Taking UCA Mainstream
Substation integration and automation system design begins with the concept of providing access to vast amounts of untapped substation data beyond what traditional supervisory control and data acquisition (SCADA) systems provide system operators. To gain access to substation data, a utility can either form a system architecture around specific products, making the communications work, or the utility can select its desired system architecture and then find ways to make the products work.
The Utility Communications Architecture (UCA
| Description | Modbus | UCA | DNP |
|---|---|---|---|
| Registered user group members | 2280 | 206 | 250 |
| Products | >200 | 52 | >48 |
| Ethernet-based conformance-tested products | 15 | 0 | Unknown |
| Free user community | Yes | Yes | No |
| Fee-based user group | Under development (UD) | Yes | Yes |
| Technical committee | UD | Yes | Yes |
| Document control | Yes | Some | Yes |
| Test documentation | Yes | UD | Yes |
| Independent testing | Yes | UD | Yes |
| Inherent time-stamped events | No | Yes | Yes |
| Master/slave communications | Yes | No | Yes |
| Client/server communications | No | Yes | No |
| Unsolicited reporting | Application dependent (AD) | Yes | Yes |
| Peer-to-peer communication | AD | Yes | AD |
| Broadcast messages | Yes | Yes | Yes |
| Custom data objects | Yes | Yes | Yes |
| Self-description | No | Yes | No |
| Select-before-operate | No | Yes | Yes |
Automation system design begins with conceptual formation. In this phase, it is essential to separate products from services and address the design issues related to serial, Ethernet or hybrid architectures. Once Ethernet and TCP/IP are selected, UCA, DNP or Modbus must be evaluated. The work required for each is based upon protocol complexity. Modbus can be simple (with perhaps complex implementations). DNP can be as simple as Modbus or as complex as UCA (depending upon the device implementation). UCA is the most complex. (Table 1)
Once UCA is chosen, additional design work includes specifying system requirements and defining the risk management plan. The utility can minimize the costs associated with these tasks by:
Adopting requirements that are not more stringent than the emerging standards or the de facto standards used by vendors.
Adopting an architecture and technology that is supported by international standards and vendors.
Selecting hardened equipment to minimize the probability of equipment failure.
Working with stakeholders to identify acceptable changes to existing corporate standards.
Most design tasks will require more initial effort, resulting in higher initial design costs, which will vary proportionally with the amount of substation equipment and independent vendors. The more devices and vendors, the more complex the project will be and the higher the initial design costs. As the number of integrated substations increases, new utility standards and practices can leverage UCA depending upon the level of integration, resulting in reduced costs as designs are streamlined.
Legacy or serial devices most likely will remain in the system architecture. This increases project risk and the initial design costs because the UCA object model presents a steep learning curve and protocol conversion will be required. Until configuration databases become more flexible, the substation configuration language is implemented and automated configuration tools are developed, integrating these devices will represent a significant challenge for automation projects. As vendors, integrators and utilities gain more experience with specific devices, the impact on project costs and risks will decline.
Purchased Components
The UCA International Users Group provides a list of UCA 2.0 compliant products at www.ucainternational.org. This list is the only clearinghouse of UCA devices, so vendors developing UCA devices are encouraged to provide product information to the Users Group. The product list has been grouped into categories for relays, meters, substation hosts, networking, operator interfaces and equipment monitors. The list price of adding UCA capability to the base device is used in this analysis.
