Enabling Industrial Internet of Things (IIoT) towards an emerging smart energy system

1 Introduction

When we are seeing the trend of more sustainable energy to be used in daily life, the time of de-centralized future energy system finally arrives [1,2]. Similarly, as Information technology (IT) [3], which has been focusing on store, retrieve, transmit and manipulate data via computer based applications, Energy technology (ET)[4,5] is going to store, receive, transmit and manipulate energy via Distributed Energy Resource (DER) based applications [6,7]. Similarly, as different data format found in IT industry, such as text, sound, picture, video, movie and others, there are also many energy formats found in ET industry, such as power, heat, cold, fuel, gas, water and others. IT industry has solved generic data conversion methodology between different formats owing to the prevalence of data operating system (DOS) that could manage the computer resources in a highly-efficient and well-organized fashion.

Unfortunately, as network based ET industry is still as its infancy stage, especially from DER application perspective, there is no genetic energy conversion methodology found between different energy formats yet. In addition, even managing the energy system with different formats on one platform is still a challenging task as so many tedious details at sub-system level are based on different physical / chemical characteristics and not suitable or difficult to be generalized. This has become a major barrier to overcome in the development of DER based energy system.

This paper proposed a novel industrial methodology for future energy system based an Energy Operating System named NicerNet, and then the Industrial Smart Energy Consortium (ISEC) Initiative is proposed as one effective mechanism to build Industrial IoT regulation and standard for large industry user.

2 Industrial IoT methodology

In a distributed energy resources (DER) based energy system, the tasks for control and management are more challenging than a single energy conversion device,therefore the Internet of things (IoT) [8,9] based techniques are being studied and distributed energy control has become one of the key interests for future de-centralized energy infrastructure, sometimes it is being referred as “energy internet” or “Internet of Energy(IoE)” [10], either way it shows the industrial trend of building the future energy world into an IoT concept based energy-centric ecological system.

2.1 Energy operating system based NicerNet

A unified distributed energy network paradigm, from multiple type sources including both fossil energy and renewable energy, from multiple dimensions sources including both power and chemical energy, is being proposed by NICE of Shenhua group, which is defined as Normalized Information Cell-driven Energy Regional Network (NicerNet). Research on energy coupling and optimization control based on Industrial Internet of things(IIoT) is in progress. Fig. 1 is the Energy Operating System (EOS) platform, which is being developed by the distributed energy system group of NICE as the key fundamental architecture for future distributed energy control & management solution to meet the challenging requirements of a future energy generating and distributing network.

Fig. 1 Structure of NICE_EOS platform

2.2 Classic cloud computing

Most recently, the Cloud computing has been well accepted as the foundation solution for IoT application. Fig.2 is the structure of Cloud control platform. It may be well suited for consumer/social/enterprise IoT application as large non-real time big data streams from device to Cloud is the major task for data analysis and management.

Fig. 2 Structure of cloud control platform

However, this is not the case for industrial IoT such as smart energy application. Sending all the energy data to the Cloud for analysis also poses a risk of data bottlenecks, as well as security concerns. New business models need data analytics in soft real time at sub-second or hard real time at millisecond level with interoperability functions between devices that are robust, reliable, flexible and scalable. The problem of data congestion will only get worse as Industrial IoT applications and devices continue to proliferate and generate a couple of orders in magnitude of data volume when compared to traditional IoT applications. These requirements have become the driving force for creation of Fog-Computing concepts.

2.3 Emerging fog computing

To support the computational demand of real time latency, sensitive applications of largely geo-distributed IoT devices/sensors, a new computing concept named “Fog computing” has been introduced.

Generally, Fog computing resides closer to the IoT devices/sensors and extends to the Cloud based computing,storage and networking facilities. In contrast to the Cloud,Fog platforms have been described as dense computational architectures at the network’s edge. Fig. 3 shown the characteristics of such platforms, which reportedly include low latency, location awareness and use of wireless access.Benefits include real-time analytics and improved security.Fog computing is critical since it enables time-sensitive,reliable operation, and removes the requirement for persistent Cloud connectivity to address many of today’s emerging scenarios.

