Abstract
Power Internet of things (IoT) is deemed as a promising network platform with widely deployed infrastructure to boost efficient information delivery in the power grid. Due to the long history of the mature power grid and increased requirements from various industries, the architecture of power IoT should be carefully investigated. Specifically, a large number of end devices in the power grid are required to simultaneously report their sensed information to the management side. However, there are few works related to the uniform communication mechanism to support various devices made by different manufacturers. In this paper, we study the architecture of power IoT for high concurrency access. For each layer of power IoT, we study its fundamental structure and functionality. Moreover, according to the architecture of power IoT, we propose an integrated gateway to support multiple end device simultaneous access. The simulation results verify that the proposed gateway can work well in different network traffic and communication protocols.
1. Introduction
The Internet of things (IoT) is an intelligent service platform that can exchange information in the physical and virtual world by connecting objects, people, and systems based on various protocols through sensors [1–4]. The characteristics of IoT can be categorized in five folds: (1) various communication models and generalized terminal interconnection; (2) automatic intelligent and efficient access; (3) large data throughput and low latency; (4) fast, flexible, and convenient data acquisition; (5) high-security requirement for internal and external network data. On the other hand, with the various construction requirement of the new type of a power system, a large number of distributed power and energy storage equipment have been introduced into the power grid. The advanced smart grid is believed to present similar characteristics by taking advantage of distributed computing systems, communication networks, and information networks. As a widely deployed infrastructure, power IoT can connect different entities in the power grid, such as power users and their equipment, power grid enterprises and their equipment, power generation enterprises, and corresponding equipment, electrical equipment enterprises and their equipment, by means of extensive information interaction and collaborative control. The energy production, energy consumption, and enterprise operation can be greatly improved by digital management in the power grid and the corresponding facilities.
As a significant issue of power IoT, simultaneous access of a large number of end devices has to be well solved to enhance the communication and data exchange in the large area of the power grid. Low-power communications, such as WiFi, ZigBee, and Bluetooth, can support multiple access in a certain area of the network [5–7]. The fourth and fifth-generation communications can be used to transmit data from the local area network to the public networks at a long-range distance. In the application of power IoT, there are quite a large number of end devices in the power grid with a long history [8]. These power devices employ different communication modules in terms of transmitting range, power, and network protocols due to different manufacturers. Thus, it is difficult to directly connect these devices to a public network without a uniform gateway or some similar mechanisms [9]. Moreover, how to process the different types of data from different devices in the uniform gateway should be carefully addressed to ensure data exchange and display. Currently, the state grid corporation of China boosts informatization in the power grid, which put IoT, mobile Internet, and the facilities of the power grid together to enhance the efficiency of data exchange [10–12]. Moreover, an integrated gateway with a high capacity of information perception, comprehensive information process, and control and management in the power grid is believed to realize real-time detection, high-accuracy localization, and construction supervision in power grid [13].
In this paper, we discuss the architecture of power IoT. Moreover, we analyze each layer of power IoT by taking the requirement of the power industry and current infrastructure into consideration. Then, we propose an integrated gateway for simultaneous access to various devices in power IoT. The performance evaluation shows that the proposed integrated gateway can work well to process multiple requirements from various end devices. The main contributions are summarized as follows:(1)Compared with traditional power networks, we study the main framework of power IoT, including different layers of power IoT associated with various functions.(2)Based on the framework of power IoT, we propose a novel integrated gateway for power IoT to support multiple power devices to simultaneously access.
The remaining of this paper is organized as follows. Section 2 presents the development of power IoT and the hierarchical design of power IoT. We also analyze each layer for power IoT in detail in Section 2. In Section 3, we present the design of an integrated gateway for power IoT and evaluate its performance in Section 4. The paper is concluded in Section 5.
