Get Smart

Making Use of Smart Grid Data

(We set out for smart grid, we are now at verbose grid.)

Smart-Grid is built on the premise that more distribution information is available and that utilities have or will have “automatic” and/or “remote” control of the distribution system equipment. With the right information and tools system equipment can be used to dramatically improve reliability, reduce system peak and reduce the cost to serve electricity.

Load Data:

Commercial and Industrial (C & I) customers have had interval load meters for some time now. Interval load meters are used to employ time of use and peak demand billing to commercial and industrial customers. Residential customers on the other hand have had very little load data other than a monthly kWh reading which gives no information about demand. With the advent of Automatic Meter Reading (AMR) and Advanced Metering Infrastructure (AMI) systems more demand load data is becoming available for the entire distribution system. The justifications for these systems have primarily been based on reduction of meter reading staff and time of use billing. Many of the meters send “counts” which represent kWh. Typically no pf or kVA information is being measured or collected for residential customers. However, kWh information on 15 min intervals is light years ahead of the preexisting data available for distribution loads. TLM systems have historically been used to estimate demand loads for distribution customers. However, TLM systems typically have errors in the 20-30% range.

Power factor (pf) is not always measured by AMI/AMR systems at residential locations. Having pf allows for greater understanding of the loads that are on the distribution system. Currently kWh counts are brought back at in 15 min intervals. Having time synchronized Voltage, kW, and kVar readings would allow for the calculation of constant power, constant current and constant impedance loads. If each meter were to have a GPS receiver and send its location it would aid in the calculation of secondary impedance and for locating customers that are out of power. Voltage, kW, and kVar readings at distribution transformers can be used for the detection of power theft, figuring out how much power unmetered street lights are consuming, the calculation of secondary impedance, and verification of the power-flow model.

A huge benefit to the verification of the electrical model and understanding of the layout of the electrical system would be to have customer meters and/or distribution transformer meters to return what phase they are currently connected to. This could be simply accomplished by putting a different very low frequency signal on the low side substation bus for each phase which would be picked up by the distribution customer meter or distribution transformer meter and returned with the load readings. The challenge will be to determine lighting phase from power phase of open delta connected transformers. Another challenge will be to prevent phase signal cross over from one phase to another at locations where there is any kind of phase to phase connection.

Station feeder and transformer load data were historically measured and recorded using circle charts. Circle charts lacked precision and were only collected periodically. Over the past 30 years or so, the measurement transducers that were feeding load data to the circle charts were connected to SCADA systems and the data was sent back to the control center via the SCADA system. In many cases the load data has been captured and stored in a data historian and accessed from desktop computer clients. The feeder and substation load data has been extremely valuable to planning engineers in their efforts to plan and layout the power system. The data is also incredibly valuable to system operators when running the power system. AMR and AMI data could be used in similar ways as the SCADA data to better understand loading on the distribution system.

Collectively these sources of load data provide the requisite load information for power flow studies which are used to gain better understanding of what is actually happening on the power system. A significant challenge to utilities has been penetration of Supervisory Control and Data Acquisition (SCADA) and AMI. It is not unusual for a utility to have 100% penetration of SCADA at substations within major cities, but then to have far less penetration of SCADA at substations outside the major cities. Installing SCADA can be an expensive undertaking and difficult to fund. With the system not fully visible with SCADA, it can pose challenges for system planners and operators to have a full view of what is happening on the power system.

The same problems face AMI systems at utilities. The challenge of course is financial expense and justification. There is a much higher cost to the utility that is difficult to quantify that being lack of knowledge of what is happening on the distribution system. As a result of the lack of relevant load data is that the distribution system had historically been over built to allow a factor of safety. Overbuilding is expensive. Once more, in recent years utilities have felt much greater financial pressure a lot of which goes back to deregulation and a more competitive environment for utilities. As a result, utilities have attempted to push their existing systems and equipment harder to squeeze every last bit of value out of what they have. However, without proper data to support the heavier burden being placed on the equipment there tends to be more unexpected over loads and equipment failures which can be quite expensive from both a financial and customer satisfaction points of view. Modern power systems have tremendous need for more and clearer information.

A utility that has 100% penetration of SCADA, 100% penetration of AMI for residential and 100% penetration of interval load meters for C & I customers is well suited for the load data needs of Smart-Grid. The next hurdle is to collect the load data on a regular basis, organize the data, validate and clean up the data. There will inevitably be missed reads, communication errors, wrong multipliers for data etc. For this there are Meter Data Management Systems (MDMS) which collect organize and validate data. In situations where there are problems with the data, missed/inaccurate reads, the MDMS estimate the missing or wrong load values. With the data sources in place and feeding and MDMS there is a strong foundation of load data for smart-grid. One of the challenges will be to process such massive amounts of data. Planners have historically used hourly load data from the SCADA system also called 8760 data which is a reference to the number of hours in the year. Even with 8760 data there are far too many data readings for humans to interpret. This problem is further complicated by AMI data having 35,000 readings per year. Techniques such as Apportioned-Time Intermediate and Long-term Data Analysis (ATILDA) help convert these piles of data into understandable information that humans can use.

