Smart Metering PLC Network Monitoring, Analysis, and Grid Conditioning
Introduction
The capacity to remotely access and regulate utility meters is the foundation of smart metering. Meter reading, connection/disconnection, and delivering vital operational information on the electrical grid (transformer load, power consumption per LV feeder, historical consumption patterns, and so on) are typical procedures.
Smart metering, also known as Automatic Meter Reading (AMR), improves overall grid control and optimizes energy consumption management (for example, through tariffs based on known, real-time use and maximum energy consumption limiting at specific periods). Infrastructure control and network integration are the two most important features of smart metering. Smart metering programs are now available in several utilities across the world. The deployment is encouraged by utilities' interests or as a result of government mandates.
Power Line Communications (PLC) is a popular and well-suited technology for smart metering. For example, refers to the use of PLC technologies in AMR systems for operational purposes, explains different PLC applications (not just for utility-related purposes), and addresses the evolution of smart metering into smart grids. AMR systems make considerable use of non-PLC technology. They are said to lack the natural interaction with the electrical grid provided by PLC. PLC approaches have a long history in utilities. Nonetheless, only when open and royalty-free system standards for narrowband PLC solutions were published was the goal of having widely acknowledged, cost-effective, and field-deployable PLC technologies achieved.
Power line Intelligent Metering Evolution (PRIME) is one such technology that is being widely used by Iberdrola in Spain and is being explored globally by standardization groups such as IEEE, ITU-T, CENELEC, or IEC. PRIME uses Orthogonal Frequency Division Multiplexing (OFDM) and was designed to operate in the CENELEC A-band (42 to 90 kHz range), with adaptive physical layer transmission capabilities of up to 120 kbps and a Media Access Control (MAC) layer that ensures the configuration of highly available "plug and play" telecommunications networks.
For decades, utilities have been waiting for real-time PLC systems on LV networks. Most contemporary systems are only capable of receiving meter readings once per few days, hence they cannot be considered real time. As a result, most AMR systems are unable to provide real-time grid analysis. These real-time data enable optimal grid operation and give instantaneous information on supply points fed by secondary substations (SS). This information might be used to identify cable cuts (faults), regulate voltage, detect meter manipulation, and provide LV monitoring capabilities comparable to those found in medium and high voltages.
Smart metering networks must give communications access to millions of points of supply (meters) that are linked to tens of thousands or perhaps hundreds of thousands of SSs. Because of the historical grid evolution, the interconnection of such networks (e.g., which transformer feeds a certain meter, which path each LV cable takes, etc.) is often unknown. This is a significant difficulty for smart metering systems.
PRIME is a PLC technology that gives smart metering networks real-time capabilities. Real-time networks need monitoring processes that may evaluate the technology's performance in various deployment settings. Monitoring mechanisms for narrowband PLC networks do not exist. The monitoring procedure proposed in this paper has no precedent in previous PLC technologies (as opposed to traditional telecommunication technologies such as SONET/SDH transmission, radio communications, or Ethernet switching, which have been well developed in standardization bodies such as the International Telecommunication Union (ITU) and the Institute of Electrical and Electronics Engineers -IEEE).
The reason for this novelty is that no narrowband PLC technology has previously been able to provide such real-time characteristics (for example, the "availability" concept in other PLC technologies is based on the ability to access a meter once in a certain period of time that can be measured in days or months). Furthermore, once the monitoring process is completed, information must be compiled in such a way that the results can be analyzed, taking into account relevant LV grid data such as distance to the SS, electrical grid hierarchy, and so on (this type of data can be derived from the basic data in Table 2).
This analysis is important for a variety of reasons, including the operational information extracted from the results (e.g., SS connectivity, and LV line and phase identification), as well as other PLC network-related aspects, such as the evaluation and comparison of different solutions provided by different manufacturers delivering PRIME interoperable products, and the validation of the criteria used for field deployment options (injection at the SS level [26]).
PLC Technology System Aspects for Smart Metering
The concentrator is a critical system component of smart metering PLC networks. The concentrator is an application level element positioned at the SS that focuses readings acquired from LV grid meters.
Concentrators are made up of two sections. One portion is in charge of the metering application, while the other is in charge of telecommunications (PLC in the example under consideration in this work). The communications part of the concentrator is referred to as the base node (BN) in PRIME, and a subnetwork is defined as a group of service nodes (SN's; meters also have two parts, one in charge of metering and another known as SN, in charge of PRIME PLC communications) that are dependent on a BN (typically in an SS). From a telecommunications standpoint, we will refer to BNs and SNs, and from a metering standpoint, we will refer to concentrators and meters, respectively.
