurements. The optimiser uses general methods, which is why it can easily be applied to other applications, or system types. The on-line production optimiser is also ‘the missing link’ between advanced process control and Collaborate Production Management system (CPM) to support the planning function for the integrated pulp & paper mills. The ABB ‘On-line optimiser’ is a very powerful tool for process control and decision support and is expected to contribute to signifi cant improvements in the process industry for years to come.

The ABB ‘On-line optimiser’ is a very powerful tool for process control and decision support and is expected to contribute to signifi cant improvements in the process industry for years to come.

Authors

Ulf PerssonThomas LindbergLars Ledung Per-Olof SahlinABB Automation Technologies ABVästerås, Swedenulf.x.person@se.abb.com

Contact

Mark SheldonABB South AfricaTel: +2711 617 2125mark.sheldon@za.abb.com

112/2005

PULP & PAPER TECHNOLOGY

Thermography is used to detect defi ciencies on the unit transformers. While Laser Alignment and balancing is utilised to provide a pro-active service to ensure proper

alignment and balancing on all rotating equipment. However, during the latter half of 2004 the focus changed to the milling plant due to the number of load losses encountered on this plant. A number of machines were having problems associ-ated to alignment. An alignment specifi cation well below the coupling manufacturer specifi cation was adopted, and this led to decreased failures of motor bearings, gearboxes and pinions and girth gears.

What is Vibration Analysis?

All rotating equipment has its own specifi c vibration, which is a product of the rotational speed. Therefore, vibration is mea-sured in amplitude at a specifi c frequency (Hz). The amplitude is measured in either displacement (microns), velocity (mm/s) or acceleration (g’s). This will then give you an amplitude at a specifi c frequency. For example, if one knows the number of teeth on a gear and the rotational speed of that gear you could determine at what frequency your Gearmesh will be in a FFT Spectrum. This calculation can be done for any component of a machine, i.e. bearings, impeller vanes, rotorbars etc.

Case Studies

1) 1B Mill

During a monthly routine survey a calculated inner race defect was detected on the gearbox input NDE bearing (Figures 1 and 2). It was calculated that the inner race defect should be at 194,6Hz. ABB made the call to replace this bearing. After the bearing was removed ABB inspected the bearing for damage and found a defect as shown in (Figure 4).

To measure is to knowABB was awarded the condition monitoring contract at the Eskom Kendal Power Station in May 2003. ABB is responsible for ensuring availability of plant and reducing maintenance costs through using condition monitoring tools such as vibra-tion and oil analysis on the turbine auxiliary plant, the boiler plant, and coal, ash and auxiliary plants.

Figure 1

Figure 2: In Figure 2 a once per revolution impacting can be seen which is due to the defect being on the inner race which is turning with the shaft. This impacting occurs when the defect goes into the load zone.

Figure 3: In fi gure 3 a clear chip mark can be seen on the inner race of the bearing.

2/200512

CONDITION MONITORING

2) 2D Mill

This gearbox was noted to have a higher than normal gearmesh frequency vibration. This mill was due for a 6000Hr service and it was decided to visually inspect the gears internally for damage. Figure 1 shows this mill compared to an exactly similar machine. Incidentally, when this machine was visually inspected it was found that the cage on the input NDE bearing had also failed. If this cage failed it could have caused signifi cant secondary damage.

Figure 5 shows the broken cage.

Figure 1

Figure 2 show the Gearmesh frequency vibration with its harmonics.

Figure 3 show the improvement in vibration levels after the gear was rotated 180 degrees so that it meshes on the other tooth profi le.

Figure 4 shows the damage on the gear tooth.

132/2005

CONDITION MONITORING

2/200514

CONDITION MONITORING

What is thermography?

Thermography is the science of making an unseen temperature visible. Thermography is also a dominating method for tempera-ture fault diagnostics and is an integral part of a preventive main-tenance programme. Thermogra-phy inspection can be performed while the plant is in operation and actual temperatures can be measured without disturbing the production process. It is accepted as the most reliable and effective non-destructing method for testing temperatures.

