Mounir Yehia, Rania Ramadan

GIS: A Technology for Performance Improvement in Electricity Supply Systems 

As modern technologies arise, power utilities are seeking new methods to improve system reliability, power quality and customer satisfaction. High quality of the power can only be achieved if the system losses are at their minimum level and the voltage level is within an acceptable range. For the last decade, Reactive Power Management has been an active research area. Electric utilities have paid more attention to this issue as a result of voltage related operational problems and lack of transmission capability. Consequently, a need for an integrated RPC solution, based on the full integration of technical and economical aspects of the problem arises. The integration must provide a decision support tool that enlarges the scope of possible RPC allocations, and allow the decision-maker to define his priorities (which could be economical, technical or socio-political...) and seek an optimal decision.

In order to reach an optimal solution of the problem presented above, it is necessary to perform a large number of senarios and output analyses for both the economical and the technical algorithms. The system models for this analysis is derived from a large database model. Thus a powerful data management tool with query, analyses and interfaces capabilities is a major necessity for the complete solution of the problem.

Automated Mapping, Facilities Management and Geographic Information System (AM/FM/GIS) provide a real solution to the management of the data required for engineering analyses. In addition, it provides tools to support the planning of spatial components and means to interface to third party analysis tools. As a solution and support for Reactive Power placement, a combination AM/FM/GIS and integrated technical-economical RPC is proposed.

Most AM/FM/GIS electric power studies and analysis applications were developed for electrical distribution systems analysis and design. AM/FM/GIS was used as a component of a full-featured Distribution Automation system [1]. Other applications tackled the problem of designing the electrical supply system for new residential development [2]. Other utilities have been also investing on process automation in order to provide their customers with high quality attendance [3,4,5,6]. GIS was also used to rebuild the design of the whole work procedures in electric utilities [7]. GIS and Global Positioning System are also integrated for mapping and analysis of electric distribution circuits [8].

It is worth to note that GIS in electric power distribution systems applications are well ahead compared to transmission system applications. Transmission systems have not been addressed widely by this new technology. This is due to the fact that transmission systems analysis involves complex electric power algorithms. This paper can be seen as an extension to the work presented in [7] and a pioneer in power analysis techniques for transmission operation planning and control.

This work uses GIS as the main engine for the implementation of the power tools including Load Flow and Reactive Power Allocation algorithms.

The RPC technical-economical methodology presented earlier in [9], is a decision support tool for the optimal distribution of reactive power. Traditionally, RPC was treated as a well-defined problem solution for voltage deficiency and excess line losses in a system. The results were always definite amounts of reactive power allocated at different busses. The trade-off methodology presents a wider solution range leaving the final decision to the engineer, who will use his own skills knowledge and judgment, assisted by the data analysis tools, in order to optimally allocate the reactive power. As a result, the new trend is a Reactive Power Management (RPM) tool where more flexibility in reaching final decision is given to the analyst. The trade-off technique provides a set of the economical solutions resulting in minimal power losses, and a set of technical solutions resulting in voltage improvement. This methodology performs a large number of scenarios for both economical and technical algorithms in order to get the solution sets. Moreover, those runs are accompanied by a set of load flow calculations needed to update the power system data in use. The system models used in the analysis are derived from a large data model and therefore will result in large data output. Once solution results are available, management, analysis and query of the data must be performed by the decision-maker in order to reach the unique distribution of reactive power in the electric network. Therefore, a need for an advanced high level information processing system arises.

The power and flexibility of GIS comes from two characteristics: it is a computer-based system that operates with geographically referenced data. GIS allows the user to perform complex analyses, making it a powerful tool for data query and management. Moreover, the interaction between the geographic and circuit information gives access to more usable information.

GIS is an important tool where the value of visual feedback is used to supplement the detailed result tables coming from circuit analysis tools. Geographic based circuit maps let us explore ways to improve circuit performance through more accurate placement of capacitors and this helps reduce power losses and improve operating voltages,see Figure 1. Moreover, GIS and circuit analysis tools such as RPC, permits analyst to test various circuit configurations and view the results on a color-coded graphical display, and other graphical representations such as graphs. Having a visual display of circuit conditions make the analysis process quicker, easier, and more accurate. Furthermore, relational database access and object-oriented programming allow applications to be highly integrated and quickly developed.

