Taku Noda, Tomo Tadokoro, Takashi Dozaki
©SHUTTERSTOCK.COM/PETRMALINAK
What is the most time-consuming part of simulations? In the past, the answer was the simulation itself. So, many experts worked very hard to develop faster simulation algorithms with less consumption of memory. It was also true that many experts worked hard on applying a new computing facility—for instance, a parallel computer—to a specific type of simulation. Those efforts were begun to make the simulation time shorter as much as possible with less memory consumption. However, now, computers are extremely fast with a sufficient amount of memory. For instance, the smartphone in my hand has a gigahertz-clock CPU with eight cores and gigabytes of memory. This is much better than even the mainframe computer shared by many students and even professors when I was a university student about 30 years ago. Laptop and desktop computers and cloud servers today have even better CPUs, memories, and storage. Of course, a faster simulation speed with less consumption of memory is still important since people want to simulate more complex and thus realistic simulation cases with a shorter or similar simulation time. However, this is not the most important aspect anymore, at least for power system simulations.
If we focus on power system simulations, the most time-consuming part of simulations today is the process of preparing a data case for a simulation. Of course, this is performed by an engineer or a group of engineers. For a transmission or distribution system simulation, the engineer has to reproduce the power system to be analyzed in a simulation program, most likely in the graphical user interface (GUI) environment of the program. To do this, the engineer has to know all the electrical data of the system. The system usually consists of a number of components.
In the case of a transmission system, it includes generators with their control systems, transmission lines, loads, and substation components, such as transformers and circuit breakers with protection relay systems to control them. If a high-voltage dc (HVdc) transmission system and/or a power electronics-based system-control device such as a static synchronous compensator (STATCOM) is involved in the system, the number of data to be prepared drastically increases. Nowadays, representing a wind farm requires a large amount of data. In the case of a distribution system, it includes a distribution substation, distribution lines, step voltage regulators (SVRs), photovoltaic (PV) power generation systems, battery energy storage systems, electric vehicle (EV) chargers, and loads. Now, power electronics-based system-control devices such as STATCOM are being applied to some distribution systems. PV power generation systems, EV chargers, and power electronics-based system-control devices also drastically increase the amount of required data. The engineer has to find and pick up an enormous amount of data from databases (DBs) maintained by the transmission and/or distribution company to prepare the simulation data case. This process takes time. Without exaggeration, weeks or months are usually required to collect and check the data for the preparation of the simulation data case to reproduce the target system.
Depending on the type of simulation, the granularity of modeling is different, and thus, the required amounts of data are different. For phasor simulations like transient stability simulations, a substation is represented by transformers and some circuit breakers. For overvoltage simulations like lightning overvoltage simulations, on the other hand, surge arrestors, busbar structures, line switches, and groundings are represented in addition to transformers and circuit breakers for which even stray capacitances are represented. Once weeks or months are spent, then the simulations are performed. To draw a conclusion from the simulation study, many simulation cases are derived by changing some of the parameters. But to perform just one simulation case requires only seconds or minutes. So, we can understand that the most time-consuming part of simulations today is the process of preparing a data case for a simulation.
To solve this problem, we have been developing a technology for the automatic generation of power system simulation data cases from utility DBs. The development project for distribution systems finished in 2021, and the system, named GriSim, is now used by transmission and distribution (T&D) companies in Japan for the assessment of renewable energy integration and other purposes (it should be noted that T&D services are together in one company in the case of Japan). The development project for transmission systems started in 2021 and is currently being conducted with the support of all T&D companies in Japan. The project is planned to finish in 2026. The system will automatically generate simulation data cases from utility DBs for phasor-based and electromagnetic transient (EMT) simulation programs. In this article, the system for distribution simulations is introduced first. Then, the system for transmission simulations, which is currently under development, is introduced with some illustrations.
The operation of power distribution systems has become complicated due to the massive integration of PV power generation systems. Voltage problems are often reported, and they include voltage rises due to reverse power flow from PV power generation systems, the variation of generator outputs due to changeable sun irradiation, and three-phase imbalance due to single-phase generation systems and loads. Harmonics is another important issue. With the presence of capacitors to support voltages, power electronics-based converters to integrate PV power generation systems, battery energy storage systems, and heat-pump hot-water supply systems may become the source of harmonics problems depending on the conditions. As a conventional problem, the lightning protection of distribution lines is still an important factor.