By using thermography we are now able to categorise faults by measur-ing the temperature of a fault, and subtracting it from the reference temperature and then recommend an action by experienced based standards.

Thermography is used today in a wide fi eld of applications for ex-ample: electrical, mechanical, civil and medical.

What is oil analysis?

Oil can be associated with the life blood of a machine. Early detection of machine defects will show very early in the oil analysis. The fi rst indications of wear will show up in the oil, even though there will be no indication in vibration analysis as not enough wear has taken place to generate a vibration associated with a specifi c component of a machine. Dirt in oil will create wear, thus oil has to be kept as clean as possible. The only way to ensure plant avail-ability is to perform both oil and vibration analysis.

ContactABB Automation TechnologiesFrancois van EedenTel: +27 13 647 9213veedenfr@eskom.co.za

Interoperability according to IEC 61850 means the capability of two or more intelligent electronic devices (IEDs) from one, or several vendors to exchange information and to use it in the performance of their functions and for correct co-operation.

Protection modelling

Data transfer for utility networks is historically a one-way procedure, with data fl owing from a simple sender to a highly sophisticated receiver, which interprets

the data.

In most cases an empoyee reads and understands the data with the help of his comprehensive background. An example is the master-slave communication commonly used in the past, for example the information interface of protection devices accord-ing to IEC 60870-5-103. Interoperability is much more than simple data transfer, but provides for information exchange between two, or more devices of similar intelligence.

The receiver has to understand not only the structure (syntax) of the data, but also its meaning, i.e. the semantics in the context of the process and of his tasks.

The modelling approach of IEC 61850

All functions performed in substations have been split into the smallest entities, which communicate with each other. These en-tities, or objects called Logical Nodes (LN) contain all function related data and their attributes to be communicated. The LNs have a standardised mnemonic name of four letters. Instances of these LNs may be implemented single or multiple in any IED. The data is accessed by defi ned services. Logical nodes for common applications are grouped into Logical Devices (LD), maybe one for control and another for protection. All these objects and services are the base of the function model.

Since functions are implemented in devices, the function model has to be complemented by a physical device (PD) model, which describes the common properties of the device. The common device properties are described in the LN LPHD (Logical Node for Physical Device). The Logical Nodes classes can be instantiated several times on the device, implementing its functionality. If not mentioned specifi cally, we refer to LNs in the context of modelling.

The data model, including its services, is mapped to a main-stream communication stack consisting of MMS, TCP/IP, and Ethernet; time critical messages directly to the link layer of the Ethernet. The system, including the confi guration, is described by the Substation Confi guration description Language (SCL) in XML.

Modelling functions

Vertical function structures

Most functions have a vertical structure, for example they reside with their core functionality at bay level, and communicate with

both the station level (e.g. with the operators work place LN IHMI) and the process level (e.g. with the circuit breaker LN XCBR). If the data model of the circuit breaker (LN XCBR) resides in a breaker IED integrated in the switchgear, or in an I/O card of the bay controller IED, depends on the existence of a serial link (‘process bus’), or parallel wires between the bay and process level.

Figure 1 - Vertical function structure for protection (a,b) and control (c) with (b,c) and without (a) process bus.

Figure 2 - Horizontal function structure for a line protection scheme with distance protection.

Horizontal function structure

Some logical nodes exist in many instances in one level and communicate horizontally with each other. One example is the distance protection. Its LN class is called PDIS. The class defi nition contains three types of data, i.e. identifi cation data like naming, status information data like ‘start’ and ‘operate’ (trip), and setting data. For modelling a device ‘distance pro-tection’, one instance of PDIS per zone shall be implemented,

152/2005

PROTECTION MODELLING

i.e. PDIS1 for zone 1, PDIS2 for zone 2, etc. The semantics of the different instances may be given in the description at-tribute of data Name Plate (NamPlt). The instances of PDIS may also communicate with instances on the other side of the line according to the protection scheme applied. Instances of PSCH co-ordinate the ‘start’ (Str) and ‘operate’ (Op) of PDIS instances and maybe auxiliary functions like time overcurrent protection (PTOC) according to this scheme. The result of the co-ordination is a trip via PTRC (trip conditioning) to the local circuit breaker (XCBR).

