Vous êtes sur la page 1sur 6

©DIGITAL VISION, LTD.

application, power application necessitates a high level


eregulation of the energy industry has resulted in

D the emergence of software that coordinates the


functions in power systems. The complexity of
these vertically split power systems demand stringent
of quality. Design discrepancies in a power software
application can result in low reliability of the energy sup-
plied to end-users. This causes monetary and other
kinds of losses to the suppliers as well as the users. The
requirements from the employed software. As the mar-
complexity of power software owes its nature to applica-
ket rules change, the regulating policies stated by FERC
tion usage and changing market rules in the deregulated
will need incorporation into the software. By using the
energy markets. Power system software for present-day
latest technology and applying software management
markets should be developed with desired capabilities
practices during development, the measurable results
and maintenance features in mind. This can be accom-
include better designs of the software coupled with high
plished by using a suitable tool for software modeling
quality and performance. A variety of commercially
and configuration management. Tool selection is deter-
available tools can be used to improve the software
mined by a combination of factors that includes, but is
development process. Software modeling and configura-
not limited to, architecture, version control capabilities,
tion management encompasses two key areas in the
change control management, process management,
development life cycle.
reporting, and economics.
This article presents an overview of software model-
ing and configuration management as applied to soft-
ware development in deregulated energy markets. It Software Development Environments
outlines the requirements for software quality, various With the advent of digital technology, the considerable
tool features, and challenges ahead in the development cost reduction of computer hardware has allowed soft-
of software for deregulated energy markets. ware engineers the opportunity to be provided with
automated tools to support the development process.
Software Quality in Power Applications Today’s software engineering environments are integrat-
The quality of the software should meet or exceed the ed around software management systems that offer sup-
customer’s expectations and be easily adaptable to port for all software process activities. All products of
changes. The quality of the software is often related to the software process can be placed under configuration
the quantity of defects. Fewer defects mean higher quali- management using the appropriate configuration tools.
ty. However, software quality is also an integral part of Software designs can be visualized best using commer-
the full product development cycle that spans require- cially available modeling tools. The use of such software
ment definition to product release. To develop quality tools results in quality enhancement of the software sys-
software, suitable tools to control the development tem and its maintenance. Various software development
process must be employed. As in any critical control environments can be classified as multiuser, cross-plat-
form, and multiproject.
H.S. Gill, V. Filipovic-Gledja, and H.P. Mukherjee are with ABB Energy
Information Systems, Santa Clara, California, USA. T.S. Sidhu and L. Cui Multiuser Environment
are with the University of Saskatchewan, Saskatoon, Saskatchewan, It is common in today’s software development process
Canada.