To cope with these problems, T&D companies use simulation programs to carry out proactive measures and to identify the cause of existing problems. In the case of Japan, the integrated power flow program CALDG is used to simulate voltage and harmonics problems. To solve more detailed problems, the EMT program XTAP is used to perform waveform-level simulations. To estimate the lightning outage rate of distribution lines, the program DLOP is used. As mentioned earlier, distribution engineers used to spend a significant amount of time preparing simulation data cases for these programs. The most time-consuming part of these simulations was literally preparing simulation data cases.
In 2016, the Central Research Institute of Electric Power Industry (CRIEPI) started to develop a computer system for the automatic generation of simulation data cases for distribution simulations from utility DBs with the support of the seven T&D companies mentioned in the “Acknowledgment” section. The developed system is called GriSim, which stands for grid simulations. Depending on the company that provides the distribution service, how distribution facility data are stored is different. Some companies keep all facility data in the DBs equipped with their automatic distribution operation systems, which automatically reconfigure distribution systems to localize the impact of a failure and to minimize outages. Some keep the data in their construction DBs in which data are entered at the time of construction. The others keep the data in both DBs. So, the GriSim system first extracts data from those DBs and stores them as its own DB in the common data format defined for this project. The next section illustrates how the GriSim system works.
The entire GriSim system is deployed in the intranet of a T&D company. As a batch process, the necessary data are collected from various existing DBs in the company and stored in the DB of the GriSim system in the common data format to update the information. This process is carried out automatically once in a period, such as a day, set by the company, and users do not notice the process. It should be noted that the GriSim system is able to keep different sets of data with different dates. They are like snapshots of the distribution system that may change due to constructions and failures. The GriSim system works in the following six steps, as shown in Figure 1.
Figure 1. The data flow to show how GriSim works.
Step 1: When distribution engineers invoke a client program of the GriSim system on their laptop computers, a menu shows up for choosing what kind of simulation the engineer wants to perform. In the case of an integrated power flow simulation or an EMT simulation, a feeder is identified by providing the name of the region, the name of the service station, and the name of the bank transformer to which the feeder is connected. In the case of a lightning outage rate simulation, the name of the region is provided. A date can also be designated to specify which snapshot is used to reproduce the distribution system in the simulation program.
Step 2: The client program accesses the DB of the GriSim system deployed on a server computer in the T&D company and extracts the data of the designated feeder or the region of the designated date.
Step 3: In the case of an integrated power flow simulation, distribution substations, distribution poles, overhead wires and cables, switches, SVRs, thyristor voltage regulators (TVRs), pole-mounted transformers, and so on are shown on a map or on a skeleton diagram.
Step 4: If some of the data are not included in the DBs, the user can supplement those data at this point. In most cases, there are some specific parameters for substations, SVRs, and TVRs.
Step 5: As an option, the user can put the designated distribution system superimposed on a Google Map. For instance, this may be used to check a construction site in advance using the Street View function.
Step 6: The extracted data are converted to a data case file for a selected simulation program.
Now, the distribution engineer is able to start simulations with the generated data case; this process takes just a few minutes. Since distribution engineers used to spend weeks or sometimes months preparing the data case for simulations, this can be considered a successful digital transformation (DX) application.
As mentioned earlier, the GriSim system keeps its own DB in the common data format. In the 1970s, Japanese utility companies started developing and deploying their automatic distribution operation systems to achieve better supply reliability and establish more efficient workflows. Generations of automatic distribution operation systems have been developed and deployed by each T&D company, and these activities are still going on. At the same time, systems to manage construction have been developed and deployed by each company. Due to such reasons, each company maintains its own DB formats for the automatic distribution operation system and the construction management system.
Considering this point, CRIEPI and the seven T&D companies supporting this project have defined a common data format to describe distribution systems. The format consists of about 200 entries. It was defined so that the data of the seven companies can be fully incorporated and described.
The following two points should be noted. Using a DB copied from other working DBs does not affect the operation of those working DB systems and therefore achieves safer development and operation. The other point to be noted is that we can divide the development of the system into two parts: 1) gathering data from existing DBs to update GriSim’s own DB and 2) extracting the necessary data from GriSim’s own DB to convert to a data case file for a selected simulation program. In this way, only for part 1, we have to develop a specific program for each company to collect data from existing DBs. The program for part 2 can be common to all companies. As a result, we can make the development process more efficient.