The PRIME subnetwork serves as the message exchange domain via which BNs and SNs communicate. The BN and SN construct a "topology," or the structure of telecommunications dependencies built among the various meters (their PLC modems), in order to connect with the concentrator (its PLC modem). With direct physical layer connection between the concentrator and the meter, not all devices narrowband PLC networks can be accessible from the concentrator. In various PLC technologies, the notion of "repeater" exists in order to design structures that can execute PLC signal repetition, allowing all meters to reach the concentrator. The notion of recurrence is included in PLC modems in PRIME technology, just as a specific state in the state machine of all devices.
When the network senses that a meter need assistance in reaching the concentrator, distinct SNs immediately start the process of possibly becoming repeaters (switches) in the network. With multiple applicants, only one will be authorized by the BN to serve as a repeater, allowing additional meters to connect to the BN.
PRIME's topology is dynamic and self-configurable, and within the subnetwork, some meters (their modems) manage to behave as switches (the "repeaters," in PRIME terminology), providing communications access to other SN's (meters) that may also behave as switches or simply terminal devices. Figure 1 depicts a PRIME PLC technology state machine.
Figure 1 depicts the PRIME Service Node (SN) state system.
Implications of Smart Metering PLC Networks for Low Voltage (LV) Grid
This paper's methodology focuses on system level and electricity-related characteristics. This PLC system analysis avoids utilizing most of the low-level specifics of the power grid (kind of wires, particular sounds, etc.) that greatly outnumber the grid's current available information.
The electrical grid, the PLC network, the metering application, and/or the electronics of the devices all contribute to operational issues in smart metering networks. PRIME nodes (meters) that do not communicate in the SS (individually, whole feeders, groups of feeders, full LV panel, full medium voltage/low voltage transformer, etc.) are examples of non-performing devices. The causes of a communication breakdown might vary greatly depending on the situation, and to make matters worse, they are not always immediately observable by typical operation and maintenance activities.
The monitoring stage will offer data on two key concepts: availability and stability, at both the subnetwork and subnetwork element levels. The analysis step will be based on the association of those ideas with the electrical grid properties where they occur.
The performance of a PRIME subnetwork is described by two main notions. The first term is "availability," which is defined as the amount of time (relative to total time) that a device is available in the subnetwork. The availability of a subnetwork is connected to the statistical values of the subnetwork components as a group. The second term is "stability," which is defined as a subnetwork element's capacity to remain unconnected. The capacity to maintain a steady number of registered items is connected to the subnetwork. The metrics that indicate the performance of the subnetwork and each of its constituents in terms of communications and applications will be the materialization of these notions in various quantifiable characteristics.
Methodology for PLC Network Monitoring and Analysis Proposed
The proposed approach for overcoming the challenges discussed in the preceding sections will give data for evaluating the performance of each PRIME subnetwork. There are two aspects to the technique.
The first step (monitoring) collects data and analyzes the findings to provide parameters that offer information on network performance, both at the communications (PLC) and application levels (metering).
The second step (analysis) examines parameters (PLC and metering; see Appendix for details) and their relationship to electrical grid data. Only when the results are connected with the physical parameters of the LV grid will the study be richer and more thorough. Although exact knowledge of LV physical data is rarely achievable in LV electrical networks, a minimal set of data (see Section 4.2) is necessary to evaluate the findings.
Stage of Monitoring
On the monitoring stage, the approach is based on purposefully straining the communication network, pushing the communication capacity to its limits by constant polling of meter data from the concentrator. Because PLC is the technology being studied, the minimum monitoring duration is one week. This length is a compromise to account for the fact that noise in the LV grid, for example, is highly dependent on network load, which is highly variable on date, time, and individual conditions. After this time, the PRIME subnetwork will resume normal operations.
The monitoring stage is in charge of two elements of the system's behavior. First, it monitors the PLC technology's behavior. The behavior of the PLC technology reveals how the network architecture develops over time. It is necessary to consider SN dependencies, SN behavior (switch, terminal, or any other non-functional state), and any timing components connected to those nodes. PRIME BN must be able to detect every single change in the subnetwork in real time (SN disconnect, terminal promotion to switch status, and so on).
Second, because metering is an application layer that must be monitored, a continuous flow of application data is established to study the overall system's activity. In non-stop loops, the concentrator is configured with cycles that query every element linked to it. The number of nodes in the network, the complexity of the network, the amount of transactions, and the size of the meter readings in each read attempt will all influence the findings collected and recorded from this application data stress of the network.