System Configuration Data:

Knowledge of system configuration is critical for a utility. This information is required in order for planners and operators to plan and operate the system. In the past, and in some cases the present, utilities used various system maps. These maps may be organized into distribution maps, transmission maps, underground maps, overhead maps, etc. Utilities have been digitizing maps in various forms over the past couple of decades. In some cases the maps were simply scanned and digitally pasted together. In other cases the maps may have been converted or redrawn to a CAD format. With the increasing popularity of Geographic Information Systems (GIS) and it's customization to the needs of electric utilities, many utilities have built models of their electric systems into various GIS packages.

Models within power flow software are often not the up-to-date with the configuration that is in the field. The distribution system is very dynamic. Switches are opened and closed on daily bases which reconfigure the system. In the past switching operations may have been recorded on paper maps or in switching logs. It is important to realize that it was not unusual for switching operation to not be recorded at all particularly during periods of storms and storm restoration. Today many utilities track switching operations using Distribution Management System (DMS) an Outage Management System (OMS). For utilities that have and use a DMS/OMS a digital record of the system configuration exists and can sometimes be used for power flow studies. For accurate and meaningful power flow studies current switching information is important and needs to be in an electronic format to be useful for moment to moment power flow studies. There have been a few cases of utilities tying a power-flow tool to the load data and/or the OMS/DMS data. As we move forward as an industry it will be commonplace for utilities to have the GIS location information, the OMS switching states, the AMI/SCADA loads all tied into a real-time load flow tool to continuously monitor the state of the distribution system.

The switches that are SCADA controlled can be tied to the DMS/OMS system. Therefore the SCADA switches will for the most part have accurate state information within the DMS/OMS system. On the other hand, for switches that are not SCADA controlled, a vast majority of switches on the distribution system will likely have their state manually entered into the DMS/OMS system by system operators. Under normal operating conditions it can be presumed that the switch status information will be maintained accurately. However, under abnormal operating conditions such as those that exist during a storm it is likely that some switch status information will become corrupt. Therefore, there is tremendous need to have a system to verify the switch status of non-SCADA switches. Manual verification is possible, but can be a substantial drain on manpower. There is a need to build a software tool that is capable of comparing expected power flow results against actual system readings. A tool that is related to a state estimator for the transmission system could be built to perform this task. As it stands the fundamentals required to build such a tool exist, but thus far no known system exists. The Topology Error Identification system would monitor near real-time loads and "known" system topology to estimate the "actual" system topology and identify which switches are in an abnormal state.

In order for utilities to take full advantage of AMI data it is important for there to be an accurate three phase system model all the way to the customer. There are a couple of short comings with respect to the data available to support such a model. It is not unusual for utilities to have inaccurate data as to which phase a customer or lateral is actually on. During the system peak it is not unusual for trouble-men to be sent out to move a lateral from one phase to another to address either a loading of phase balance problem. These activities were not always accurately recorded in the past and therefore the records are not always entirely accurate. For this reason, it would be useful for the meter at the customer premises to send back the phase designation along with the load readings.

The other issue concerning data to support a model to the customer is that utilities typically do not have accurate records concerning the last 100 feet of wire to the customer meter (the secondary). This detail is frequently left to the installing electrician and is often not recorded or included in any electrical system models. The big problem here is that the secondary is where the greatest voltage drop and losses are seen, which can be quite dramatic compared to any other part of the electrical system. While it would be impractical to send people out to measure every secondary on the system, it may be a comparatively trivial matter to include accurate time synched readings for kW, kVar, and Voltage in the customer meter as well as meters on the distribution transformers. With this data the impedance of the secondary services can be calculated easily enough.

Power Flow Software:

Planners and Operators frequently require the ability to run power flow studies. While GIS typically has a model of the physical layout of the electrical system, it does not always have the electrical characteristics of the electrical equipment in the system which is required for running power flow studies. Therefore, many utilities have some form of power flow software in which a model of the distribution system exists. The system configuration and electrical characteristics are critical for accurate power flow studies.