2. Hierarchical Design of Power IoT
2.1. History of Power IoT
The development of power IoT evolves within three stages. In the first stage, a large number of sensing devices are deployed to enhance on-site digitization. The power grid is characterized by digital substation, dispatching automation, electricity information collection, distribution automation, intelligent station area, online monitoring of the status of power transmission, and transformation equipment. Within several years of digitization, the power grid side has basically met the requirements of the interconnection of things. The whole power grid is at the level of “observable”. That is, the abstract process from the “physical grid” to the “logical grid” has been realized on the computer. Furthermore, the “digital twin” [14, 15] foundation of industry 4.0 has been realized based on the infrastructure of power IoT. From a comprehensive energy aspect, the power IoT is still in a preliminary stage because there is a huge digital bottleneck in the automation of industrial parks and the digitization of electricity using management. The second stage is referred as to interaction among things. More specifically, on the premise that the full digitization of power IoT and digital twin realization, how to use this digital information to improve the efficiency of management is an important issue. In a nutshell, things in power IoT can efficiently communicate with each other. On the power grid side, management information has been basically realized through several rounds of information technology investment, such as the informatization progress of production, marketing, scheduling, finance, and safety supervision. Specifically, the problem that needs to be solved is data interaction. The most typical data is marketing information and production information. Due to the division of departments, complete information and data systems are not easy to establish. As a result, it is difficult to put these two large systems together with efficient and unlimited interaction. Therefore, the power of IoT takes “data unification, operation, and distribution” as an important role. On the service side of comprehensive energy, the basic informatization of management is quite low. There are many small management systems existing in power grid companies. Due to the lack of full business standards and information models, these management systems form information isolated islands [16]. Artificial intelligence (AI) is involved in the third stage. In power system, for example, intelligent data analysis and intelligent unmanned aerial vehicle (UAV) applications can be utilized for the reduction of labor cost and decision-making [17, 18]. These artificial intelligence applications can enhance the perception and interaction in the power grid and self-healing capability. Furthermore, various data from different resources can be analyzed by artificially intelligent algorithms to improve the collaboration among the source power, network, charge, and storage in the power grid system. On the service side of comprehensive energy, within the establishment of the spot market, power grid companies can generate more services. For instance, allowing virtual power plants to participate in the spot and ancillary services trade, using price signals and service demands in each subject, link, and platform and finally forming the intelligent energy network ecosystem.
2.2. Hierarchical Design of Power IoT
Based on the application requirements of collaborative interaction between the source network and storage, we design a four-layer architecture, namely, the perception layer, network layer, platform layer, and application layer, which also follows the main business requirements of power IoT and various commercial network framework. The detailed structure is shown in Figure 1.
2.2.1. Perception Layer
The perception layer includes end-side devices and side-side devices. The end-side devices are responsible for providing basic data of the distribution network such as the operation status, equipment status, environment status, and other auxiliary information to the platform. End-side equipment includes source nodes, network, load, and storage structured data collection of each link of the intelligent device. Based on the principle of on-demand deployment, side-side devices realize local data aggregation of all kinds of perception information in the region, which form open platforms to support multi-container, multi-channel, and multi-protocol modes. End-to-end collaboration is completed through data exchange to realize full data collection, full perception, and full control. Standardized processing and uploading of collected data are realized based on the object model. The real-time full-duplex interaction of key operation data was completed to achieve edge-cloud collaboration by taking the advantages of cloud computing and edge computing.
2.2.2. Network Layer
The network layer serves as an up-down data transmission channel, see Figure 2. It can be divided into remote communication networks and local communication networks. The remote communication network provides the data communication channel between the source, network, load, and storage interactive application and the edge computing node. The local communication network provides the data communication channel between the edge computing nodes and the terminal units. The communication technology 5 G/6 G and the visual private network (VPN) constructed by the power grid system are utilized to transmit perception data to the platform.
2.2.3. Platform Layer
The platform layer adopts big data, micro-service, container management, artificial intelligence, and other technologies to realize the comprehensive cloud and micro-service of the master station. This can meet the business requirements of plug and play of massive terminal devices, rapid online application, and effective data fusion of multiple platforms. The platform also can realize unified online management and remote operation and maintenance of all kinds of sensing layer devices and IoT applications.