Concentrators, both in the metering and communications portions (the BN), contain the information needed to monitor the PLC subnetworks' activities. These elements must be able to log application information relevant to network status from both the application data and the PLC control information sides.
The components that must be documented for PLC topology network behavior analysis, as well as the application level, are summarized in Table 1. Because of the volatile nature of power line situations, PLC networks continually modify their topological configuration. The goal of these ongoing adjustments is to enable the BN to access the meters engaged in their subnetwork. Throughout the week, these changes and accompanying information must be recorded in real time, alongside the metering application data obtained through system polls.
Table 1 shows the application layer parameters that are necessary for performance monitoring.
To associate both groups of parameters and outcomes, the meter number and MAC address parameters in Table 1 must be matched. As meters are identifiable in utility databases, PRIME performance results may be immediately correlated with the meters and their location in an electrical system (address in a city).
Stage of Analysis
Data processing can identify the behavior of each individual meter (application level) and SN (PRIME communication level) after the results are gathered. The connection of the results with the physical aspects of the power network can assist to better understand the causes influencing the outcomes and the regions where problems may exist. Table 2 contains a list of meter-related data in the grid that may be used to understand the performance of PLC systems.
As an example, using the referred table, we may determine if the observed behavior is shared by one meter concentration, one of the phases, the entire electrical line, or the entire LV panel or transformer at the SS. Location and distance, single or three phase meters if three phase injection is utilized, and other factors will assist explain differing actions and discover database issues.
Table 2 shows meter and electrical grid data for PLC analysis.
Performance Evaluation of Smart Metering PLC Networks
Table 3 contains the bare minimum of concepts to consider when developing relevant criteria for a smart metering PLC implementation study. They mix the global network performance perspective with that of individual elements, and they also correlate meter and SN behavior.
The concepts in Table 3 can be stated using various numeric parameters and/or graphics, which can increase the interpretability of the derived parameter.
The ideas are described in the following subsections, with derived parameters specified in the Appendix. Each of these characteristics gives unique information on features of smart metering systems that must be regulated.
Table 3 shows the concepts required for the study of Smart Metering PLC networks.
Note 1: C = Communications; A = Application (Metering).
Evolution of Subnetworks
When operating a subnetwork, this collection of parameters gives information on the network view that both the concentrator and the BN (application and communications, respectively) have.
Stability of Topology
The amount of units registered in a certain concentrator in the SS determines the stability of the subnetwork architecture (see Figure 3). This setting is connected to PRIME subnetwork upkeep. It does not take into account application data.
Evolution of Application Data
Smart metering systems typically have a master-slave architecture with one or more polling cycles. When continuous polling is enabled, the number of registered nodes at the start of each cycle defines the number of cycles (read attempts) and the duration of each cycle.
Figure 3 depicts the progression of registered PRIME nodes in a real-world test subnetwork with 58 SNs over the course of two days.
PRIME Nodes individually
These metrics are part of a group whose significance is connected to how the meter functions as an SN, or a communications device.
Combined Total State Duration
These characteristics are based on the total time spent in each relevant state of the network, including the terminal, switch, and any other non-operational states. When doing PRIME analysis, these factors are important.
Total Amount of PRIME Level Disconnections
The cumulative number of changes from the registered state to the non-registered (disconnected) state is described by this parameter.
Availability of Subnetworks
Information about the viability of access to any SN (PRIME communications) or meter (applications perspective) in the subnetwork is provided by this collection of parameters.
Prime Time Access
Regarding the PRIME state machine, availability is determined by comparing the proportion of time that a node (terminal, switch) spends in an available state to the overall evaluation/observation time (see Figure 4).
Meter Readiness
Measuring meter availability is the proportional number of successful read attempts while taking the results of cycle testing into account. Non-successful efforts are those that fell short or were never made because the state was not registered.
Figure 4 shows a real-world field test subnetwork using a sample PRIME availability concept. In the position shown by the caption, each bar corresponds to one SN.
Similarity in Availability
This characteristic consists of the PRIME nodes' (communications level) relative availabilities and related meters.
The availability of PRIME Node and Meter level Subnetworks
Total average availability data taking into account every component in the impact zone of the subnetwork/concentrator.
Topology of a subnetwork
The PRIME network's setup and development in connection to its goal of serving as a communications network for smart meters will be described by these factors.