ISSN 0895-0156/01/$10.00©2001 IEEE October 2001 33


have specific requirements (rules, policies, etc.) that
Project A Project B
must be implemented in the software. As such, suitable
1 3 development environments for parallel projects has
5 become inevitable. To keep track of required software
2 4 data and program files, companies using parallel pro-
jects that are engaged in the development process need
sufficient tool support. Moreover, part of the develop-
ment activities may be common to these parallel pro-
Figure 1. Parallel projects environment
jects. To avoid duplication of effort, merging the work
among these projects is a desirable feature. This envi-
ronment can be visualized as depicted in Figure 1.
Two projects (A and B) share common files as repre-
app Applications
sented by the numbered triangles 3 and 4. These pro-
jects have other files that are used by each individual
project (files 1 and 2 for project A and file 5 for project
Architecture B). Project A team members have read-write access to
arc
files 1 and 2 but have read-only access to files 3 and 4
(common files). Similarly, project B team members have
read-write access to file 5 but have read-only access to
dbm Database Management the common files. The possibility exists for a team mem-
ber to have read-write access to one of the common files
and not the other. All such combinations may compli-
cate the parallel project development environment. In
gui Graphical User Interface
Figure 1, files are used to illustrate the environment. In
practice, these files may be replaced by components
that constitute a number of files. The project team mem-
bers may have read-write access to a particular compo-
req Software Requirements nent or part of the component and a read-only access to
other components or parts of a component. These fea-
tures make the development process extremely com-
plex. An assessment of all such permutations and
doc Documentation operating scenarios should be considered when design-
ing a power software system. To ensure product quality
in the development of power system software, software
Figure 2. Types of components
modeling and configuration management requires spe-
for a group of people to be working on the same software cial attention. The following sections describe the appli-
system. Each team member works on a module that is cation of tools used in the development process of
periodically integrated during the development process. power system software.
Being complex in functionality, the power system soft-
ware needs such an environment. The team members
Configuration Management
independently develop the software modules that repre- Due to the onset of deregulated markets and the com-
sent the functions of the software. These modules are plexity of the power system software, the application of
then integrated to build the final system. any configuration management tool requires a detailed
view of the associated data and program (source code,
Cross-Platform Environment object, executable) files. These files can be grouped into
Power system software is based on software packages components, based on their functionality, to aid configu-
that function on different operating systems. Coordina- ration management. Figure 2 shows an example of cate-
tion of these software packages on multiple operating gorizing these files into components. These components
systems, in turn, need middleware Many software com- include files related to specific functions as follows:
panies use this environment to carry out their efforts, ■ Applications: Bus scheduler and network builder,
which also facilitates development in accordance to the power flow module, ancillary services and conges-
preferences of development team involved. tion management, security analysis, etc.
■ Architecture: Database and middleware applica-
Multiproject Environment tions programming interface, event-logging mod-
Most power software companies are engaged in develop- ules, utility tools (porting software, process
ing software for different power pool areas. These areas scheduler), etc.

34 IEEE Computer Applications in Power


■ Database management: Database population and
data management, utility tools (database tuning,
app doc
schema modeling), etc.
■ Graphic User Interface: Interface source files and
library functions, information about configuration
and utility tools (modeling, memory leak), etc.
■ Software requirements: This includes software
requirement documents that keep track of the sys Data req
periodic changes occurring in the software. Storage
■ Documentation: Release notes, software variance
log, user manuals, help files, etc.
Each component is a storage unit in a configuration
management tool, as shown in Figure 3. The contents of
these components can be accessed using the load and dbm mmi
view features of the configuration management tool. Fur-
ther, these components will have unique substructure.
Figure 3. Data storage for configuration management
The subdirectories can be kept the same for uniformity
with the possibility of empty subdirectories. These sub-
structures can vary from one software to another. Figure The object-oriented
4 shows an example of one such component’s substruc- approach of software
ture as viewed using the features of the configuration modeling seems promis- app
management tool. The src, inc, and run directories repre- ing for designing software src
sent the collection of source code, include, and exe- systems for deregulated
cutable files, respectively. energy markets. Because
inc
Team members working on the power system soft- power systems comprise
ware project can load the components necessary to a large number of com-
carry out development efforts. The configuration man- ponents that operate in
agement tool tracks all changes made by team members. conjunction with the op-
As such, any version of a particular file can be retrieved eration of power systems
Figure 4. Substructure of a
and used by any member in the team. Suitable baseline carried out under defined
component
labels can be assigned to keep track of significant mile- system constraints, the
stones accomplished during development. software developed for this purpose must have the
capability to consider these requirements. Deregulation
Power System Software Modeling has further complicated the operation scenario due to
Modeling power system software helps the development the involvement of numerous entities (transmission ser-
teams comprehend, organize, and develop complex sys- vice providers, power suppliers, customers, etc.). Addi-
tems for the present as well as the future. As depicted in tionally, operation is conducted under the policies and
Figure 5, constructing models will help designers focus regulations set forth by electricity regulatory authorities.
on the project on a larger scale without having to get To outline the importance of software modeling, a sam-
into the specific details of individual elements. ple of a system for the deregulated energy markets is dis-
Different modeling approaches are possible for any
software system. The two most commonly used
approaches are algorithmic and object-oriented. The Software Modeling
algorithmic approach is based on procedures and func-
tions. Maintenance of software systems built using algo-
rithmic focus is very difficult. On the other hand, the Requirements Software
Comprehension Maintenance
object-oriented approach has proven to be immensely Design
useful in building systems involving all degrees of size Improvements
and complexity. These commercially available tools
model the software by defining the software components
and their attributes and establishing mutual interaction
among the components. This results in a set of scenarios
that provides a way to describe the system behavior. Quality Enhancement
The dynamic behavior of the individual components can
also be defined in terms of different states of compo-
nents and the transition between these states. Figure 5. Software modeling and software quality