| Description | Vendor | Modbus | DNP | UCA |
|---|---|---|---|---|
| DPU2000R | ABB | Yes | No | Yes |
| F60 | GE | Yes | Yes | Yes |
| C60 | GE | Yes | Yes | Yes |
| D60 | GE | Yes | Yes | Yes |
| L90 | GE | Yes | Yes | Yes |
| L60 | GE | Yes | Yes | Yes |
| T60 | GE | Yes | Yes | Yes |
| T35 | GE | Yes | Yes | Yes |
| F35 | GE | Yes | Yes | Yes |
| B30 | GE | Yes | Yes | Yes |
| G60 | GE | Yes | Yes | Yes |
| 9745 | RFL | Yes | ||
| 9300 | RFL | Yes | ||
| 421 | SEL | No | No | Yes |
| BCD-G | ZIV | Yes | ||
| Minimum cost | $744.72 | $744.72 | $744.72 | |
| Average cost | $744.72 | $744.72 | $848.52 | |
| Maximum cost | $744.72 | $744.72 | $1840.00 |
| Description | Vendor | UCA | DNP | Modbus |
|---|---|---|---|---|
| Signature System 5530 DataNode | Dranetz-BMI | Yes | No | No |
| PowerServe | Bitronics | Yes | Yes | Yes |
| Average cost | $265 | |||
Relays
Relays have a variety of classifications and vendors, but only three vendors support UCA via Ethernet in 12 relays. Some vendors are releasing UCA-based relays in 2003. Relays that support UCA may also support DNP or Modbus over Ethernet. The cost of adding UCA to some relays is the same for all of the supported protocols. (Table 2)
Currently, only two meters support UCA and are from different vendors. While this is a concern, many relays now provide nonrevenue accuracy-metering values. The requirement for additional metering or power-quality features can be obtained by adding the meters that support UCA. In addition, one meter could be used to migrate to UCA by using the product as an intermediate platform to add control, status and analogs without requiring the replacement of relays that do not support UCA. (Table 3)
Networking Components
Since UCA supports Ethernet, networking components support UCA. Networking costs vary with the type of equipment used, from hubs, switches, managed switches and routers, to the number of ports and hardware required. Considering the substation LAN to include only hubs and switches simplifies the analysis by removing costs associated with extending the corporate WAN to the substation. The choice of a networking device must include the applied standards (substation hardened vs. commercial grade). This is not a trivial decision, as many substation environments present challenges to commercial-grade networking components. A US$100 commercial hub can be replaced with a $4300 hardened, multimode fiber, managed switch. This price difference demonstrates that networking equipment should not be chosen based upon price alone. (Table 4)
| Description | Vendor | Multi Mode | Single Mode | Copper Ports |
|---|---|---|---|---|
| RS8000 | RuggedCom | Yes | Available | No |
| RS8000T | RuggedCom | Yes | Available | Yes |
| RS1600 | RuggedCom | Yes | Available | Yes |
| RS1600T | RuggedCom | Yes | Available | Yes |
| FSR200 | OnTime Networks | Yes | Yes | Yes |
| FST200 | OnTime Networks | Yes | Yes | Yes |
| 6K25 | GarrettCom | Yes | Yes | Yes |
| Minimum cost | $212.00 | $416.00 | $104.00 | |
| Average cost | $392.43 | $992.83 | $194.42 | |
| Maximum cost | $650.00 | $1375.00 | $425.00 |
| Description | Vendor | UCA Client | UCA Server |
|---|---|---|---|
| AXS-4MMS | SISCO | Yes | Yes |
| BASE | GE | Yes | Yes |
| Average cost | $ 2,748.33 | ||
| Description | Vendor | |
|---|---|---|
| iSNMP | CIO Software | |
| SNMP-OPC Gateway | Obermeier Software | |
| Average cost | $1456 |
Operator Interfaces
Inherent in any comprehensive substation automation system design is the substation computer. System design analysis will reveal whether the substation computer is a critical component that must be reliable and highly available, or a trivial one that can be easily replaced. Again, this results in a decision regarding hardened versus commercial-grade computers. This analysis uses a noncritical computer, but if an industrial-grade computer must be used, expect a 5 to 10 times cost increase.
A substation computer has an operator interface, and if OPC-compliant (Object Linking & Embedding for Process Control), it can have an OPC client/server for UCA. Presently, there are two OPC servers for UCA. The capabilities of both products are similar in some areas but the pricing structure is different. (Table 5)
The utility also should consider the software to monitor the substation Ethernet network. Currently, there are two OPC servers available from different vendors that monitor Ethernet traffic using Simple Network Management Protocol (SNMP). (Table 6)
Substation Host
A substation host includes data concentrators, PLCs, communications processors and RTUs. Nine vendors offer 11 products of this type. These devices provide complex programming capabilities, analog inputs/outputs, digital inputs/outputs, protocol conversion for SCADA master protocols or serial device protocols, database connectivity, Web servers, internal security implementations, routing, communications monitoring, UCA client, UCA server and multiple communications ports. The utility should make equipment choice based on functionality more than cost because the choice impacts all aspects of system design. (Table 7)
Equipment Monitors
Equipment monitors exist for circuit breakers, transformers, LTC controllers, CT/PTs, batteries and bushings. An LTC monitor/controller from Beckwith Electric (Largo, Florida, U.S.) supports serial UCA, while one CT/PT/bushing monitor from On-Line Monitoring supports Ethernet UCA. These limited choices in equipment monitors will require converting the native protocol to UCA.