Fog computing - a term originally coined by Cisco - in many ways has been successfully extending the Cloud computing to the edge. However, Fog computing is not equivalent to edge computing, since at the edge of the network, computation intelligence must be residing at closer to the boundaries of sensors and actuators due to varied reasons such as proprietary algorithms,communication power usage, data congestion reduction or simply cost-saving, bandwidth limitation etc. This has been driving the creation of Mist-Computing concept.

2.4 Proposed mist computing

Mist computing, although has been discussed from academic perspective occasionally - is a term systematically defined by NICE of Shenhua group to include all edge to extreme edge data analytics of IoT network, that is not or cannot be covered by classic Cloud and emerging Fog computing, which is shown in Fig. 4.

Mist is a more dispersed version of Fog, just as Fog computing is a dispersed version of the Cloud. The proximity of the Mist is closer to data sources, where real time analytic and data handling are required when there are no benefits to be gained by moving it to the Cloud remotely.

Fig. 4 Structure of IoT network

The Internet of Things is about more than just connecting devices to a common network. The value comes from the collecting of data, analysis of data through analytics, and the ability to make critical decisions quickly.

The strategy of connected analytics in IoT is targeted at making this process operate at demanded speeds from application with optimized values propositions all way from Cloud, Fog to Mist and even further down as far as the self-similarity continues.

The iteration feature of this connected analytical structure has been summarized as Fractal-Computing by NICE of Shenhua, which drives a cost-effective dataanalytical infrastructure design for future IoT based smart system design.

2.5 Fractal-Computing based Smart-OS

From a connected analytical perspective, Cloud, Fog and Mist computing all show self-similarity [11] from the following perspective at different spatial, time and type categories：

· Latency and Response time is often a critical part, and normally application oriented and going down from Cloud,Fog to Mist direction.

· Data Bandwidth and Capacity is very often costdriven and normally data bandwidth and capacity are going down from Cloud, Fog to Mist direction.

· Security and privacy are a big concern and a more local and decentralized approach increases the level of security from Cloud, Fog to Mist direction.

· Power consumption (environmental issue) is more of an issue in the Cloud-computing world. And the Cloud-Fog-Mist direction synchronizes with the de-centralization of energy structure with less power consumption and lower low-carbon economy.

· Data obesity - the percentage of amount of untreated data are going down with Cloud, Fog to Mist direction as the analytical intelligence is pushed to the edge.

Therefore, a FRACTAL-Computing [12] could be defined as a complex, non-linear, interactive, network based data-analytical mechanism that has the ability to adapt to a dynamic changing IoT environment with autonomous and self-intelligent behaviors growing into applications based on many layers of Cloud-Fog-Mist liked fractal data analytical layers as needed.

NicerNet, as proposed by NICE, is adopting Fractalcomputing based Smart Operating System Concept into de-centralized energy network by introduction of Nanostructure as fundamental fractal element of the system. Fig.5 is the structure of NicerNet. If we use energy storage as one example, we can find that modern energy system is a fractal energy network full of fractal storage assets at different scales with self-similarity from nano-structure as defined in figure below.

Fig. 5 Fractal computing structure of NicerNet

Fractal system evolves by random mutation and selforganization based on self-similarity, it transforms and iterates into a complex and self-sustained network based their internal fundamental fractal element with simple rule-set from natural selection. Examples include living organisms, the nervous system, the immune system, the economy, corporations, societies, and so on.

By introducing fractal-computing concept, we can not only build smart Operating System (OS) around future energy network as Energy Operating System (EOS), but also benefit us to find a more cost-effective way to realize the smart goal in so called “industrial internet” and “industry 4.0”.

Working with OpenFog consortium led by CISCO,NICE is actively planning to create an OpenMist working group to bring Shenhua to be the leader at the forefront stage of the coming IoT millennium and also pioneer the smart energy application based on the effort from fractal computing related research and development work.

NICE has built an open architecture driven and multienergy based Nano-grid platform and advocate to form an Industrial Smart Energy Consortium (ISEC) to drive for the success of smart energy application in large industrial application.