Data dependencies

The Logical Nodes (LN) defi nes the data and the related server. To model the data fl ow, inputs from other LNs can also be described by the Substation Confi guration description Language (SCL).

Designing devices

In this paragraph we will look at the design of a device, by investigating how state-of-the-art functions can be applied to the IEC 61850 functional and data model. Here we are not referring to mapping the IEC 61850 model of parts 7 to a protocol – this is already done in IEC61850 parts 8 and 9 in a standardised way. Further, we limit ourselves to the modelling functions of a device. Functions within the system have been handled above.

Application of the device

Each IED is a physical device. IEC 61850 provides the logical node LPHD for modelling the physical device:

• Physical (Hardware) health.

• Communication problems with statistical counters.

• Power supply health with statistical counters.

As all LNs must reside in some logical device (LD), this applies also for above LNs. If you have only one LD, then naturally they reside inside it, otherwise they reside in the specially named logi-cal device LD0, together with a LLN0, which contains common data for all LNs of a LD. In this second case it is recommended to reserve LD0 only for this purpose, and for defi nition of all communication related data sets and control blocks. This

leaves the other LDs completely open for functional modelling, and you do not have to touch them at all for communication and system engineering at the IEC 61850 level (see Figure 3).

Mapping existing functionality

If you have IEDs with an existing functionality and want to map this to IEC 61850, you have to cope with some problems when things do not match:

• Some general mandatory functionality like LD level Mode functionality does not exist.

• Some mandatory data points do not exist.

• Some mandatory data attributes do not exist.

• Data or data attributes exist, but with other data type or other coding.

• Services cannot be supported the way specifi ed in IEC 61850.

The provision of missing, but mandatory features can be pro-vided at the following levels:

• IED internal programme, fi rmware or application level.

• IED communication driver for IEC61850.

The selection between these possibilities depends on the generality of the problem in the IED, and on the amount of application confi guration effort for a real project. If the conversion is restricted to a certain LN type, then it should be solved near the LN implementation. If it concerns a specifi c common data class (CDC), or is valid for data objects (DATA) like Mod, which occur in more than one LN, then it is better solved at IED driver level.

Example: Modelling an IEC 60870-5-103 device

The following illustrates some of the problems and possible solutions with a mapping of the IEC 60870-5-103 fault indi-cations (information number IN) of the distance protection function to IEC61850.

At the fi rst glance this looks good. All IEC60870-5-103 signals belonging to distance protection have been mapped (the miss-ing information numbers IN 86 – 89 only belong to transformer differential). One problem emerges for IN 74 and 75: two different binary signals are mapped to one enumeration type signal. Furthermore, we see that a lot of LNs are used for all signals of the same distance protection function. The difference between the distance backup I>> (PIOC1) and other possible independent over current functions (PTOC2, PIOC3), or earth fault functions (PTOC4, PIOC5) can no longer be seen by standardised identifi cations. However, is this necessary in this case? If yes, grouping by prefi x (see later) offers a possible solution. On the other hand, IEC 61850 LNs have a lot more mandatory data objects, which should be handled:

• The mandatory settings can be provided just as ‘static’ values to be readable via the IEC 61850 stack. The needed values can be provided via the IED SCL fi le. Figure 3 - Typical model of a combined device for protection and

control

2/200516

PROTECTION MODELLING

• The mandatory general data objects Beh, Mod, Health, and NamPlt can be provided as read only values. An exception in our 103-example for Mod is the PSCH, for which beneath a status indication also a control command exists (both IN 17). Furthermore, the protection off command and indication (IN 18) should be mapped to LLN0.Mod. Health can be mirrored from LPHD.PhyHealth, if nothing else is available, and be reduced to the states ok and alarm. An exception here again is the PSCH, where the IN 39 (teleprotection disturbed) can be mapped to PSCH.Health. Mod can be statically on, if not provided specially as described before. Beh then is calculated from Mod and Health as defi ned in 7-4. NamPlt is also stati-cally provided.

