We’re back for round two in a series offering practical advice to prepare your business for the Industrial Internet of Things (IIoT). In this context, IIoT entails connecting and monitoring equipment, vehicles, and other hard assets within a company’s aggregated information network to enable additional organizational insights, efficiency, and control. The last article summarized the value proposition for IIoT. This article presents an approach to select and implement an IIoT solution to connect your enterprise.
The recommended IIoT implementation strategy is typically iterative and consists of the following five stages:
A business case with costs, benefits, and a payback period is usually the first step to justify an IIoT implementation project. It’s important that the business case maps the organization’s strategic objectives to the expected benefits delivered by an IIoT implementation. This allows the project team to determine where to focus their effort during implementation and facilitates measuring the actual benefits delivered when the IIoT system and processes are operational.
You should first identify what organizational insights are needed to fulfill the goals of the business case. This will determine which company assets need to be monitored and what sensors, network infrastructure, and IIoT platform are required to capture the intended data (e.g. temperature, tank level, etc.). Massive amounts of data can be generated, especially if many devices are connected or if the sampling frequency is high (e.g. one second between sensor readings). A scalable data storage, migration, and archive strategy are important, which should include whether the data will be hosted in the cloud. The platform should be flexible to allow additional sensors and other functionality over time.
A lot of attention is given to what is being implemented, but it’s just as critical to determine how your project will be implemented. The overall approach should be iterative so that real-world feedback can be obtained frequently and to allow complexity to be added over time. Throughout the implementation project, scope decisions should be based on the expected value delivered as defined in the business case.
One of the first decisions you’ll face is whether you will self-install the system or have a third party install it. It’s possible you will have thousands or millions of sensors to deploy, so planning the logistics is crucial. The supplier of your system should provide implementation support, including installation instructions, training materials, and on-call technical expertise.
The data captured from the sensors must be understood at both the individual device level and at an aggregate level. To understand the overall system and the impacts on the business, data from each device should be aggregated alongside other organizational data and summarized to provide meaningful operational and organizational insights. As more data is captured, pattern recognition based on predictive models can provide a new level of insight to enable more proactive decision making. You’ll want to make sure the system you select supports predictive models.
Gleaning intelligible insights from massive data volumes can be daunting, so it’s important to access or build an internal data analytics skill-set or seek outside help. Once you have a good understanding of the data, there may be opportunities to collaborate with other organizations for benchmarking, or to incorporate contextual data in value-added ways. For example, weather data could be used to predict future temperature measurements.
Defining how you’re going to react to system data is an important part of realizing value from the system. You’ll want to make sure you have the ability to generate custom reports based on your unique and evolving requirements.
There are two types of system interaction: pull and push. For example, real-time alerts sent by email, text, and phone are push notifications that can notify you when a sensor threshold is breached (e.g. temperature is too high), allowing corrective action to be taken. Another push-based interaction comes in the form of scheduled reports. These reports can include information about system performance, compliance, quality control, and equipment reliability. Pull interactions include on-demand export of data, as well as on-demand reporting.
The challenges of implementing a comprehensive IIoT solution can be minimized with advanced planning and an iterative and flexible approach. The next article in this series will focus on important IIoT considerations, such as setup, maintenance, data security, and organizational impact.
Subscribe to our blog to get the next post in this series straight to your inbox.