3 Industrial smart energy solution

Fig. 6 shows the China Power Consumption in 2016 [13].In China, 70% of power consumption is from industrial users. There are urgent needs for large industry users to update their energy infrastructure as well as production and manufacture process based on the latest industrial Internet of things (IoT) concept and methodology. With its diversified portfolio, Shenhua is at the position to create a novel multienergy system based on combined power & chemical distributed grid. Fig. 7 is the structure of power & chemical distributed grid.

Very recently, both GE and Siemens have released their own Industrial IoT Operating System platform which is called Predix and Mindsphere, separately, as well as many other companies releasing similar IoT platform products.The disruptive change that industry analysts are forecasting behind these products will have profound consequences for energy & manufacturing, the global economy and the living culture.

Fig. 6 China Power Consumption in 2016

Fig. 7 Structure of power & chemical distributed grid

However, the vision is bigger than just one company and the success in IoT is way beyond one particular platform or specific system from one company.

3.1 ISEC initiative

NICE, as the innovation hub for Shenhua group, is making efforts to setup an open architecture Industrial IoT platform, defined as Industrial Smart Energy Consortium(ISEC) and build partnerships with other technology companies, academia, consultants, and systems integrators.The effort is to share its expertise and knowhow from system specification and industry regulation perspective,and co-innovate to drive important advances in Industrial IoT test-bed functionality and harness the potential of the Industrial Internet to deliver powerful customer outcomes in Chinese industry, with clear short-term goal to be defined as serving un-manned underground mining operation in smart coal mine application.

ISEC will be a purely technical collaboration joining group with the clear vision to adopt Industrial IoT methodology to drive and pioneer IIoT regulation and standard to build smart energy and further smart operation in Chines industry, starting from mining, power, railway &harbor and chemical, which Shenhua business is focusing on.

ISEC will be test-bed centric platform by using NICE as the open architecture Industrial IoT laboratory to jointly develop solution based projects under different working group, initially categorized by the background of H4 founding members and option to expand in the future.

· IIoT chip & module (Huawei)

· IIoT equipment (Huaxia Tianxin)

· IIoT System Spec. & Design (Shenhua - NICE &Shendong)

· IIoT Theory and Algorithms (Tsinghua)

Most recently, NICE has implemented many test-beds under the open architecture platform defined as Nano-grid and could be used as the examples and test cases to further establish related regulations and standards for the industry.

3.2 Open platform of Nano-grid

As the first combined power & chemical distributed grid, NICE has enhanced the energy network R&D focus towards understanding in depth, the inter-operability between devices and the performance dependence of their functions from multi-perspectives of material stability,reaction efficiency, heat & cold control, power and energy optimization based on multi-disciplinary data analytics that is driven by Industrial IoT infrastructure. Fig. 8 is the features, configuration, algorithm and advantage of Energy,information and market three into one analysis and design platform [14,15].

The goal is to have one unified design platform to be able to optimize the energy technology (ET), possibly extendable to operation technology (OT), under the adoption of information technology (IT) evolved IoT technology in the most cost-effective way. The most challenging aspect is how to build a smart communication and control architecture to address the following issues：

· Interoperability: Data interoperability for a vibrant competitive market

· Security: Cybersecurity is a founding architectural principle

· Practicality: Incrementally buildable, upgradable,expandable, compatible with legacy installations and protocols

· Edge Intelligence: Intelligent system control at the edge, driving automation and rapid response

· Data Integration: Enabling data sharing between grid and Cloud/Fog/Mist analytics

Fig. 9 is the solution, which is the adoption of Data Distribution Service (DDS) bus. DDS is a data-centric middleware standard with roots in high-performance defense, industrial, and embedded applications. DDS can efficiently deliver millions of messages per second to many simultaneous receivers. Because it targets device-todevice communications, DDS differs markedly from the other protocols in QoS control. In addition to reliability,DDS offers QoS control of “liveliness” (when you discover problems), resource usage, discovery and even timing.

Fig. 8 The features, configuration, algorithm and advantage of Energy, information and market three into one analysis and design platform

Fig. 9 Adoption of Data Distribution Service (DDS) bus

3.3 One network solution

Thanks to the concept of NicerNet, it is the generic concept to deal with energy assets at different time scale and different dimension. By introducing hydrogen as one kind of storage [16], unified solution could be found to deal with both fossil and renewable energy.