The integrated system consists of four components:

1. GIS software package (Esri's ARCVIEW): this is used as the system engine. GIS is considered as highly efficient software in terms of user interface, graphical output, and data analysis and query. However ARCVIEW script language, AVENUE, is a high level object oriented language, inefficient in terms of algorithmic mathematical computations. This leads to the second component.
2. Power analysis modules: These are the Load Flow (LF) and Reactive Power Compensation (RPC) algorithms implemented using MATLAB. The runs of the algorithms are done user-transparently. Only the results are presented to the user. However, the programs are accessible and changes can be easily introduced, if this is found necessary. Thus, the user is provided with data required for analysis, without having to bother about the complexity of the algorithms. The results of the runs are stored in special data files and transferred to and from ARCVIEW via DDE, supported in both MATLAB and ARCVIEW.
3. Database: the data needed for LF and RPC is initially fed by the user through GIS digitizing or data entry facilities. The data consists of both coverages, representing the electric network components and the geographic regions, and attribute and coordinate tables. These tables are naturally related to the coverages they present. Other data, necessary for the power algorithms are fed by simple tabular data entry forming DBF files, or are first entered as text files, in the appropriate format supported by both MATLAB and ARCVIEW, and then converted into DBF files through ARCVIEW facilities. Once the needed data is available in both graphical and DBF files representation, DDE is used to take care of their transfer between GIS and MATLAB. The latter uses a special text file format of the same DBF file, and conversion is necessary.
4. DDE routines: these are application integration techniques supported by both ARCVIEW and MATLAB. It is a Client/Server mechanism allowing applications to exchange data on the same machine. The link created is called DDE conversation, with an application name such as ARCVIEW or MATLAB, and a topic the name of the file that contains the data or the general topic named "system", as in our case. The data actually being passed during a DDE conversation is called a DDE item. Exchanging data is made through several commands such as "Execute", where client asks server to perform some function, or "Request", where client asks server to return the value of an item, or else, "Poke", where client asks server to update one of its items with some data. In this work, ARCVIEW DDE commands are used in order to:

• Open a conversation with MATLAB as a server.
• Transfer converted DBF files (into text files) to MATLAB.
• Start MATLAB RPC or LF algorithms.
• Get the results back.
• Close the conversation.

The components of the integrated system are shown in figure 2. Note that the user is considered as an evident and indispensable "component" of the overall system.

The integration of GIS ARCVIEW and the RPC and LF algorithms uncovers an easy to use and powerful software package. The different features of the integrated system reveal several important modules requirements needed for the completion of the package:

1. The first objective to be tackled is the file operation module. The foreseen package is intended to manage all projects operations ranging from creating new projects, to opening existing ones, to saving changes, to saving as another project... ARCVIEW offers all these facilities in its main file menu. However, our software is addressed mainly to engineers in an electrical utility. This makes it necessary to customize the module in such a way to show the user the best procedure in fulfilling his objectives. One simple example would be replacing the terminology used in ARCVIEW with a more technical one directed towards electrical engineers.
2. A second requirement is the editor module. This module should offer the necessary facilities to edit, copy, delete, rename, add, move, or drag (etc..) graphical or tabular data and related elements such as tables or views. The editor must be direct, with no complexity, in such a way to guide the user in an efficient and easy manner through his problem solution. Another issue is that the editor must offer the possibility of building new projects by use of digitizers or normal graphical and tabular data entry. Again, ARCVIEW has many of these features, however some tools such as add/delete nodes or add/delete links have to be developed. Moreover, customization of the software is necessary to make it user-friendly.
3. The third requirement is the power analysis integration facilities. ARCVIEW and its object oriented script language, AVENUE, support the interface to the reactive power compensation and load flow modules implemented using MATLAB. Dynamic Data Exchange link to MATLAB should open the way for conversation between both packages. Data preparation and transfer in between both MATLAB and ARCVIEW made the integration implementation possible. It should be emphasized here that the link and conversation are totally user transparent, only the end results of computations are available to the user as DBF files and tables. The user will then be able to choose his data, run algorithms, edit the data, and send it again to the power analysis module for another run. See fig. 3.