The majority of stable states
For each PRIME SN, two values will be evaluated: the maximum individual time spent in a state and the maximum individual time added to the total time even when the same state occurs at different times during the monitoring period, taking into account topology dependencies up to the point where the BN is considered a component of the state. The first figure would be 16 hours as terminal, and the second would represent 20 (16 + 4) hours as terminal if an SN is watched for a period of 24 hours as terminal dependent directly on the BN for the first 16 hours, switch for the following 4 hours, and terminal again for the final 4 hours.
Topology for Instant Dependencies
The PRIME instantaneous subnetwork "tree" is represented by this parameter.
Information from a Database
These elements will be used to communicate communications results achieved in the field, which will be coupled with existing database information to offer, for example, a representation of the increased connection in the electrical grid (which meter links to which SS, and in which line and phase).
Low Voltage (LV) Grid Connectivity
These metrics represent the database's current information, as well as PLC pertinent LV network information (like as availability), on a per-meter basis and taking element concentration into account.
Characteristics of the Meter
PLC signal injection capabilities are determined by meter properties.
Grid Conditioning for the Real-Time Acquisition of Operational Data in a PLC-Based Smart Metering System
The operation of the low voltage grid necessitates a thorough understanding of the electrical grid, at least comparable to that found in other portions, such as the medium or high voltage grid components. Utility-accessible LV databases offer varying levels of dependability. Some lack any data at all, while others have data that is insufficiently accurate. Today, despite the fact that many utilities assert to have knowledge of the precise relationship between a meter (point of supply) and an SS, this knowledge is not entirely reliable. In general, the LV line and the phase to which every meter is linked are significantly more challenging to register and update if more granularity is needed.
All performance data for the PLC smart metering system that could be obtained from a monitoring procedure was displayed in the preceding section. As described in earlier sections, this monitoring method is a one-time effort designed to comprehend how the network behaves. However, PLC-based smart metering systems may also include these metrics in their real-time information. As a result, the LV grid might also supply real-time information.
Additionally, the smart metering system might benefit from the function mentioned in section 2—signal repetition with the use of switches—if the PLC system being used is PRIME.
This feature was also researched in [35], where it was discovered that in order for SS multi-transformer environments to function properly, all of the LV busbars need the aid of so-called auxiliary nodes. Meters are better accessed by PLC components that are electrically close to them and have a good communications path towards the concentrator, necessitating this setup. As a result, it has been shown that the meters in their LV busbar are electrically dependent on that particular busbar, and the auxiliary nodes in these situations essentially act as switches that enable access to those meters.
Deeper granularity may be applied to this idea. It would be conceivable to install one auxiliary node per LV line rather than one auxiliary node per LV busbar, providing the benefit of recognizing which meter corresponds to which LV line. A PRIME system's technique for allocating a meter to an auxiliary node is based on internal PRIME stability data for this reason. As an extension of this strategy, the voting choice made at the PRIME internal level would result in the actual LV line connection of the meters in the SS if the PRIME system also relies on real received PLC power level information (auxiliary nodes require both voltage and current information).
Both radial and mesh low voltage distribution grids may use PRIME PLC technology. PRIME architecture will always be radial from a communications standpoint, and SNs will be linked to the BN with a stronger PLC signal. Thus, PRIME meters may be connected to LV lines in both grid types.
Detection of Lines
Additional computations must be added to the subnetwork's BN in order to construct a line detection method for an LV PRIME network. The chance that each SN (meter) in the subnetwork is a part of each of the LV lines is allocated. The likelihood is determined by an algorithm that is mostly dependent on the PLC signal power (calculated from the maintain alive messages given by SNs) that is received by each auxiliary node connected to each LV line. By comparing the quantity of messages received at each auxiliary node from all of the SNs, it is possible to determine if an LV line is connected to the SN. Table 4 provides a case study.
Table 4 shows the data the base node (BN) stores about line detection.
Nodes for Line Detection
The implementation of a PRIME PLC based LV line detection system necessitates the use of one auxiliary node per LV line, as indicated at the beginning of Section 6. A new kind of node must be included in the PRIME network for this purpose. Line detection node (LDN) will be the name given to this new node. LDNs, which are regular PRIME SNs, are in charge of determining the PRIME signal strength arriving from each of the meters at the head-end of each of the LV lines exiting the LV busbar. The measurements required are inductively coupled ones of the received current. Since the procedure is based on the application of Kirchhoff's current law (KCL), certain LDN current values are required.
