Because smart meters are attached to homes and businesses, the most obvious usage of the data that they collect is understanding customer power usage trends. But understanding customer usage is far from the only way to use smart meter readings. With the proper analysis, smart meter readings can yield insights into the performance of the electrical grid as a whole.

In determining network health, it is as important to analyze when smart meters are not responding as it is to analyze their actual signal. Lack of response from a group of meters over a period of hours or days could indicate an outage event. It’s common for utility companies to have outage management systems that keep track of outages both through smart meter data and reported outage information. However, the information that can be gathered about network performance does not stop with recording simple outage information. Careful analysis of smart meter data can yield more subtle information about problems on a power network.

To understand how smart meters interact with the grid as a whole, it’s necessary to understand how they transmit their readings to the utility. A couple of solutions have been employed to transmit smart meter data to the utility for analysis. One is to use wireless technology or existing cellular networks to enable the meters to communicate with each other as well as with central locations. Another method is to use existing power cable infrastructure to carry information from the smart meters using a system referred to as a power line communication (PLC) system. By modulating the existing carrier signal, sent along existing power lines it is possible communicate information in both directions from the smart meters to a central location. In locations where a PLC is in use the signals transmitted to and from the smart meters are interacting with the grid as a whole and there is the potential to use those signals to understand grid health.

Power line communication systems have been used to transmit data over power lines for over a century. The first patent suggesting a power line communication system was filed in 1897.1 In Japan a PLC system was implemented in the 1920’s and was used to carry voice communication along power lines.2 More recently, trials have been done in several countries, testing the use of of power line communication systems for carrying broadband internet.

Because of their integration with power line systems, PLC communication systems are a natural choice for transmission of smart meter data. However as power lines were not designed for transmitting information, there are challenges associated with using them to transmit information. Transmission channel characteristics can vary widely due to changes in power load along the line which can lead to signal attenuation or distortion. In the context of smart meter data, noise within the system or power line disruptions can cause interruptions in smart meter data transmissions. These interruptions, called blinks, result in the loss of some smart meter information, but they also provide information about disruptions along the power system.

Although interaction of communication systems with the power transmitted along the line can be problematic, it also yields some intrinsic advantages in that it can provide insight into the health of the smart grid. Particularly, if communication from a certain set of smart meters is interrupted it can provide a clue as to mechanical problems along the network, even if those problems haven’t become severe enough to cause an outage. For instance, if all meters downstream of a particular transformer experience a blink, it could be an indication that there is a problem with the transformer. Likewise, if all meters downstream of a certain line segment experience a blink, it could indicate that something such as a tree limb is interfering with the line. Conversely, if a blink occurs in an isolated meter, it’s reasonable to guess that the problem occurred with an individual meter. In more complicated cases in which it’s not clear that a set of blinks originate from a certain piece of equipment, machine learning can be useful in determining the cause of the event.

In some cases the timing of blinks in integral in determining a cause. Differences in the number of meters that experience a blink may be due to small differences in when the meters record a reading. If the number meters blinking changes during an event the change in blink patterns may also give important clues as to the source of a problem. Identifying equipment failures that have caused similar blink patterns can be useful in identifying the possible cause of a set of blinks.

GridCure’s blink module is built to quickly identify which piece of equipment and section of line is most likely to be responsible for a set of blinks. It also keeps track of patterns in blink evolution that might not be completely obvious and identifies previous causes of similar blink events. Using subtle information in smart meter data, GridCure’s blink module may be able to point to network problems before they cause an outage event.

1 Marumo, N. Simultaneous transmission and reception in radio telephony. Proc. Inst. Radio Eng. 1920, 8, 199–219.

2. Routin, J.; Brown, C.E.L. Power Line Signalling Electricity Meters. British Patent 24833, 1897.