Before completing the development project, we carried out operation tests in the seven T&D companies that were a part of the project. Figure 2 shows some photographs of the operation tests. Basically, the GriSim system was installed and operated in a server computer inside the company’s intranet, as shown in Figure 1. For companies whose security policy did not satisfy the operation of GriSim at that time, we set up a stand-alone PC for the test. In that case, the distribution system data were collected by a program, and the gathered data were manually connected to the stand-alone PC, as shown in Figure 3. In the case of the former, the test was performed not only in the headquarters but also in the branches. It was also confirmed that the client program worked from PCs in operation stations. The operation station is located at the end of the distribution business, and it actually does the operation—in other words, the maintenance and construction of the distribution systems for which it is responsible. It is quite important that the GriSim system and the simulation tools work at the end of the distribution business.
Figure 2. (a) and (b) Operation tests of GriSim in T&D companies.
Figure 3. An alternative method to perform an operation test of GriSim when the company’s security policy does not satisfy the operation of GriSim at the time of the test.
It is worthwhile to mention the following experience. In the tests, of course, the GriSim system revealed some bugs and failed to generate a simulation data case. But even if GriSim works correctly, generated results sometimes show errors. Then, we started looking at the data and found errors in the registered data and inconsistency of data among branches. In this way, deploying the GriSim system was a good opportunity to review and correct the existing data.
As mentioned earlier, T&D companies have been developing and deploying their automatic distribution operation systems. The latest generations of automatic distribution operation systems have voltage and current sensors equipped at the switchgears, and the systems collect voltage, current, and other data using high-speed communication channels. To utilize those collected data, we have added a feature to the GriSim system together with the integrated power flow simulation program CALDG. All of the latest automatic distribution operation systems developed by different T&D companies measure and collect voltages, currents, and power factors in common. From those data, GriSim is able to extract load curves using a vector decomposition algorithm developed at CRIEPI. Then, GriSim sets the calculated load curves as simulation conditions, and CALDG calculates the voltage profiles in the designated distribution feeder according to the obtained load curves.
Traditionally, various simulations have been used for the planning, design, and operation of power transmission systems. Simulation methods used for transmission systems can be roughly classified into phasor-based simulations and EMT simulations. The phasor-based simulations include various types of power flow, fault, transient stability, voltage variation, voltage stability, and small signal stability simulations.
Traditionally, the EMT simulations include overvoltage, inrush current, abnormal oscillation, and power quality simulations. Power electronics-based converters are now used for various purposes. They are used to integrate renewable energy generation systems, especially for wind energy in the case of transmission systems. Integrating battery energy storage systems into power systems is another application. Power electronics-based converters are also used for long and submarine transmission using HVdc and system stabilization. The latter is conceptualized as flexible ac transmission systems (FACTS). Since power electronics-based converters control power by using high-speed switching by semiconductor devices, EMT or waveform-level simulations are crucial even for system-wide studies in addition to the traditional studies mentioned previously. This is the reason for the recent high demand for EMT simulations. For transmission systems, the software package CRIEPI’s Power System Analysis Tools (CPAT) is mainly used for phasor-based simulations, and XTAP (mentioned earlier) is used for EMT simulations in Japan.
In 2021, CRIEPI started to develop a computer system for the automatic generation of simulation data cases for transmission simulations from utility DBs with the support of all T&D companies. In those companies, the data of transmission facilities have been kept in various formats defined by sections of transmission-line construction, substation construction, and power system operation. They are different section by section. In the past, typical formats used were Excel and comma-separated variable (CSV) files, and some data were just handwritten and then scanned and saved as PDF files. Each section has developed a computer system for construction, maintenance, and operation, and the data have been entered in their own formats.
Figure 4 shows the developed system. The system has the main DB indicated by (I) in Figure 4. The DB uses a relational DB technology and is deployed in a server computer in the intranet of a T&D company. It includes the data on the power system facilities, such as generator stations, substations, and transmission lines, as well as how they are connected. Those data are collected from existing DBs, as described in the next paragraph. In laptop computers used by engineers belonging to one of the sections related to the transmission systems mentioned earlier, the client program, indicated by (II) in Figure 4, is installed. Using the client program, the engineer is able to select a type of simulation supported by CPAT and XTAP and also to specify a portion of the transmission system to be simulated.