Figure 5 depicts the installation of LDNs. In PRIME single phase injection systems, it should be noted that all LDNs must be connected to the same LV line phase of the system, which must correspond with the phase where the BN couples to. For an LDN connected at the head-end of Line 1, it can be inferred from Figure 5 and KCL that:
IPRIME1,1 = IPRIME1,CT + IPRIME1,2 + IPRIME1,3 + … + IPRIME1,N
Therefore, any PRIME SN attached to any LV line will always be detected more strongly by the LDN connected to that same line, according to KCL.
Figure 5 shows the position of line detecting nodes in an LV network.
Algorithm for Detecting Lines
They will only be listening to a certain kind of PRIME traffic, the ALV (stay alive) control packets, in order to consistently apply KCL and compare similar signals in all the LDNs. these assumes that every node in the PRIME network sends these kind of packets on a regular basis for link maintenance purposes and that the BN is aware of and controls this frequency.
Because information coming from SNs closer to the network (connected in level 0 or level 1) is more reliable than information coming from distant nodes, LDNs will maintain an internal log where they will record the power of the most recent ALV received from each node level combination (Switch Identifier and Local Node Identifier in PRIME terminology). A sample of the information each LDN stores is shown in Table 5.
Table 5 shows the power data that the line detection nodes have saved.
The BN (concentrator) will periodically ask each LDN for the most recent power data they have on a certain SN (meter) of the subnetwork. In order to verify that the data given by each LDN is current, the BN will make sure to request this information each time it receives a new ALV packet from the same node. The LDNs were supposed to be listening to this most recent ALV message, and if one of the LDNs did not deliver any new information, the most recent ALV packet was lost.
The BN will update the line detection information algorithm—which determines the possibility of a node being linked to each of the lines associated with each node in the network—using the information it has received from the LDNs. particular sufficient rounds of this method, the information will be consistent and the BN will be able to precisely determine which LV line every particular meter is linked to, even if there will always be some random noise and jitter in the measurements taken by the LDNs.
PRIME Topology's Distant Nodes and Line Deduction
A SN that is sufficiently far from the SS might not be immediately observed by any of the LDNs. This SN might be located extremely distant from the BN and be registered in the PRIME network through a series of switches. Thus, line identification by direct power information comparison would be absolutely impractical.
Fortunately, the PRIME network architecture, which is heavily reliant on the electrical topology of the subnetwork, offers enough information via communications dependence to identify the LV line to which a remote node is attached. Nodes registered through level 2 or higher level switches belong to the same line as the chain of switches, and as the PRIME network topology fully converges, the probability that a distant node registers through a switch connected to a different LV line tends to zero (existing PRIME field experience shows that it is not realistic that an SN is far enough not to be detected by any LDN while also being capable of registering through a switch connected to a different LV line).
PRIME's Phase Detection
There are various more attributes that may be identified in order to give further functionality now that we are able to know which is the SS and LV line that a meter belongs to. We concentrate on the LV line's phase with respect to the linked meter. In order to reduce the so-called technical losses, the transformer phase consumption balance depends on this information.
The PRIME PLC's synchronization capabilities are also available. There are many functional levels in PRIME. The physical and MAC layers are the most significant time domain layers. Since the PRIME MAC layer is inherently synchronous, it is simple to leverage this property to offer synchronization services to higher levels.
The PRIME standard also provides a PHY layer primitive (PHY_ZCT [28]) that enables a meter to communicate with higher layers at the precise moment when the voltage level signal in the phase to which it is attached crosses zero. Because of this, if each SN (i.e., meter) in the PRIME network is capable of detecting the zero crossing of its respective phase and the zero crossing event can be tagged relative to the timing of the MAC layer super frame structure, then the BN can assign the appropriate phase correspondence to each meter in its subnetwork by sending a query to each SN and asking for the zero crossing relative to the super frame time. As a result, the system is inherently resistant to latency,
Conclusions
The performance of real-time PLC-based smart metering systems is evaluated using a brand-new, comprehensive approach provided in this study. Using information gathered in the field and static grid data from utility databases, this approach may be used to track and evaluate the effectiveness of PLC smart metering systems that are being installed all over the world. The approach makes it possible to install PLC-based smart metering networks correctly.
In addition, this article proposes grid conditioning, which when combined with PLC-based smart metering data, will make it possible to get precise data on the LV grid, especially on the LV feeder and phase connections of each smart meter. This data offers a strong foundation for more precise fault location, LV grid layout, and point of supply identification. These functions are based on data on each meter's connection to each LV feeder and phase (identified by its address), allowing for an accurate and thorough mapping of the LV distribution grid architecture.
0 Comments