 



Figure 3. Power Analysis Menu

 

4. Thematic mapping is the major module contributed by ARCVIEW. Creating maps is not enough to make ARCVIEW attractive for this application. However, adding tabular data about electric or other features to the electric and other geographical coverages, and symbolizing the data available would certainly make it more than appropriate. For instance, capacitor banks can be labeled with their setting value, and symbolized in a rectangular shape. Different colors would be used for different ratings or types of capacitors. This would transform the electric network from a simple single line diagram to a geographically referenced, naturally readable and understood electric network map. In addition, moving around the map with different zooming and panning facilities makes viewing graphical data much more flexible.
5. Finally, data analysis module requirements should enable users analyze and query data, derive statistics about attributes, select and identify a set of features based on specific data... Again ARCVIEW offers many facilities including joining and linking tables, building queries, charting data...

Note that customization option available in ARCVIEW makes it possible to reach optimal results, especially regarding the integration, interface and expandability of the software package in process. The system module requirements are depicted in figure 4.

3.1.3 Integrated System Data Exchange

The software in process is intended to use the data exchange facility offered in both MATLAB and ARCVIEW. As mentioned earlier, both packages support Dynamic Data Exchange (DDE) function. This function enables starting a communication link between them. Once the data files containing electrical and technical information about an electric network are ready in ARCVIEW tables format, the files are automatically saved as data files in a format readable by ARCVIEW. This format is the Data Base File or DBF files. The next step would be to make these files readable in a MATLAB format. This is the text-delimited file or TXT format. ARCVIEW has the ability to transform the input DBF files into input TXT files. AVENUE, the script language in ARCVIEW, takes care of this step. The input TXT files are saved in a special directory accessible by MATLAB. Then, the link is opened using DDE link in AVENUE. Once this link is formed, MATLAB will start running, and the required input TXT files are searched from the specific directory. The RPC or LF algorithms are then called, and MATLAB uses the input TXT files in the running. It should be mentioned here that some modifications are made on the TXT files so that to delete unnecessary data and prepare the required matrices. The outputs of the runs are then saved in output TXT files in the same TXT file directory. Note here that MATLAB would take a little more time than expected for the runs, depending on the electric network under investigation.

Meanwhile, AVENUE will not wait and will proceed in the next statement of code. Therefore, a special checkpoint was built in AVENUE, where a flag is checked before jumping to the next statement of code. This flag is set in MATLAB at the end of the runs, giving AVENUE green light to proceed. Next, the link is closed, giving end to the DDE communication link. Finally, the output TXT files are transformed in AVENUE to output DBF files, again readable by ARCVIEW.

GIS-RPC software has many capabilities mainly intended to assist electric utility analysts and operators in the assessment, planning and operation of the power networks.

When a power system suffers from voltage violations and excess line losses, operators at the utility are urged to take proper control actions in order to solve the problem. These actions are usually the result of careful studies done in the corresponding utility department. However, GIS-RPC offers the operator the possibility of quickly viewing the operating status of the system. Using the queries available in the Data Analysis Module supported by the Graphical User Interface, data such as total reactive power allocated, total power losses, and voltage level at each bus, are conveniently accessible for direct evaluation of the system operating condition. For instance, a graph showing the voltage level at each bus is easily and quickly drawn following any Power Analysis Tool run. The graph is useful in directly showing voltage values at weak busses. An even faster method to tackle this same issue would be via the Thematic Mapper query that gives a distinguished color for under-voltage busses. Labeling them with the voltage values is also possible. Proper actions are consequently taken: the operator, based on his experience and knowledge of the network, interactively allocates capacitor banks at the convenient bus locations presented on the graphical representation of the network.

On the other side, Studies and Planning departments of the electric utility can benefit greatly from GIS-RPC assessment ability. The analyst is offered the competency to carefully and efficiently study and examine, not only the data output from LF or RPC runs offered by the Power Analysis Module, but also any proposed change to the network status.

The system planner can apply any modification to the power network status by graphically editing the related data in order to simulate any situation. For example, as in contingency analysis, the effect of removing a transmission line, a generation unit, or a load bus can be studied by running the LF Power Analysis Module. The output of the Load Flow can then be viewed by means of the graphical interface: the Thematic Mapper can directly show the power flows and the voltage levels; the value of the resulting losses can also be viewed by using queries offered by the Data Analysis Module. Accordingly, the analyst will assess the feasibility of the action he took. Then, RPC modules can be used in correcting the weak nodes' voltages and reduce losses if required. The use of the Technical-Economical RPC module presents a range of solutions, and according to his priorities, the analyst chooses the most suitable one to be implemented.