Fig. 10 is the schematic of Hydrogen coupled energy system. It is the generic concept to deal with energy storage at different time scale and different dimension.By introducing hydrogen as one kind of storage, unified solution could be found to deal with both fossil and renewable energy. This feature is especially important for the energy infrastructure in China as it generalizes energy conversion concept from power engineering to chemical engineering so that the expertise of multiple disciplines can be brought together into one platform solution for the best optimization result. It is also the first time in one system that the carbon foot front could be quantified and controlled with real time IoT intelligence.

Fig. 10 The schematic of Hydrogen coupled energy system

3.4 Control interoperability

Fig. 11 is the framework of NICE_EOS, it is designed with open architecture framework for future energy internet system.

Cloud Layer: Energy Cloud layer is defined for globalized energy data analytical function so that energy transaction based application could be developed as software module at this layer.

Fog Layer: Energy Fog layer is defined for localized energy flow analysis so that energy dispatch based application could be developed as software module at this layer.

Mist Layer: Energy Mist layer is defined for extreme edge energy control analysis so that energy stability based application could be developed as software module at this layer.

Fig. 11 Framework of NICE_EOS designed with open architecture framework for future energy internet system

Fig. 12 Application of NICE_EOS targeted for the future low carbon energy internet system

Fig.12 is the application of NICE_EOS, which is targeted for the future low carbon energy internet system driven by Shenhua Group.

Alternatively, the OS concept could be further expanded into multiple Industrial IoT applications. It can directly benefit Shenhua core business such as un-manned underground mining, coal mine safety smart management,smart power plant, smart railway, smart harbor, smart chemical factory, etc. This is also the driving factor to have the ISEC initiative so that system specification and related equipment could be certified at an open architecture based laboratory platform to drive correct industrial regulation.

4 Conclusion

The growing popularity of distributed energy is analogous to the historical evolution of computer systems.Just as the smaller size and lower cost of computers has enabled individuals to buy and run their own computing power, so the same trend in energy technologies is enabling individual business and residential consumers to purchase and run their own energy systems.

Implementing distributed energy can be a more complex system. There are so far, no analytical tools available from mathematical equation perspective to be able to describe the behaviors of distributed energy system with large quantities of assets.

More recently, the industrial internet or industrial internet of things (IIoT) has been introduced to energy industry and all the assets in a distributed energy system are treated as a smart industrial device so that a similar industrial distributed control could be applied. It can also use the latest IT tools such as big data or Cloud computing methodology to handle the large control signal flow of distributed energy assets, but the following issues remain：

· System tends to be over-designed with more management function at larger time-scale rather than control function at smaller time-scale as needed.

· Industrial IIT methodology has the benefits to bring the information together in a real-time fashion so that a decision making could be done in a much faster speed than normal network. It is normally considered as a digital network based control tool sets added to a distributed energy entity, which has no intelligence to identify and solve the critical issues in a decentralized energy environment such as stability etc. As we simply lack a simple and efficient theoretical based rule sets for the system to learn, the complexity of adaptive control for the dynamic structure evolvement of distributed energy system is not able to be learned by the system itself from IIT perspective.

· If the foundation of distributed energy control is not unified, it is very hard to design a self-learning system for distributed energy, therefore, the energy internet won’t be able to work in a self-disciplined fashion, which means extra effort on system integration and breakdown maintenance will dramatically increase the overall system cost and make the distributed energy less compelling and competitive.

NICE_EOS, under the concept of NicerNet, is being developed at an open architecture framework Nano-grid for energy control and management. Supported by the real application from Shenhua group, the largest energy company in China. This methodology has been proven to be a feasible system solution to solve the issues above,thanks for a Fractal based Mist-COMPUTING layer and related DDS communication layer for effective connected analytical in real time between devices.

The Industrial Smart Energy Consortium (ISEC)Initiative is proposed as one effective mechanism to build Industrial IoT regulation and standard for large industry user as Shenhua, so that application driven test-beds could be built and tested in NICE to bring Shenhua into the leading position for IIoT technology development and application.

Acknowledgements

This work was supported by National Key Research and Development Program (2016YFE0102600); National Natural Science Foundation of China (51577096,51477082).