• The mandatory PTOC Str and PDIS Str can be provided statically, values are then always FALSE. The mandatory RFLO FDkm can be calculated from FDOhm and some additional parameter.

• IEC 60870-5-103 has some more commands in control direc-tion. From this e.g. the autoreclosure on/off together with the appropriate status indications (IN 16) should be mapped to the RREC.Mod data object.

Above examples illustrate most of the principle problems. The mapping of services is not relevant in this context, but only for a protocol stack mapping.

Packing of LNs

As we have already seen in the 103 example above, it is some-times necessary to group LNs according to some common

functionality, like the distance protection handled above, or all LNs handling the same primary device like the circuit breaker, or for common management like Protection on/off. For grouping of LNs within an IED the IEC 61850 provides the following mechanisms:

• Logical devices.

• Logical node (LN) prefi xes.

Grouping with Logical Devices

The logical device allows via its logical node LLN0 common management functions for all contained LNs. This is specifi -cally:

• The data object Mod, which allows to block or switch off all LNs of the logical device at once.

• The data object Loc, which shows the local / remote state for all process controls within the LD.

Typically an IED contains beneath the LD0 for the physical device separate LDs for protection and control according to Figure 3, and very often, additionally for disturbance upload. If an IED controls switches as well as a transformer, then again this is typically separated into different LDs. Simple IEDs containing e.g. just protection might have only one LD, which then manages all contained LNs. As long as this grouping is function oriented, it is recommended to name the LDs (beneath the mandatory LD0) according to IEC 61346.

If you make top down design of the SA functionality, then you might also group LNs according to LDs, if you want to

IEC 60870-5-103 IEC 61850-7-4, -7-3

IN Description LN DATA Attribute Comments to LN or DATA

64 start/pick-up L1 PTRC Str phsA Protection trip conditioning

65 start/pick-up L2 PTRC Str phsB

66 start/pick-up L3 PTRC Str phsC

67 Start N/ pick-up PTRC Str neut

68 general trip PTRC Op general

69 trip L1 PTRC Op phsA

70 trip L2 PTRC Op PhsB

71 trip L3 PTRC Op phsC

72 trip I>> (back-up operation) PIOC1 Op general Instant. Overcurrent

73 fault location X in ohms RFLO FDOhm mag Fault location

74 fault forward/line PTRC Str dirGeneral DATA no Boolean another value in the same attribute

75 fault reverse/busbar PTRC Str dirGeneral

76 teleprotection signal transmitted PSCH ProTx stVal Teleprotection scheme

77 teleprotection signal received PSCH ProRx stVal

78 zone 1 PDIS1 Op general Distance protection

79 - 83 zone2 - 6 PDIS2 – PDIS6 Op general One PDIS instance for each zone

84 general start/pick-up PTRC Str general Protection trip conditioning

85 breaker failure RBRF OpIn general Breaker failure

90 trip I> PTOC2 Op general Time Overcurrent

91 trip I>> PIOC3 Op general Instant. Overcurrent

92 trip IN> PTOC4 Op general Time Overcurrent

93 trip IN>> PIOC5 Op general Instant. Overcurrent

172/2005

PROTECTION MODELLING

have a common management function for them. These func-tional LDs are then later allocated to IEDs as needed and as available. After this process the LNs for the physical IEDs are added as needed.

Grouping with Logical Node Prefi x

The prefi x allows a functional grouping for engineering or naming purpose within a LD, without any common manage-ment functions. This corresponds to what has been called the horizontal function structure above. Seen from the device modelling and engineering these might contain pre-engineered function parts within an IED fulfi lling one common purpose. A typical grouping has already been mentioned above: all LNs for distance protection get a common prefi x, e.g. 21 accord-ing to IEEE / ANSI notation, and the independent backup overcurrent function gets another (or no) prefi x to separate it. The prefi x can also be used to distinguish and denote different usages of the PTOC related function, e.g. fi rst step or second step overcurrent, earth current protection etc. Related to the 103 example this could give the following groups:

Prefi x (acc. IEEE/ANSI) Grouped LNs

21_ PDIS1, PDIS2, PDIS3, PDIS4, PDIS5, PDIS6, PIOC1, PTRC, RFLO, PSCH, RBRF

51_ PTOC2, PIOC3

67N_ PTOC4, PIOC5

Table 1 prefi xing for protection related LNs.