Testing
The utility should conduct testing in four separate stages: conformance testing, conceptual testing, factory acceptance testing (FAT) and site acceptance testing (SAT). Ultimately, system testing depends on the criteria established in the initial project design.
Conformance testing includes environmental conformance, which can be problematic for networking devices. Finding equipment that meets P1613 (IEEE Std. “Standard Environmental Requirements for Communications Networking Devices Installed in Electric Power Substations”), as well as existing corporate IT standards, will be difficult initially. Vendors can perform testing.
Conformance testing also includes communications testing. Testing protocol performance can be problematic, as the IEEE has discovered that there is no coherent communications modeling, terminology and communications test scenarios for the evaluation of communications networks involving substations. To help solve this problem, the IEEE has balloted C37.115 “Test Method for Use in the Evaluation of Message Communications Between Intelligent Electronic Devices in an Integrated Substation Protection, Control and Data Acquisition System.”
Conceptual testing includes testing device interoperability and system performance in a laboratory setting. If these tests are delayed until the FAT, or never completed, project risks increase and may result in increased costs. If this testing is completed successfully, it shortens the FAT and SAT requirements.
The FAT occurs once the system is assembled at the factory. With the previously discussed tests, the FAT becomes a relatively simple test that confirms the system point list. The FAT includes the typical panel checkouts and testing, which can be minimized with a complete UCA implementation. This analysis does not include this potential savings in FAT testing.
The SAT starts once the automation system is delivered on site and completely installed. The SAT is based upon the FAT with some additional tests. The SAT ensures that the system has been connected and installed properly, and all points are functioning as required. Since the SAT is based upon the FAT, the same cost savings achieved in the FAT are applicable to the SAT.
Construction
Panel fabrication costs typically include panel construction, equipment mounting, inter-panel wiring, intra-panel wiring, creation of wiring diagrams from schematics (optional) and panel testing. Most of these costs do not depend upon the protocol but on the amount of integration. The cost difference in panel fabrication results from using “soft contacts” to implement Generic Object Oriented Substation Event (GOOSE). The initial programming required in the devices for GOOSE messaging and software testing will be more than the cost reduction obtained from the initial panel fabrication cost reductions. Over several substations, the cost analysis may reveal that implementing GOOSE will result in panel fabrication cost savings.
| Description | Vendor | UCA Client | UCA Server |
|---|---|---|---|
| StationMANAGER/Substation Host | Telegyr | Yes | Yes |
| C30 | GE | Yes | Yes |
| D25 | GE | UD | Yes |
| D20 | GE | UD | Yes |
| 2030 | SEL | Yes | Yes |
| RT integration server | LiveData | Yes | Yes |
| e-terracomm control 3.0 | Alstom EMM | Yes | Yes |
| ePAQ-9100 | QEI | Yes | Yes |
| RTU-9100 | QEI | Yes | Yes |
| SOS supervisor | On-Line Monitoring | No | Yes |
| Minimum cost | 0 | ||
| Average cost | $2204 | ||
| Maximum cost | $10,000 | ||
| Description | Small | Medium | Large |
|---|---|---|---|
| Switch | $6279 | $6279 | $12,558 |
| Operator interface | $4204 | $4204 | $4204 |
| Substation host | $2204 | $2204 | $2204 |
| Relays | $3394 | $6788 | $13576 |
| Revenue metering | $0 | $0 | $0 |
| Design and engineering | $53,760 | $60,480 | $68,600 |
| Testing | $45,720 | $73,760 | $87,840 |
| Construction | ($27,500) | ($49,500) | ($93,500) |
| Maintenance, support and recurring costs | $0 | $0 | $0 |
| Phaseout | $0 | $0 | $0 |
| General management and support | $17,334 | $23,057 | $28,347 |
| Additional cost to add UCA | $105,395 | $127,272 | $123,829 |
Maintenance, Support, and Recurring Costs
Cyclical costs include equipment maintenance, equipment replacement, software and hardware support, as well as upgrades in hardware, firmware and software. Recurring costs also should include all training and facility/component maintenance, and all software licensing costs. Using equipment monitoring and self-diagnostic equipment reduces equipment maintenance and related costs. The costs to upgrade firmware, hardware and software need to include testing those upgrades. Recurring costs tend to ramp-up sharply during the first few substations, but after a few substations are commissioned for operation, the sustaining cost for operation should stabilize.