October 2001 35
• Public Market • State Estimator proper implementation. An approach
Information • Load Forcast to modeling this scheduling system
• Generation and • Historical Information
Load Schedule would be to visualize each portion of
• Ancillary Service Bids the system into smaller portions. Fig-
ure 7 shows the load forecast mod-
ule divided into submodules. The
load forecast module requires input
ISO User Interface
Power Management System
of weather data, actual load, and
user parameters. This module
includes the functionality to compare
Saved Generator the ISO load forecast with the aver-
Scheduling Cases Schedule Data age load forecast provided by the
(Previous System Scenario)
scheduling coordinators.
(Available Capacity)
User preferences are obtained, his-
(Facilitate Formulation torical load data is searched, and a
of Network Model) sample load curve is generated. Later,
Network Model Builder it compares the actual load from the
automatic generation control (AGC)
• Foward and Real-Time Markets subsystem and gives error estima-
• System Dispatch tions to the user about the load fore-
• Status of • Congestion Management
casts. To model this component, the
Circuit Breaker
• System
complexity of the system must be
Interconnections considered. Individual objects are
• Loss Factors defined as shown in Figure 8.
Full Network Model Various Scheduling Applications The figure shows the load forecast
module to be composed of a number
Figure 6. ISO scheduling system of components, each of which has

cussed in the following section.


Figure 6 shows a block diagram of the schedul-
ing system network of an independent system
operator (ISO).
A detailed network represents the transmis-
sion network for performing functions such as
congestion management, determining transmis- Weather Forecast and
sion congestion prices, forecasting transmission Actual Weather Data
loss responsibilities, and conducting transmis-
sion assessment. The network model builder is
able to construct the network model for use by
Load
various scheduling application functions. For Forecast
each network model, load and generation condi- User Settings Predicted Load
tions are specified, and any scheduled out-of-ser-
vice transmission equipment are mentioned. The
network model provides the capability to repre-
sent the network conditions over all 24-hour peri-
ods. Additionally, the network model reloads
previously stored saved cases and uses model
data from online (power management system
state estimator) models. Historical Load
From the description of the network model, it
is clear that this component of the scheduling Figure 7. Load forecast module of a network builder
system is very complex and involves diverse
functionality. Its realization in software needs extensive several attributes and operations. The exchange of infor-
design, hence, the need to model the power applications. mation among the components takes place in the opera-
It is obvious that modeling of the component will be tions. The interface provides abstraction for
useful in better understanding its requirements and implementing functions that can be provided by individ-

36 IEEE Computer Applications in Power


To ensure product quality in the development of power system
software, software modeling and configuration
management require special attention