Moreover, GIS-RPC can be effectively used in long or short term planning. The study of the effect of adding a transmission line, a generation unit, or a load bus, on the voltage level and line losses, helps the analyst reach a confident final decision, especially when a multi-scenario problem is at hand. The availability of the querying and graphical interface enhances greatly the performance of the user.

The editing facility offered by GIS-RPC enables the user to change any output variable related to any electrical feature provided on the graphical presentation of the electrical power system. The effect of varying a generation power, bus load, or transmission line characteristic on bus voltage or a transmission line power flow can thus be investigated.

Power Analysis Modules, such as LF, can be used for this purpose. The results are then viewed by means of Thematic Mapper queries. Color-coded nodes and directed line flows are involved.

RPC modules available in the Power Analysis tools can also be used in sensitivity analysis. The idea is to allocate a specific amount of reactive power at different locations in the system, and the effect is studied by examining the resulting output.

In this case system losses and consequently cost variables are used in evaluating the new system design. The cost of allocating capacitor banks versus the saving from power loss minimization is one-assessment criteria used in this direction.

Finally, it is important to mention that an advanced user of the software can further enhance the analysis capabilities of GIS-RPC, by developing other related modules.

In this section, GIS-RPC is used in the assessment and planning for the reinforcement of the Lebanese Electrical Power System (LEPS) operation. The LEPS has 58 load busses and 11 generation busses. It is considered as an ill-conditioned network [10], with under-voltage problems in some regions.

The first step done in the assessment is running the Load Flow (LF) on the network data. This would help locating weak nodes (i.e. those busses with voltages less than 0.95 p.u.), and calculating active power losses. The results are depicted in Figure 5.

In order to enhance the voltage level and minimize losses, and to assist the planner in reaching a better solution, VAR distribution is affected by means of the Technical-Economical RPC. This will provide the planner with a range of VAR allocations that form a decision range on how much and where to allocate VAR resources. The trade-off methodology is run for a range of minimum attainable voltage values between 0.9 and 0.95.The results are shown in figures 6. Figure 7 present the total power loss versus the VAR allocations corresponding to each solution (i.e. technical and economical range of solutions).

The results of the technical-economical RPC show that all the weak nodes voltages have been gradually raised above 0.95 p.u., except when allocating 34 MVAR using the economical RPC, only one node (node 63, as shown in figure 6) out of the four weak nodes remains with a voltage of 0.94 p.u., which is relatively not low. The node is located in the NorthEast region of Lebanon. This proves that the technical solution is better shaped for voltage improvement.

Figure 6 shows that the total power losses decrease gradually with every increase in VAR allocation. It is obvious that the losses are always lower for the economical solution when using the same amount of VAR allocated.

Finally, concerning the LEPS, if our objective is to improve the voltage all through the network, the best solution would be to allocate 33.7 MVAR resulting in a minimum voltage of 0.952 p.u. and power losses of 0.252 p.u. (10.95 % less than the original system losses). The results of this solution are shown in figure 8.

Figure 9 and 10 show the Trade-off Methodology VAR allocations for 33.7 MVar for the technical RPC and the economical RPC, respectively. The Economical distribution of VAR resources is over a wider geographical and topological range of busses, thus the system is weakly connected leading to losses being spread over a wide area. The technical distribution of VAR resources is more constrained to the region of voltage deficiency, thus it is the optimal solution for the problem at hand.

Integration of GIS with the power quality improvement tool technical-economical RPC results in advanced customized and user friendly decision support package that provides a state-of-the-art data management query, analysis and graphical interpretation of the information. It provides the utility engineer with a wider scope to enhance the decision -making process. The optimal allocation of reactive power compensators is not a result of attaining unique objective; it is rather an outcome of complete understanding of the physical system, economic situation and customer requirements. Furthermore, this package has many capabilities to assist the utility analysts and operators in the assessment, planning and operation of the power network. Application to the Lebanese power system is presented where different scenarios for reactive power allocations are provided and it is up to the decision maker to set the priorities and choose the appropriate solution.

Mounir Yehia, Associate Professor

Rania Ramadan, Graduate Student

Faculty of Engineering and Architecture

American University of Beirut

P.O.Box 11-0236

Beirut, Lebanon

Fax: (212) 478-1995