IoT Caters to Manufacturing Industry: Predictive Maintenance

Predictive Maintenance: Key Point for Service Transformation

Although manufacturing enterprises are highly valuable in the huge service market, traditional manufacturing faces dim prospects. Take product maintenance as an example: traditional maintenance modes mainly comprise posterior maintenance and preventive maintenance. We can think of the former as a group of firefighters who rush to the scene when failure has already occurred and damage is done; the latter mainly relies on scheduled maintenance and artificial field maintenance, while deficient maintenance still exists. Excessive maintenance wastes of energy and results in soaring of human resources costs. In other cases, certain fault symptoms are easily overlooked and some maintenance workers lack sufficient knowledge.

In fields such as engineering machinery, industrial washers, industrial air conditioners, and numerical control machines, costly heavy-duty machines must be distributed over a wide sales area. Traditional maintenance carries risks of unpredictable emergencies and requires prolonged preparation and maintenance cycles, all of which impede customers’ sound business operations. Just imagine: in a metropolis, a sudden breakdown of elevators may inconvenience dozens of happy families. In factories, a one-hour interruption of intelligent machinery may delay the delivery of billions of projects all the way down the supply chain. In the countryside, faults in heavy-duty harvesting machinery may reduce harvest efficiency, leaving crops decaying in fields.

Predictive maintenance brings about a revolution in the servicing mode and process. This solution connects products and monitoring sensors through IoT infrastructure to monitor the running and usage of products in real time. It uses a cloud-based Big Data analytics platform for predictive analysis, predicting device faults, locating potential faults, and offering remote services. It makes maintenance easier while ensuring device reliability and reducing costs.

Predictive Maintenance: Start from Edge Intelligence

However, enterprises also face some challenges when they start predictive maintenance. Take industrial machinery as an example; the devices demanding maintenance may be scattered all over the world, each producing over 10 G of data. Collecting and uploading all this data to the cloud for analysis puts a heavy burden on the network and is costly. In addition, the real-time requirements of key services cannot be met because data analysis and control logic are all achieved in the cloud. The quantity of connected devices has been exponentially increasing in recent years, so the arduous task of dealing with connections and management of massive terminals is becoming ever more crucial. Moreover, real-time analysis and settlement of large-scale data are also obstacles lying ahead, which must be overcome to smoothly implement predictive maintenance.

Edge Computing IoT (EC-IoT) effectively supports predictive maintenance. EC-IoT comprises the terminal communication module, edge computing gateway (Huawei AR500 series), and Agile Controller. The terminal communication module supports intelligent interconnection of IoT terminals and the sensor network; the edge computing gateway provides intelligent services in immediate areas; and the Agile Controller allows interconnection with third-party application systems through open API/eSDK. It also enables intelligent connection and efficient management for massive unattended terminals through cloud management. EC-IoT provides excellent adaptability and streamlining services and innovative business modules for customers, serving as a solid foundation for businesses’ foray into the exciting possibilities offered by predictive maintenance.

What role does EC-IoT play in predictive maintenance?

The EC-IoT Solution is an innovative application of the edge computing architecture to the IoT field. At the network edge close to a device or data source, the solution deploys the terminal communication module and edge computing gateway that integrates network, computing, storage, and application capabilities, providing platform support for device, network, data, and application domains for edge computing. The device domain supports real-time intelligent connections of onsite devices through the terminal communication module. The network domain provides real-time connection and management services for system connections, data aggregation, and transmission. The data domain provides edge data aggregation and optimisation services, ensuring data security and privacy. The application domain uses open interfaces to achieve edge industry application deployment and support edge service operation.

This architecture helps customers from many industries deploy custom applications at the network edge and apply predictive maintenance data analysis models to meet their service requirements, improving efficiency and reducing costs. The applications and models implement real-time data cleansing, data analysis, and pre-defined service response policies based on the data analysis results, helping customers quickly detect potential device faults. In addition, the IoT gateway supports local service survival. If a connection with the cloud fails, data can be saved and processed locally. After the connection is restored, high-value local data is automatically synchronised to the cloud, ensuring complete product operation information is maintained in the cloud.

Edge computing and cloud computing have different use purposes, but can coordinate to benefit users in brand new ways. Cloud computing is applicable to non-real-time and long-term Big Data analysis and is useful in services such as routine maintenance, fault and risk identification and analysis, and product health check. Edge computing is applicable to real-time and short-term data analysis, supporting real-time alarms, fast fault identification, and service response in milliseconds. The EC-IoT Solution implements collaboration between cloud computing and edge computing. The edge computing gateway is deployed close to a device and collects high-value data required by the cloud, supporting Big Data analysis on a cloud application. Based on analysis results, the cloud computing application optimises service policies and delivers them to the edge computing gateway for service processing optimisation.

Take the elevator industry as an example. There will be about 20 million elevators worldwide by 2020. The EC-IoT based elevator IoT solution deploys edge computing gateways connecting to elevator controllers and sensors to collect elevator running data in real time. Based on local light-weight data analysis models, the gateways pre-analyze data such as elevator noise frequency and strength, helping predict and quickly detect potential faults on elevators. They also send the data to a cloud-based Big Data analysis platform, through which a customer can obtain the status of each component of elevators. The gateways then receive the analysis results from the platform and optimise the data analysis models to implement more intelligent prediction.

The EC-IoT Solution uses the cloud management architecture. The Agile Controller deployed on the cloud platform implements unified management and automatic deployment of computing resources, protocols, applications, and data of the edge computing gateway. It provides cloud-based network management. It supports seamless extensibility and can uniformly manage tens of millions of IoT terminals in the cloud. With cloud management, the Agile Controller provides full lifecycle management covering network planning, deployment, and O&M. Through the GIS-based management component, it monitors the status of the entire network in real time. It provides plug-and-play functionality for masses of devices and automatic service deployment, reducing the OPEX by more than 50%.

Huge Predictive Maintenance Market, Extending Industry Value Chains

As the core driver to keep up with demand for service transformation, predictive maintenance reduces maintenance costs and creates value. The transformation in services and operation modes driven by predictive maintenance extends industry value chains and makes manufacturers more profitable.

According to statistics from a third-party organisation, predictive maintenance shortens the duration of unexpected service interruptions, reducing risks of VAS operation on devices. Equipment suppliers can transform their business model from equipment sale to long-term equipment leasing and maintenance services. For example, through remote predictive maintenance, a controller platform can effectively monitor the running status of mechanical sugar beet harvesters, improving their reliability and lowering the machine failure rate by over 70%. Continuous proper machine running ensures high-quality output, which strengthens enterprise competitiveness.

Additionally, predictive maintenance can collect real-time running status data on devices in different areas and environments and establish a product database based on in-depth data mining and analysis. Such a database helps vendors optimise their product design to reduce defect rates, and provides sufficient quantitative support for digital precision marketing. Predictive maintenance plays an important role in transformation of vendor operation modes.

So far, the EC-IoT Solution has been successfully applied to fields such as elevator IoT, power IoT, city and lighting IoT, smart energy, smart manufacturing, engineering machinery, and the Internet of Vehicles (IoV). The EC-IoT Solution will continue to revolutionise the predictive maintenance market and enable extraordinary service transformation.