Another typical grouping, which is more control related, is the following:

Prefi x (acc. IEC 61346) Grouped LNs

QA1 (circuit breaker) CSWI1, CILO1, XCBR1

QB1 (busbar isolator BB1) CSWI2, CILO2, XSWI2

QB2 (busbar isolator BB2) CSWI3, CILO3, XSWI3

QE1 (earthing switch) CSWI4, CILO4, XSWI4

Table 2 prefi xing for control related LNs.

Preconfi guration

IEC 61850 defi nes semantics on LN and data object level. However, the detailed functionality of a LN also depends on its binding to the process. A typical example here is the inter-locking logic (LN CILO), which looks different for a bus bar disconnector than for an earthing switch or a line disconnector. One way to show a special CILO implementation according to the place of the switch in the single line diagram is to use prefi xes, as shown in Table 2 above. Another, more general possibility is to model the appropriate part (e.g. bay) of the plant with links to the LNs in the substation section of an SCL fi le for this IED. This SCL fi le then just contains the IED type description, with an empty IED name, as IED section, and the relevant part of the SCL Substation section.

Process interface

The process interface is modelled in IEC 61850 with T, X, Y and Z type logical nodes. These are typically found on process near IEDs, which are often called PISA (Process Interface for Sensors and Actuators), and are connected to protection and control functions via a process bus. If however a protection device is connected directly with cables to the process, then the question

arises, if the modelling of the interfaces is necessary at all. There are situations, where the necessity is clear: some inter-facing LNs include some basic functionality like monitoring, supervision and command blocking. If the bay level device also provides this functionality, even if in some rudimentary form, then the modelling of the interface is necessary. A typi-cal example is to use TVTR for VT fuse supervision. The rule should be that everything to be visible outside the IED shall be modelled as defi ned in the standard. This might also open up possibilities not so obvious.

An often-used feature is a local/remote switch per bay. This is typically modelled by an LLN0.Loc data object in the control logical device. However, if additional operation directly at the switchgear is possible, and bay level (local) control shall be inhibited, the appropriate Mod or Loc data objects at a LD modelling the switchgear itself, i.e. just containing X and Y type LNs, can be used.

Setting groups

The standard allows the defi nition of setting groups and the switching between different setting group value sets. However, there is at most one setting group per LD. So, if a device sup-ports different setting groups, the LNs having data in the same setting group have to be put into the same logical device, and for each setting group there must exist a logical device. Notably it is not possible to have a part of LN data in one setting group, and another part of the same LN in another setting group.

In case that an IED has setting groups where the value set acti-vation can be changed online, but value change of parameters within the set is not possible, then an IED can just contain a setting group control block, but identify no parameters as belonging to it. Naturally a setting group might contain change-able parameters as well as hidden parameters, i.e. parameters not visible in the data model. However, the hidden parameters can then not be changed in a standardised way. If a change is not wanted, but the actual value shall nevertheless be documented in a standardised way, this can be done via the SCL fi le. For parameters outside a setting group this can also be done by only providing read only access to them.

Conclusion

IEDs with existing functions can be modelled according to IEC 61850 with some restrictions, which should not apply for new IEDs. The free allocation, combination and connection of LNs allow to optimise the system architecture and also to implement new functionality. This supports in a standardised way e.g. combined devices for protection and control, which goes beyond IEC 60870-5-103. The IEC 61850 also provides extension rules, which allow also to introduce in a standardised way new LN classes and data for today unforeseen semantic (future proof). This was already explored in the domain of wind-power plants (IEC 61400-25).