2.2.4. Application Layer
The application layer contains advanced applications that are suitable for collaborative control of the source network load and storage. Advanced applications should be developed based on collaborative interaction strategies. The detailed application requirement is described as follows: (1) The Collaborative and interactive application. The core business of the application layer is collaboration and interaction for power source, network, charge, and storage. Real-time coordination and optimization of the collaborative and interactive application of source, network, charge, and storage are based on massive information, such as state of charge (SOC), power generation output of new energy, and power grid operation. Based on the recognition of power grid operation mode in a stable power grid, the real-time closed-loop control of the power of common connection points, controllable power supply, charge and discharge of energy storage, and the flexible load is realized to ensure the coordinated optimal operation of the power grid. (2) Energy storage optimization. According to the spatial-temporal coupling correlation and complimentary matching characteristics of layered energy storage, the dynamic autonomous control strategy and distributed collaborative control strategy of energy storage are formulated to optimize the charging and discharging mode of energy storage and improve the power supply. (3) Optimization of electric vehicles. According to the current operating status of the system, the coordination control system in the charging station is adopted to control the charging of power electric vehicles. The system is also required to optimize the charging behavior of electric vehicles, which can enable electric vehicles actively participate in the power balance regulation of the system. (4) Management of demand side. Load demand response management mainly manages and evaluates users who participate in the demand response. The system stores user names, user addresses, contacts, contact information, and maximum power response capacity. After the demand response control is delivered, the system releases the execution process of users participating in the demand response and then optimizes the implementation process of the demand response in real-time based on the actual situation.
3. Access Solution of Multiple Power Devices in Power IoT
In this section, we propose an integrated gateway to receive and process signals from various devices in power IoT. At first, we identify the requirement of multiple device simultaneous access in power IoT. Then, we present the framework of the integrated gateway, including central processing and peripheral modules. Furthermore, we specify the data format to support simultaneous accessing of multiple devices. In addition, we discuss the process of data parsing and protocol conversion.
3.1. Requirement of Multiple Devices Access in Power IoT
As discussed in Section 1, the state grid boosts the infrastructure establishment of power IoT in the past several years and there are a large number of end devices required to access the public networks. According to the suggestion and requirement from the state grid, we summarize different access requirements from source-grid-load-storage of the power grid in terms of device type, communication protocol, data content, and area as shown in Table 1.
3.2. Integrated Gateway Design for Multiple Protocols
As discussed in Section 3.1, there are various devices from source-grid-load-storage in power IoT. Therefore, the power IoT can be considered as a heterogeneous network to support different devices simultaneous accessing and data exchange. The proposed integrated gateway includes central processing and peripheral modules. As shown in Figure 3, the central processing module consists of data identification, buffer, data processing, and signal transmitting modules. The peripheral module contains ZigBee, WIFI, Bluetooth, LoRa, NB-IoT, Ethernet, and USB modules. Furthermore, power, reset, and display modules are designed. The peripheral module is connected to the central processing module by a two-way misconnection that is used for the whole system reset and running states display. The communication modules, such as ZigBee, WIFI, Bluetooth, LoRa, NB-IoT, Ethernet, and USB modules, are connected with central processing modules by a one-way link for data transmission. The central processing module also conducts the protocol conversion and cache data.
3.3. Data Preprocess and the Buffer Module Design
The data identification module is used to distinguish received signals from different end devices. In the proposed framework, ZigBee, WIFI, Bluetooth, LoRa, and NB-IoT are supported to report data to the central process modules through multiple communication ports. The received signals (i.e., data frame) can be in the buffer to mitigate data congestion. As shown in Figure 4, if the data processing module is idle, the received data are sent to the data processing module. Otherwise, the received data are stored in the buffer.