ual components. In Figure 8, the central load-forecaster nia market structure, combined with an imbalance of
component has attributes such as location, day, hour, supply and demand, has led to higher than usual whole-
error percentage, etc. Also, the operation forecast load sale prices.
collects information from all other components such as In addition, FERC urged California regulators and
UserData, WeatherData, and HistoricalLoadData to gen- other policymakers to resolve issues that have placed
erate the forecast load. The ForecastedLoadInterface consumers at risk. This includes delays in citing new
helps to create an abstraction for the load forecast com- generation and transmission capacity, and the imple-
ponents for a 24-hour period. Similarly, all other compo- mentation of additional retail demand response pro-
nents used in the network model builder (Figure 7) can grams. Implementation of deregulation related rules and
be modeled. policies would require changes to the existing software
Software modeling helps to understand the relation- systems. In the near future, software developed using
ship between the objects inside the parent module and modeling and configuration techniques will be able to
the components of other modules, where it sends or easily adapt to the changing requirements.
receives information. The maintenance of the design
structure and its ongoing changes can be handled Open Platforms
effectively by this approach. Over time, this results in With the advancement of software technology, new soft-
higher productivity with enhanced software quality. ware packages have emerged on the horizon that are
compatible with multiple operating systems. The prolif-
Challenges Ahead eration of operating systems has necessitated that the
With the evolving power mar-
ket scenarios and the emer-
gence of diverse software, WeatherData
software quality is bound to location
dryTemperature
invite more attention than illumination
ever before. This poses strong humidity
rainfall
challenges to companies who otherWeatherData

are involved in developing getWeatherData() ForcastedLoad


dataRefinement() hourNumber
deregulated energy system etcWeatherData() loadMW
peakLoad
software. peakHour
Information Collected on otherForecastedLoad
Hourly or Daily Basis
Changing Market Policies generatedForecast()
UserData etcForecastedLoad()
The U.S. Federal Energy Regula- location LoadForecaster
Instantaneous Hourly and
tory Commission (FERC) recent- range
Information location Intermittent
relaxationFactor day
ly approved several actions to Collection Information
smoothingFactor hour
otherUserData errorPercentage
revamp California’s power mar- otherLoadForecaster
getUserData()
kets over the next 2 years. This setUserData() forecastLoad()
ForecastedLoadInterface
calculateSmoothingValue() etcLoadForecaster()
is to ensure that sufficient sup- etcUserData()
hourNumber
peakLoad
ply is available to meet load peakHour
otherForecastedLoad
and reserve requirements. One-Time Information
FERC also approved changes to Collection transferForecast()
etcForecastedLoad()
protect wholesale customers HistoricalLoadData
from unreasonable rates during location
day
the time it will take to adopt peakLoad
peakHour
long-term market remedies. otherHistLoadData
FERC has proposed these storeData()
actions to contain the high retrieveData()
etcHistLoadData()
prices faced by California con-
sumers in the summer of 2000.
Its report noted that the Califor- Figure 8. Logical model of the load forecast module

October 2001 37
With the evolving power market scenarios and the emergence
of diverse software, software quality is bound to invite more
attention than ever before

power system software support the operating system. A pp. 43-49.