Phaseout
Phaseout costs are incurred when devices reach the end of their expected life. Life-cycle lengths can change the analysis results and justification, so it is important to consider typical life-cycle periods: software 1 to 5 years; computer hardware 1 to 5 years; network hardware 5 to 15 years; and physical plant 30 years.
The life cycle of each component needs to be considered separately and depends upon the technology in the device. For example, software and hardware based upon Microsoft products should be considered to have a five-year life. For other hardware and software, using the product warranty period can be indicative of an appropriate length.
Costs for Mythical Substations
As a simple example of a single substation cost analysis, look at Mythical Small, Medium and Large Substations. Table 8 shows the onetime cost at one substation. Over time, a comprehensive automation program will see significant cost reductions with some tasks. The cost analysis is based upon the following factors:
Minimum 16-port switch configuration.
Operator interface costs include OPC servers for UCA and SNMP.
Average cost for substation devices.
Metering data is brought back from relays.
Design and Engineering costs are based upon 14 tasks taking X amount of time for each Ethernet device. X is reduced across the number of devices: 8, 6 and 1.75 hours for each substation.
Testing costs include $5000 for conformance testing costs for each of the primary relay, backup relay, substation host and switch. Testing costs include an additional amount Y of testing task time for each serial and Ethernet device. Y is reduced across the number of devices: 8, 6 and 4 hours for each substation.
No reduction in relay-testing costs for the FAT and SAT.
Construction costs include the savings in panel fabrication. Each line and transformer is considered to take one full rack with cost savings of Z per rack using a full implementation of UCA. Z is reduced across the size of the substation: $1000, $900 and $800.
General management and support is 15% of the previous costs.
Average hourly labor rate taken as $80.
The cost of protocol conversion for serial backup relays, breaker monitors and transformer monitors is not included in the cost.
The analysis shows some important results for a single substation. First, the cost increase to add UCA devices to a substation is not driven by equipment costs as they account for only an average of 13% of the total increase. Second, testing costs average 39% of the total increase in costs. This significant cost could be reduced if vendors provide testing to acceptable standards. Third, the design and engineering costs average 35% of the total increase in cost. With additional substations, these costs will substantially reduce. And finally, construction savings average 31% of the total increase in costs and are substantially more than the material cost of adding UCA to the devices.
Conclusions
Many factors impact the cost of a UCA substation and make that cost a huge question mark. Costs depend upon substation type; utility experience with integration and automation technologies; development or revising of substation standards; substation retrofit or greenfield; training requirements; integration of legacy devices; SCADA master communications; the extent of integration and automation required; and testing. Unfortunately, the known costs are few, namely substation equipment hardware and software. Looking at the additional cost of adding UCA to a device does not address the substantial costs involved with changing to a new device standard that supports UCA.
Ultimately, the cost depends on the number of devices to be integrated and the number of substations where the design will be applied. As the number of devices and substations increases, the amount of construction, testing and other work also increases. As the number of substations increases, some costs will decrease: material, design, testing and training. The number of devices is an independent variable of substation type: some transmission substations will have fewer devices than some distribution substations. Ultimately, the number of devices drives the cost model.
Craig Preuss is a project engineer working in substation integration and automation in the power-delivery division at Black & Veatch Corp. (Overland Park, Kansas, U.S.).
preusscm@bv.com
Want to use this article? Click here for options!
© 2008 Penton Media Inc.
