Meter Data Analytics: Role of meters in discom operations

Role of meters in discom operations

Power distribution companies are increasingly deploying smart grid technology for improving their operational performance. As per global experts, the market is expected to expand significantly by 2018. This is mainly because this technology enables extensive monitoring of power systems and helps utilities to manage huge networks, installations and distribution equipment.

The pace of deployment of smart grid technology must be matched by utilities in terms of intelligent handling of metering data, extraction of the consequential results and implementation of best practices.

The utilities face several challenges ranging from theft and quality issues to outage management. Another issue that most global utilities are focusing on lowering the transmission and distribution (T&D) loss levels. Collecting data from meters and using it efficiently poses another key challenge.

Almost a decade ago, electromechanical meters were replaced with static meters in almost all Indian utilities under the Accelerated Power Development and Reforms Programme. However, most utilities still face high T&D losses and lack information about network quality and supply status.

Data collection parameters

The key function of consumer meters is to generate consumption data for billing and tariff purposes. However, meters generate a lot of useful data, which, when combined with defined logics, can be used to derive information about supply quality at the consumer end, installation quality, meter quality, load quality, electricity theft and technical losses.

Supply quality at consumer end

  • Outage management system (OMS): The data collected from meters may be categorised as offline data (historical) and online data. OMS combines call centres and distribution management system tools to identify and diagnose fault locations and further, to isolate and restore supply with the help of online data. In case of an outage, a geographic information system is employed to determine the location of consumer feeders and distribution transformers. A study of the variation in key factors of meter data provides information about the fault location.
  • Study of tail-end voltages: Information on tail-end voltages can significantly help in diagnosing the fault location. Online monitoring of tail-end voltage levels is required for integration with OMS, as it indicates supply availability. Based on a conditional analysis in combination with other factors such as distribution transformer loading and calling data from consumers, the fault location can be identified. Tail-end voltages can also provide information about technical loss levels and overall health of the network.
  • Power factor: The power factor is a measure of electrical efficiency of a power system and is useful in reducing network operating costs. Keeping track of consumers with low power factor and encouraging the installation of power factor correction equipment benefits not only consumers by reducing their electricity demand charges but also utilities by reducing line losses and enhancing system capacity.
  • Voltage unbalance: Voltage unbalance occurs when the root mean square line voltages on a polyphase system are not equal. Voltage unbalance can create a current unbalance 6-10 times the magnitude of the voltage unbalance. Electric supply systems should be designed and operated to limit the maximum voltage unbalance to 3 per cent when measured at the consumer meter end. Any unbalance above this limit is indicative of load imbalance, overloading or joints in service cables. A study of phase-wise voltage available in meters for an area-wise consumer base helps to find out whether a problem exists locally at the consumer end or at the higher distribution transformer level. It also helps to look out for theft, wherein a consumer may deliberately try to reduce the meter’s phase voltage, thus resulting in lower energy recording.
  • Supply availability: Meters provide various types of data including cumulative monthly power-off hours, cumulative daily power-off hours and long power-off durations. Analysis of such data helps the utility determine the reliability of power supply, especially in seasons of peak energy demand. The availability of supply may be extended to theft analysis by comparing data at the DT and consumer levels or between adjacent consumers along the same feeder. A mismatch among these is indicative of theft.

Installation quality

The quality of installation is judged using phasor diagrams and certain instantaneous parameters. Phasor diagrams help locate wrong wiring, abnormal phase association or loose connections. Similarly, persistent abnormal values of instantaneous voltages in the range of 0-400 volts indicate electricity theft or wrong wiring at the time of meter installation.

Meter health

Meters are equipped with certain diagnostic indicators like real-time clocks (RTCs). A repetitive disturbance in the RTC indicates failure of the battery inside the meter.

Technical loss estimation

  • Network stress: Voltage drop is directly proportional to the stress in the network and accounts for technical losses in the power distribution system. It is linked to the tail-end voltage method, wherein the consumer at the tail end of the feeder is recognised and the feeder’s voltage is then compared to the DT voltage. A drop in voltage levels signifies power loss in the system. Other causes of voltage drops are poor joints and terminations, hotspots, undersized conductors and overloading of lines.
  • Power factor planning: A low power factor leads to idle capacity in the network, an increase in the current level, higher voltage drops and high technical losses. Thus, it is important to monitor the power factor and encourage the installation of equipment to compensate for the same.
  • Network health: Network health at the feeder and transformer levels with respect to overloading, unbalancing and voltage fluctuations can be monitored on a routine basis for planning preventive and predictive actions. This, in turn, can result in a significant reduction in breakdown costs.