ContactABB Power TechnologiesRoss BothaTel: +27 11 236 7610ross.botha@za.abb.com

2/200518

PROTECTION MODELLING

Pro

jec

ts &

Pro

du

cts

Pro

jec

ts &

Pro

du

cts

192/2005

ABB in South Africa has signed a voluntary Energy Effi ciency Accord, a move that will save the country 12% of energy in the next decade.This agreement was signed as part of Energy Effi ciency month, launched by the national Department of Minerals and Energy in May 2005. The accord was signed by ABB, together with several South African business organisations and industry associations.This voluntary accord has a strategy that targets a 15% reduction in ‘fi nal energy demand’ for the industrial sector by 2015, and a 12% improvement in energy-effi ciency for the nation as a whole by the same date. “As a major business player in this country we see it as our responsibility to contribute towards this effort by promoting and developing energy effi cient alternatives,” says Clive Govender, Vice President, Sustainability Affairs, ABB South Africa. ABB has started working towards this plan through the implementation of ISO 14001 at its sites. “Environmental management plans for all our sites have built-in energy effi ciency programmes that are monitored regularly,” says Govender.“We will also assist our customers to meet these targets by supplying them with products and systems that will improve effi ciency, productivity and that will help in saving energy.”Participation in this initiative is currently voluntary, but will become mandatory when legislated in the next 18 months.

ABB South Africa in bid to save national energy

ABB has won a R57 million order for the manufacture, supply, delivery, installation, testing and commissioning of the electri-cal protection and data communication network for the satellite substations at the City of Tshwane Metropolitan Municipality (CTMM).ABB will provide a data communication network that has the ability to remotely monitor, control and supervise the various components and systems in the electricity distribution network located throughout the Tshwane municipal area. The data com-munication system will service the 132kV networks, as well as the main 33kV and 11kV satellite substations which cover the areas of Pretoria, Centurion, Mamelodi, Atteridgeville, Soshanguve and Acacia. The contract will be managed by DelportDuPreez (Pty) Ltd, Consulting Engineers on behalf of the CTMM.“ABB is pleased to be able to provide the customer with an innovative solution that can be fully integrated with the existing monitoring and communication facilities,” says Carlos Poñe, CEO of ABB in South Africa. Some of the devices that the data network will communicate between are control centres, quality of supply recorders, the Supervisory Control and Data Acquisi-tion (SCADA) system, remote energy metering and TCP/IP communication over the extended Wide Area Network (WAN) of the Electricity Department.“One of the great strengths of the ABB proposal was the ability to offer an economical teleprotection solution in the form of the TEBIT card (Teleprotection and Binary Unit),” says Ndivho Lukhwareni, Strategic Executive Offi cer of the Electricity Department of the CTMM.

ABB wins R57 million order for biggest online utility network in Africa

ABB’s Automation Technologies division recently hosted a group of business people in South Africa from one of the world’s largest mining groups, KGHM Polska Miedz SA of Poland.A group of mining managers from KGHM, the world’s fourth largest copper producer and second largest silver producer, held discussions with ABB South Africa and ABB Poland’s automa-tion heads. They also visited mining operations in South Africa where ABB has excellent references in mine hoists.The combined mining capabilities of both countries emphasised ABB’s global portfolio and technology leadership.The visit provided an opportunity to see the differences in tech-nology between Polish and South African mines and to explore future cooperation.Representatives from KGHM were impressed with the Kopanang mine where ABB is currently working on mine hoists. Ko-panang, an Anglogold Ashanti (Pty) Ltd mine, is located in Orkney, Free State. The mine has fi ve hoists in one shaft, which is an unusually high number together. The hoists comprise: 1 mechanically coupled Blair Rock winder, 1 electrically coupled Blair Rock winder and 3 double drum man/material winders.ABB has completed the electrical upgrade of the electrically coupled Blair Rock winder and is currently upgrading one of the double drum man/material winders.The recent upgrade of the Blair winder by ABB included new DCS 600 thyristors, control panels, drivers’ desks, AHM (Advant Hoist Monitor), and a MCC (motor control centre).ABB Poland is an ABB centre of excellence in power electronics for transforming and rectifying power for railway transportation (used in mines) in applications of up to 3000 V DC.

ABB hosts mining delegation from Poland

The delegation from Poland visiting a South African gold mine.