We also design a buffer module to avoid data congestion in high data traffic. More specifically, the proposed buffer module provides a variable table, device type, protocol conversion, port number, IP address, and timer to store the received data, which can be used when the received data is requested by the corresponding modules. The processing flow is presented in Figure 5.
3.4. Data Processing and the Transmitting Module
The proposed data processing module is used for parsing and recombining data packets according to different protocols. For example, we can encapsulate received LoRa data after analyzing and storing data frames. Furthermore, the data processing module can reply to and coordinate different requests from other modules. We present the main flow of the data process in Figure 6. In Figure 6, the number of threads presents the performance of concurrent processing. When the volume of received data is larger than a certain threshold, the rest of the data is stored in the buffer. The new thread is called for data processing if the volume of data in the buffer is larger than the threshold. We also configure the maximum number of thresholds. If the data traffic is reduced, the threshold of data processing is also reduced when the volume of data in the buffer is small. In addition, the data transmitting module is utilized for communication between the proposed integrated gateway of power IoT and the base station. The data transmitting module can send data according to different communication types.
3.5. Peripheral Module Design
The peripheral module contains ZigBee, WiFi, LoRa, NB-IoT, and Ethernet modules, where ZigBee, WiFi, and Bluetooth can be used for short-range communications. LoRa and NB-IoT are employed for long-range communications. We observe that NB-IoT can realize large network coverage with a large number of connections and low costs. LoRa is an alternative for long-range communications with unlicensed frequency bands. We present the peripheral module in Figure 7.
We also design a power module to support the central processing, display, and peripheral modules. We adopt 3.3 V/4.2 V as working voltages for different scenarios. The reset module is used to reset the integrated gateway. The main resetting flow is shown in Figure 8.
3.6. Data Parse and Protocol Conversion
As we discussed in Sections 3.1 and 3.2, there are various end devices tended to access power IoT to report various data. However, it is worth noting that different end devices and the corresponding data format are specified by manufacturers or institutionalized.
Thus, we design the process of data conversion in Figure 9. We first identify the validation of the received data. If the received data is valid, we conduct the data extraction and parsing. Otherwise, we discard the invalid data packets. Then, we convert data according to the proposed uniform data format and deliver data according to different communication protocols.
In the proposed framework of the integrated gateway, we focus on data parsing and protocol conversion to process different types of data in a uniform data format. Table 2 presents the proposed data frame to uniformly process data packets.
The start field is the version number of the self-defined protocol format, which provides global authentication during version update and iteration of node data format. The authentication results are submitted to the management layer for processing. The communication parties must use the same protocol version for data exchange. The gateway system discard message with invalid format. In general, the amount of data transmitted in the Internet of Things application network is generally small. In order to make the gateway high compatibility, it is still necessary to consider the problems that occur when transmitting long data frames. To this end, a message sequence number is added into the protocol format to represent the sequence of network data transmission. The sending terminal sets the serial numbers according to its own data volume, and the receiving terminal integrates and splices data according to the defined format when receiving different serial numbers. During the network terminal equipment operation, distinguishing different each kind of data sources is required to distinguish. The unique information of each kind of data source is reflected in the control field. The control field is used to specify the message frame of specific types of control. In the concrete implementation, the basic types include the data frame and the command frame, such as the start frame and the end frame type.
4. Performance Evaluation
We verify the dynamic loading access protocol modules for multiple end devices and corresponding middleware access based on the simulation framework with heterogeneous gateway protocols. The end devices adopt a shared library and a thread pool of Linux to enhance the reliability of simultaneous access. Specifically, we exploit the generalized object factory with the Linux shared library for a standardized data process, which can achieve the parse of heterogeneous application protocols with a low development cost. The simulation results show that the middleware gateway of power IoT can support access to multiple end devices simultaneously and avoid the heterogeneity of communication and application protocols. Furthermore, the gateway of power IoT can work well when a large number of new end devices are accessed, which can be utilized in different applications. We conduct the simulation and evaluate performance in terms of dropped packet ratio and system communication delay.
