The Internet of Things (IoT) is creating an unprecedented level of opportunities across verticals helping companies monitor the performance of assets. Businesses have leveraged the IoT technology to make or save money, positively impacting the bottom line. However, a major barrier to adoption is the most viable use case selection getting started. Predictive maintenance (PDM) is a popular and proven use case of IoT in the industrial setting but then PDM is not a metric neither does it directly account for productivity. According to McKinsey, operations optimization provides the greatest potential for creating value. Plant owners have been able to reduce cost and eliminate waste using Lean, Six Sigma processes. IIoT complements these advancements helping plant owners increase productivity. This article will help businesses looking to get started with IIoT establish proof of concepts (PoC) and pilots leveraging the issue tree.
The issue tree is a problem solving technique used by top consulting firms- McKinsey, BCG and likes. It helps create more questions and answers to a basic question. It is important to note here that before creating an issue tree, companies need to understand that data strategy and IIoT deployments needs to stay true to a business outcome thus a business problem which is the basic question needs to identified. Often times, this question/problem is a deficiency of a key metric. A problem solving worksheet can be developed to evaluate the vision, best practice, solution space, risks to manage, context, success criteria, stakeholders and sources of insight. When this is done, an issue tree can be developed. Asset Performance Management (APM) is a key metric understood by production staff, senior management and C level executives. It can be presented in different contexts, however, all contexts points to how assets are contributing to business profitability.
A sub metric used in APM is the Overall Equipment Efficiency (OEE), the OEE has got three sub metrics- Availability, Performance and Quality. Evaluating the OEE sub metrics will help identify other losses which help in identifying an IoT proof of concept most viable. Considering availability which is the percentage of available runtime compared to the time production equipment is scheduled to operate. Availability enables answers to questions as these:
1. Are there equipment failures? If yes, which and when.
2. Are there stoppages due to set ups, adjustments and changeovers? If yes, what frequency and what duration.
Equipment performance is the process rate for same equipment. It is the speed at which production runs as a percentage of its ideal cycle time. Answering the performance questions will identify answers to questions as these
1. Are there idle equipment or stoppages? If yes, which
2. Are there products which required stoppage?
3. Any material caused delay?
4. Are there slowed cycle times?
5. Are there operator challenges?
Quality is the ratio of total unit produced to scrap/rework. It is a measure of profitable production yield. Considering the quality sub metric will answer the following questions.
1. Are there process defects?
2. Are the production procedures being followed?
Answering these questions will create more answers to additional questions. The questions and answers are evaluated in choosing which PoC to establish with IoT. This means a PDM solution can be deployed to eliminate failure in order to increase availability or to reduce slowed cycle times to increase performance. A product monitoring solution can be deployed to identify defects or procedures not being followed, this helps optimize inventory and logistics. It is quite important to evaluate each of these metrics critically as their impact on bottom line differs. For instance, whilst there may be stoppages and failures, there may be standby equipment to continue operation but defect products directly impacts revenue. This is because material is wasted and inventory forecast is affected thus affecting sales. Examining the impact on bottom line also helps in justifying IoT investment.
In the above diagram, the brown boxes indicate a YES answer to problem question and the lemon boxes a NO. There are no quality issues but availability and performance. A good IoT use case pilot will be predictive maintenance as this will help prevent failures and down times also identify the cause of anomaly in the lubricant consumption. Deploying a solution to help increase yield or identify defects may not provide immediate and visible results as a quality problem isn’t high value as seen in the diagram.