Load quality

A study of load survey of the consumer provides information on the following:

  • Peak load occurrence time and quantification: While designing any demand-side management scheme, it is important to refer to peak load events.
  • Load pattern: With utilities switching to time-based tariffs, the load survey study imparts vital load quality data at any time of the day. Historical data helps benchmark the expected load pattern for a particular consumer and their expected usage in the future.
  • Phase-wise load to identify unbalanced load: Load unbalancing causes higher technical losses and voltage drops in the network.
  • Load power factor: The power factor will depend on the equipment connected as load. All inductive loads have a lower power factor, mainly in commercial and industrial establishments, which needs capacitive compensation. A low power factor reduces the rated line and transformer capacity.
  • Harmonics: Harmonics are multiple frequency voltages and currents that appear on the power system as a result of non-linear electric loads that draw a larger current than required. Electronic equipment acts as a harmonic current generator that distorts current wave shapes and causes the distortion of voltage waveform, which then affects other connected loads such as motors, computers and power equipment. This may cause transformers and line conductors to overheat and lead to unacceptable downtime.

Electricity theft analysis

Utilities require extensive IT facilities to trace electricity theft. The analytic level is changing with time in order to improve the strike rate of theft identification.

  • Level 1: This is a preliminary level of theft analysis based on the consumption pattern. However, the strike rate is generally low and consumption trend alone cannot be treated as evidence of theft.
  • Level 2: A lot of information about customers can be obtained from other data sources such as consumption patterns of similar consumers. The biggest advantage is that it takes into consideration the variations in consumption due to common factors like weather and market demand.
  • Level 3: Several newly designed meters directly identify abnormalities and log events caused due to theft. Generally, such leads have a high strike rate, but can only identify theft according to a predefined method. Since all logged events are not theft, they need to be filtered based on logics.
  • Level 4: Apart from energy consumption and tamper event logging, meters also log instantaneous parameters such as voltage, current and power factor. In addition, they log power on/off data and represent the true electrical system through phasor diagrams.

Futuristic meters

The biggest drawback with the above analytic levels is that logic filters have to be defined, reviewed and amended beforehand to analyse the data and recognise consumers involved in the theft. In brief, it is the quality of logics and data filters that decides the strike rate. This analytic level is based on two basic principles. One principle is that the outcome of two events with similar trends is expected to be similar.  The other principle states that utilities hold huge banks of meter data, consumption data and secondary data information through four levels of analytics.

The working principle of this model involves identifying cases where theft was observed in the past, plotting data to identify the trend, filtering consumers showing a similar trend for a given set of parameters, recording outcome (success/failure) and using employee metering knowledge to refine the logics.

The biggest advantage of artificial intelligence is the self-improving nature of software. The aim is to develop a system from experience and continued analysis. The basic function of artificial intelligence is to use already confirmed theft or quality issue cases (called the test population) and subsequently identify common patterns using complex algorithms. These patterns are then used to recognise similar patterns in the target population. The target population is also checked for abnormalities in their electrical data. Accuracy and strike rate in the entire process are above 95 per cent. Upon establishing a final list of cases to be checked, the feedback can be fed into the system to improve accuracy in the future and ensure continual improvement.

Summing up

Installation of smart meters is not sufficient for revenue protection. Utilities should focus on utilising the meter information in a more productive manner. There is a need to manufacture analytical meters. Based on BSES Rajdhani Power Limited’s (BRPL) experience in the field, we recommend the following points to be considered for manufacturing analytics-friendly meters.

  • Grouping of meter data: There is a need to review data grouping systems based on analytical logics so that the data for a particular analysis is available readily.
  • Faster data acquisition: Steps must be taken to ensure the availability of desired information at a faster rate.
  • Local intelligence: Rather than passing raw data, the data should be processed locally and the information should be displayed. This will help avoid the transmission of bulk meter data.
  • Consumer application: The meter should be able to generate reports according to consumer requirements.
  • Standardisation of meter data management: BRPL has tried to develop a meter in which various theft logics are implemented. Therefore, rather than storing the meter data to analyse the theft, the meter will analyse the data as per defined logics and display the outcome. Thus, only the outcome will need to be transmitted to the server.