Figure 4. An automatic data case generation system for transmission simulations. API: application program interface.
The conversion program, indicated by (III) in Figure 4, translates the transmission system data stored in the common information model (CIM) format into the data format used in the main DB. It is assumed that all T&D companies maintain their transmission system data in the CIM format. This is partially true, and the companies are on the way to completing this. Some of the data may be in different DBs of existing systems, and the others are in Excel, CSV, and even PDF formats. Some of the PDF files may be created by scanning handwritten diagrams and data. Some T&D companies have already created modern DB systems of their transmission systems based on the CIM format for other purposes. If this is the case, the translation application program interface (API), indicated by (IV) in Figure 4, is used to translate the data in the CIM format into the data format used in the main DB.
The data format based on the current CIM specifications lacks some detailed information, such as the thickness of inner/outer semiconductors, the resistivity of metallic sheathes, the geometrical dimensions of the duct banks and trenches, and the grounding types for sheath protection in the case of an underground cable transmission line. It is also difficult to describe detailed specifications of generators and transformers. Therefore, the format of the main DB has been extended to include all necessary information to perform phasor-based and EMT simulations.
The client program has an elaborate GUI. As shown in Figure 5, the transmission system is superimposed on a map, and a portion of the transmission system can be specified on the map for a simulation. If the engineer magnifies a generator station or a substation, its one-line diagram shows up. Then, the engineer is able to further magnify a certain facility to check and modify its data. If the engineer selects a transmission line, then its conductor arrangement appears, and its line constants can be calculated. According to the type of simulation, the granularity of data changes. By specifying a portion of the transmission system and a type of simulation, the client program generates a simulation data case for CPAT and XTAP considering the granularity.
Figure 5. The client program of the automatic data case generation system for transmission simulations. The Japanese characters used in the GUI are left as they are, and the data shown have been modified for publication.
To shorten the time required for preparing simulation data cases, we have been developing a technology for the automatic generation of power system simulation data cases from utility DBs. The development of the distribution system has already been completed, and the developed system, called GriSim, is being used by T&D companies in Japan. We are now developing a similar system for transmission systems in a project that will finish in 2026. This article has introduced the two systems, and it should be noted that those development projects are due to the solid cooperation of an experienced research institute and the T&D companies that transmit and distribute electrical power to the customers. We are sure that similar development projects are occurring in different countries and regions, and we expect interesting comparisons in the future.
The project for distribution systems was supported by Hokkaido Electric Power, Tohoku Electric Power, Chubu Electric Power, Kansai Electric Power, Chugoku Electric Power, Kyushu Electric Power, and Okinawa Electric Power Companies. The project for transmission systems is supported by the same companies plus Tokyo Electric Power Company Holdings, Hokuriku Electric Power, Shikoku Electric Power, and Electric Power Development Companies. We are grateful to Satoshi Uemura of CRIEPI for his kind and generous support of the project.
T. Tadokoro, T. Noda, K. Ishimoto, N. Okada, S. Uemura, and Y. Shumuta, “Automatic generation of input data for distribution system simulation programs,” in Proc. IEEE Innovative Smart Grid Technol., Asia (ISGT Asia), Singapore, 2018, pp. 25–29, doi: 10.1109/ISGT-Asia.2018.8467719.
Taku Noda (takunoda@ieee.org) is with the Grid Innovation Research Laboratory, Central Research Institute of Electric Power Industry (CRIEPI), Yokosuka, Kanagawa 240-0196, Japan.
Tomo Tadokoro (tadokoro@criepi.denken.or.jp) is currently with the National Renewable Research Laboratory, Golden, CO 80401 USA, on leave from the Central Research Institute of Electric Power Industry (CRIEPI).
Takashi Dozaki (dozaki3750@criepi.denken.or.jp) is with the Grid Innovation Research Laboratory, Central Research Institute of Electric Power Industry (CRIEPI), Yokosuka, Kanagawa 240-0196, Japan.
Digital Object Identifier 10.1109/MELE.2023.3320521
2325-5897/23©2023IEEE