4.1. Simulation Configuration
We use Visual Studio as the main IDE (integrated development environment). “makefile” is used to define, compile, and link rules of the entire system and “CMake” is used to parse the commands in the makefile and to achieve automatic compile. We also run the system in Linux with C language due to the high requirement of computational and spatial complexity of protocol conversion. The simulation framework is shown in Figure 10. In Figure 10, according to the design objective, there are three modules in the framework, including end devices management, data processing, data output, and saving module.
The end device management module is for heterogeneous communication protocols, which is consisted of concurrent data collection with multiple protocols, the shared library of the heterogeneous communication protocols, and original data processing and analysis. To be more specific, data generated by heterogeneous sensor networks and sensors are collected by the end device management module, which is transmitted to the data processing module for parse. This end device management module is able to collect concurrent data from the perceived layer in different power IoT. Meanwhile, the integration gateway enhances access to new networks, such as sensor networks and IoT.
The standardized data processing module consists of rules of data processing, data conversion, and standardized data processing, which is used for heterogeneous application protocols. The collected data from the end device management module are extracted, compressed, and mapped for the required standardized data format. The standardized data processing module can immediately identify the data source with which sensors when receiving different data from different sensors. The data is also mapped to meet the application requirements. The data output and storage module has two parts, such as data output and database, which store the standardized data in the database and output the standardized data to applications.
4.2. Performance Evaluation
We verify the functionality of the integrated gateway for power IoT by simulating delivery data to the middleware of the gateway with multiple protocols. We generate data by running simulations on windows. The middleware gateway is working on the virtual machine that runs Ubuntu. The virtual serial port software is connected to COM1 and COM3. COM1 is on a PC and COM3 is on a virtual machine. Moreover, multiple protocol data generators can achieve data delivery between PC and gateway using TCP and UDP. The data generator can simulate six different sensor data. These different data can be parsed by functions implemented in the middleware gateway.
As shown in Figure 11, the proposed integrated gateway can achieve multiple sensors simultaneous access. The communication delay of the middleware gateway is defined as the time duration from receiving one frame at serial ports to storing this frame in the database. We deploy 6, 12, 24, and 30 sensors in the end device management module, respectively. Each sensor transmits 100 copies. There are six communication approaches in these sensors and the transmitting period is 5s. We conduct 7 experiments in 5 groups. The simulation results are reported in Table 3.
It is obvious that the average delay is caused by protocol conversion, data processing, and queuing delay. With the increase in the number of sensor nodes, the average delay increases. The average delay is less than 100 ms defined as IPTD in CCSA. We also observe that the protocol conversion and data processing results in the average delay, whereas the queuing delay is not the main factor of average delay. In practice, we should consider the propagation delay of the signal (which can be ignored when the transmission delay is short).
Figure 12 shows that the average dropped packet ratio increases when the number of communication protocols is larger. Moreover, the average dropped packet ratio also increases when the number of sensors increases. Time-out caused by data processing and queuing are the main factors in discarding packets.
5. Conclusion
In this paper, we study the typical architecture of power IoT and the corresponding layer in terms of functionality and characteristics. Based on the comprehensive understanding of power IoT, we propose an integrated gateway for power IoT. The simulation result shows that the proposed gateway can support multiple protocols and multiple end devices for simultaneous access. In the future, we will implement the prototype of different layers of power IoT and conduct field experiments.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Authors’ Contributions
Conceptualization was done by Fei Yu; methodology was done by Fei Yu, Wei Rao, Chang Liu, Jin Wang, and Liang Zhou; software was done by Li Tian and Jie Wang; formal analysis was conducted by Jiangpei Xu and Chang Liu; resources were carried out by Chang Liu; data curation was done by Chang Liu; writing-original draft preparation was written by Fei Yu and Wei Rao; visualization was done by Jin Wang and Liang Zhou; supervision was supervised by Rao Wei. All authors have read and agreed to the published version of the manuscript.