practice by companies developing power system soft- R.F. Hirsh, Power Loss: The Origin of Deregulation and Restructuring
in the American Utility System, Cambridge, MA: MIT Press, 1999.
ware was to provide in-house platforms for implementa- D. Sparks, “The California Electricity Market,” IEEE Power Engineer-
tion of the power software system. This required ing Review, Jun. 1999, pp. 11-12.
customer utilities to spend additional resources on the
infrastructure to support these in-house development
systems. The present day trend is to provide open plat- Biographies
Harjeet Singh Gill is an applications engineer at ABB Energy Informa-
form systems that have the capability to run on any spe- tion Systems, Santa Clara, California. He received his B.Sc. degree from
cific operating system without the need by the utilities to Government College, Chandigarh, India, in 1985, B.E. (Hons.) degree
incur additional financial burden. One such platform has from Panjab University, Chandigarh, India, in 1989, and the M.Sc. and
recently been developed by ABB that will change how Ph.D. degrees from the University of Saskatchewan, Saskatoon,
Saskatchewan, Canada, in 1997 and 2000, respectively. He is involved in
the deregulated energy systems are presently operating. developing software systems in association to deregulated energy mar-
Concerted efforts are being directed in developing such kets and administering configuration environment of products. His
innovative and versatile system infrastructure compo- interests include software engineering, system administration, and
nents, which are related to the software engineering con- power system analysis.
Visnja Filipovic-Gledja is the Customer Focus director at ABB Ener-
cepts used as the foundation. Attention has been drawn gy Information Systems. She received her Ph.D. from McMaster Univer-
to power system engineers to make optimal utilization of sity, Ontario, Canada. Her research on developing a probability-based
such concepts in the deregulated energy markets. model for predicting the loading of transmission lines is highly
acclaimed and has earned her a patent for her work. Her vast experi-
ence in the software and engineering disciplines has garnered her posi-
Diverse Software tions of heavy responsibilities. She is focusing her efforts on process
With different vendors constantly being added to the improvements to ensure the quality of software products delivered by
software solution providers, the range of available power the company. The implementation of the capability maturity model
system software packages is increasing. Due to this constitutes a prominent part of her priority initiatives. Software engi-
neering and power system analysis form her major working interests.
trend, compatibility between the software packages can Himadri Prasad Mukherjee is a consulting engineer at ABB Energy
be an issue, making software development environments Information Systems. He received his M.S. in electrical engineering
more complex than ever. Software quality is a major con- from Arizona State University with a dual major in power systems and
cern when different software packages are working hand- computer science in 1996, and B.Tech. degree from Regional Engineer-
ing College, Kurukshetra, India. He is a member of the IEEE and has
in-glove with each other. over 5 years of professional experience in power systems software
It is expected that software requirements in deregulat- development and switchgear manufacturing.
ed energy markets will be stringent. With the importance Tarlochan S. Sidhu received his B.E. (Hons.) degree from the Pun-
of software quality being paramount, the use of tools for jabi University, Patiala, India, in 1979 and his M.Sc. and Ph.D. degrees
from the University of Saskatchewan in 1985 and 1989, respectively. He
software configuration and software modeling will be a worked for the Regional Computer Center, Chandigarh, India, develop-
basic necessity. ing software for computer-based systems, and Punjab State Electricity
Board in India in distribution system operation and thermal generating
Acknowledgments station design. After obtaining his Ph.D. degree, he joined Bell-Northern
The authors thank the Research and Development group of ABB Ener- Research Ltd., Ottawa, Canada, and worked on a software development
gy Information Systems in Santa Clara, California, for their ideas and project. In 1990, he joined the University of Saskatchewan, where he is
suggestions that contributed to the creation of this article presently professor and graduate chairman in the Department of Elec-
trical Engineering. He is a senior member of the IEEE, registered profes-
sional engineer in Saskatchewan, and a chartered engineer in the UK.
For Further Reading His research interests include applications of computers and newer
M. Shahidehpour and M. Marwali, Maintenance Scheduling in Restruc- technologies for achieving improved protection and monitoring of
tured Power Systems, Norwell, MA: Kluwer, 2000. power systems.
“Hemisphere Report,” Forbes Magazine, 15 Dec. 1997. Lan Cui received her B.Sc. and M.Sc. (Hons.) degrees in electrical
T. Quatrani, Visual Modeling with Rational Rose 2000 and UML, 2nd engineering from the Shandong University of Technology, Jinan, China,
ed., Reading, MA: Addison-Wesley, Feb. 2000, p. 7. in 1988 and 1991, respectively. Currently completing an M.S. degree
I. Sommerville, Software Engineering, 4th ed., Reading, MA: Addison- from the University of Saskatchewan, she has experience in designing
Wesley, 1992, ch. 18. substations and associated protection schemes, strategy analysis of
G. Booch, J. Rumbaugh, and I. Jacobson, The Unified Modeling Lan- reactive power optimization in distribution systems, developing
guage User Guide, 5th ed., Reading, MA: Addison-Wesley, Oct. 1999, ch. 1. automation software for resource management, power system security,
ABB Energy Information Systems, Energy News (Source: Reuters), and contingency analysis. Her professional interests include software
November 2000. engineering, computer programming and algorithmic analysis, and
“Restructuring the Thin-Stretched Grid,” IEEE Spectrum, Jun. 2000, power system reliability.

38 IEEE Computer Applications in Power

Vous aimerez peut-être aussi