Historically power flow software has used a single-phase equivalent model. Approximating a model as a single-phase equivalent model is a reasonable approximation for Transmission systems that tend to be very phase balanced. Distribution systems on the other hand tend to be very imbalanced. In fact, one of the goals of smart-grid is to have the tools to better phase balance the distribution system. Some of the power flow software packages that are marketed for use on distribution system models use three-phase models. These packages typically use a backward-forward sweep power flow algorithm. The backward-forward sweep algorithm is well suited for radial systems, but is not as well suited for networked distribution or distribution systems that contain distributed generation. The current injection power flow algorithm is far better suited for systems that have radial and networked topologies and is also better suited for systems with distributed generation. Unfortunately very few packages use the current injection algorithm. Another challenge for most power-flow software packages is the ability to handle combined transmission and distribution models. Almost all packages concentrate on either distribution or transmission models.

With the increasing prevalence of solar, wind, and micro-turbine generation (distributed generation) on the distribution system it is becoming increasingly important to use a power flow package that can handle such situations and be able to analyze power flows that go both directions on the distribution system. There has been a long standing need to handle networked systems; many large cities have secondary networks in their central business districts and some utilities use networked primary distribution feeders.

Distribution system models for power flow software tend to be very large compared to transmission models as a result of having lots of equipment and lots of laterals within fairly small geographic areas. The cost of large complex models is processing time. Even with the speed of modern computers, the processing time for models that cover a seemingly small number of electric customers take minutes in order to converge. While there are some packages that can handle very large models without dramatic simplification techniques and still converge in seconds or milliseconds, these packages are the exception and not the rule.

Another challenge of contemporary power flow software is that they are typically populated with a static model of the distribution system. The loads and/or configuration of the system of a specific time is loaded into the power flow package. The specific times typically include the annual system summer peak, winter peak and perhaps the quarterly or monthly peaks. In most cases the system configuration is not updated in the model to coincide with the loads. Therefore, if the quarterly loads are updated in the model but the current system switching configuration is not, the results will be flawed. Once more, the loads that the software is populated with are typically the substation feeder loads. The feeder load is allocated down the feeder to all the distribution transformers. This process usually involves proportionally scaling all loads on the feeder by the ratio of the new feeder load divided by the old feeder load. The ratio is presumably close to what the new customer load is, but is usually not incredibly accurate.

Some utilities have started loading data to the distribution system models with load data from a meter data management system (MDMS) which is large step forward from simply allocating substation feeder readings down the feeder. The MDMS is the warehouse system of AMI customer loads. The MDMS validates meter readings and attempts to estimate missing or wrong reading values. Populating the power flow model with loads out of the MDMS leads to improved accuracy in the results, but the issue of system switching configuration at the time of the load readings is still an issue. To address the issue of system configuration it is necessary to tie the switching data from the DMS/OMS system to the power flow model. Once both the MDMS and the DMS/OMS data are time synchronized and imported into the power flow model truly accurate power flow results can be achieved.

Once the MDMS load data and DMS/OMS data are directly tied to a power flow engine regular power flow studies can be run and archived. The results of the regular power flow studies can be used for several purposes such as having historical records of loading on sections of feeder (sectional loading). While section loads can be calculated at any time having a historical archive of the sectional loading can be incredibly useful for operators, planners, construction and maintenance personnel. For example during a storm an operator can use historical section loads when restoring service to sections of faulted feeder. Using the historical trend over the previous week or two will allow the operator to know the max load for the section in the very recent past. This allows the operator to maximize the load that is picked up without running the risk of overloading the feeder that the operator is using for temporary restoration.

Distribution Equipment:

The equipment that has the potential for control include Generators, switching devices, capacitor banks, tap changing regulators, static VAR devices, Synchronous Condensers, DC converters, energy storage devices, loads, distributed generators and inverters. The control of this equipment can have a dramatic impact on the performance characteristics of the power system. With respect to the distribution system, the devices that have the greatest potential for impact include switching devices, capacitors, loads, distributed generators and tap changing regulators. With greater penetration solar and wind generation as well as storage devices and their associated inverters will have a substantial impact on distribution system performance.

At this point in time it is not unusual to have SCADA control over various types of switching devices. It is also not uncommon to find different types of automatic and remote control over tap changing regulators and capacitor banks. Regulators and capacitors are very important with respect to controlling voltage and controlling system losses which are useful when trying to control the cost to serve electricity. Various efforts have been implemented to try to leverage the control of these devices in order to reduce the cost to serve electricity. One significant effort has been Conservation Voltage Reduction in which the voltage is reduced in an effort to reduce total power consumption. These techniques have shown benefit on the distribution system to reduce total consumption at times of system peak demand.

With respect to reliability, SCADA controlled switching devices are being used in automated system restoration schemes. SCADA controlled switches are strategically placed to allow them to be used at times of system interruption to isolate faulted sections of feeder. These systems are typically referred to as Distribution Automation (DA) schemes or Automatic Service Restoration (ASR) schemes. There are several issues with these types of schemes a couple of which include protection coordination and limiting system configurability.

Distribution system reclosing devises when in series with each other on the same feeder backbone need to have coordinated protection settings if they are to be used as reclosing devises as well as being used as switch points with a DA or ASR scheme. In this case the tendency in the past has been that synchronizing more than two or three reclosing devises with the feeder circuit breaker can prove to be quite challenging. Therefore, if more devises are to be used it is likely that many of them will strictly be used as SCADA controlled, motor operated switches, and/or they will require dynamic protection settings which means that logic will need to be developed in order to “automatically design protection settings”. Given that the distribution system tends to be fairly simple from protection point of view this may be within the realm of reason. Furthermore, implementing a setting update at the time ASR system reconfigures the system should not be a huge leap, but will require doing.

Placing reclosing devises arbitrarily on the system limits the configurability of the system by putting a fixed break point in the system. Planners need the flexibility to put new break points (switches) in the system in order to move load from a feeder that is forecasted to be overloaded to a feeder that has excess capacity. This activity often moves the “mid-point” of the feeder which occasionally forces the planner to relocate the mid-circuit reclosing devise or find an alternate solution to fixing the overload. Adding a switch to move load is far and away the most cost effective solution to dealing with overloads. When a reclosing devise is in the way the cost of a $3000 project can suddenly increase tenfold. With the advent of DA and ASR there are more reclosing devises on the feeder which are intended to shift load under emergency conditions presumably to last a matter of hours or days. Planners need the ability to move load and have it last for years. 

DA and ASR systems require far more than one mid-circuit reclosing devise which places much greater limitations on the flexibility of system configuration and thereby increases planning project costs. Both emergency and permanent system configuration are needed and must be balanced against each other and an overall cost must be found in order to weigh one need against the other. For this reason it is necessary to have a tool that can find an ideal permanent system configuration that takes into account cost, capacity planning, contingency planning, and system growth. Methods and tools exist to look at system growth such as spatial load forecasting which have been incorporated into GIS systems. Methods and tools to estimate cost have existed for some time. Although limited, tools exist and are in development for planning ideal system configuration for capacity planning. The challenge will be to incorporate those tools with a tool that plans for contingency. While this is within the realm of possibility a tool of this nature still requires development.

Distribution System Operation and Planning:

The interaction of the operations department and the planning department will continue to grow closer and closer as Smart Grid continues to progress. Many utilities have already started having these two departments working more closely together than what has traditionally been the case. Smart Grid will put great stress on both departments to keep the system up and running. One of the tools that they are both and will both be heavily dependent on is power flow software. While power flow software will tell the user what is presently going on or what has happened it does not do much to tell the user what needs to be done in order to improve things. This is where a true 3-phase AC Optimal Power Flow (OPF) software comes in. OPF software has the ability to tell the user where the stresses are in the system. DC OPF software is already being used in order to find pricing for electricity markets. The DC OPF engine is unable to handle losses system losses or reactive power flows. Therefore, an AC OPF is the only tool that can truly help in distribution system planning and operations.

The AC OPF has the ability to tell the user where the system is stressed, which means that it will pinpoint where a voltage collapse will originate. The AC OPF will tell the operator which generator is most important to start up or which capacitor bank is most important to turn on in order to stabilize the power system. An AC OPF can tell the planner where new resources need to be located and ranks all locations as to their need for new resources such as capacitors or generators. The operator can use the tool in order to see what setting a voltage regulator or LTC needs to be at in order to minimize losses in the system and get greater efficiency. The Curtailment Service Provider (CSP) can use the AC OPF in order to get the maximum benefit from a Demand Response program. It has been shown that it is possible for a CSP to reduce generation by 1.3 MW by curtailing 1 MW of load. These reasons only scratch the surface of what an AC OPF can do for planning and operating the distribution system.

If the distribution system begins to be operated based on the recommendations of OPF tools then there is a very good chance that switches will be operated more frequently and other equipment will be exercised at greater cycles than in the past. However, electrical equipment is not designed to be continuously switched on and off. The operating of switches wears them out. For this reason it will be necessary to have tools that take into account how often or that last time that a piece of equipment has been operated. Energy markets use a tool that plans the best time to turn on a generator and when is the best time to turn it off referred to as unit commitment. The unit commitment algorithm could be customized to take into account the cycling of different pieces of power system equipment so as to not over work and wear out the equipment.

There is tremendous potential for Smart Grid technologies to improve reliability, reduce system peak and reduce the cost to serve. The techniques outlined in this discussion are some of the components that will be of high importance in the coming years as Smart Grid gains more traction. Many of these tools do not yet exist for the distribution system. However, many of these techniques and tools do exist and have been done on the transmission system. In theory, the techniques, tools and practices used on transmission systems can be further developed and extended to meet the needs of the distribution system.