Optimization is not the only algorithmic tool that supply chain companies need to solve their planning and decision-making problems.More May 2, 2019
5 steps to achieve predictive maintenance excellence
Authors: Özgün Aydın from ICRON and Lieneke van Boxel from CQM
Predictive maintenance is a hot topic among executives from companies across various sectors – from manufacturing to public transport to mining to logistics. A cutting-edge maintenance strategy, predictive maintenance (which is also widely referred to as condition-based maintenance) harnesses the power of IoT, data, and advanced analytics algorithms to monitor and gauge the condition of a company’s assets (such as production machines, airplanes, trains, trucks, and cranes) on an ongoing basis, predict precisely when these assets will need maintenance, and recommend the optimal window of time in which to schedule and perform maintenance activities.
A growing number of companies are discovering that predictive maintenance-based planning and operations – which allows them to foresee asset failures and carry out maintenance only when it is utterly necessary (rather than conducting routine preventive maintenance according to a prescribed schedule or corrective maintenance after an asset actually breaks down) – can enable them to minimize asset downtime and maintenance costs. Many of these companies are interested in implementing predictive maintenance strategies, tools, and technologies, but they don’t know how to make it happen.
Indeed, it sounds almost magical – the ability to see into the future and predict and prevent asset failures. How – without a crystal ball – is this possible? As it turns out, there is nothing magical about achieving predictive maintenance – you only need to put in place the right people, processes, and practices as well as the right algorithmic, automated planning and optimization system.
In this article, we highlight the five key steps that your company needs to take to enable predictive maintenance-based planning and operations. By following these steps, your company can begin to realize the operational and financial benefits of predictive maintenance.
Step #1: Determine exactly what you want to predict.
As is the case when starting any project, it is imperative to clearly and precisely define the scope of the implementation of the predictive maintenance-based planning system as well as the goals of the initiative. One of the common pitfalls of these types of project is that companies fail to determine exactly what they want to predict – and instead have only a vague, general goal of wanting to predict and prevent asset failures.
Before embarking on the implementation of an algorithmic predictive maintenance-based planning system, your company must identify exactly what you want to predict. You may want to, for example, conduct predictive maintenance for a particular asset (such as a brake system, a heat exchanger, a semiconductor manufacturing machine, a power network, or an airplane engine) or a group of assets. For each asset, you will need to identify which specific maintenance tasks and processes are in scope.
You will then want to make a business case and estimate the expected gains from the project in terms of your overall efficiency and costs – so that you can set clear and measurable goals and also, hopefully, get buy-in from key stakeholders in your organization.
Step #2: Ascertain which data you need and assess the quality of this data.
The next step in implementing a predictive maintenance-based system is to collect historical and real-time data from various sources – including IoT devices, back-office systems (such as ERP, Excel, and MES), and web services – on the past performance, failure and maintenance history, and current condition of your assets.
To ensure that the data is relevant and of the highest quality, implementation consultants working on the project should collaborate closely with domain experts – who have a deep understanding of the data and can help intelligent collate and analyze it.
Step #3: Extract insights from the data.
After collecting relevant data on the past and present performance of your assets, you will need to feed this data into a best-of-breed automated, algorithm-based planning and optimization system – like ICRON – that has a robust advanced analytics engine.
With advanced analytics, you can automatically process your data and transform it into actionable insights, projections, and plans. So, using historical and real-time data on the performance of your assets, the system will be able to predict when the next failure of your assets is likely to occur and will propose the ideal time to perform particular maintenance tasks – to prevent failures from happening, increase asset efficiency, and decrease maintenance costs.
Step #4: Generate optimized plans and schedules that enable predictive maintenance.
Fueled by the data and advanced analytics, the algorithmic planning and optimization system will then automatically generate optimal maintenance plans and schedules for particular assets that:
- Optimize the timing and duration of your company’s maintenance activities,
- Take into account your assets’ past performance and current condition,
- Are seamlessly integrated with your company’s operational or production plans and schedules,
- Optimally group together related maintenance tasks and so that you can perform them simultaneously,
- Ensure that the necessary equipment, materials, and specialized workers necessary to conduct specific maintenance tasks are available at the right places and times,
- Enable you to predict the future performance of your assets, perform preventive maintenance only when it is absolutely necessary, and prevent costly asset failures and breakdowns – thereby reducing asset downtime and maintenance costs and improving customer satisfaction.
To create such optimized, predictive maintenance-based plans and schedules, you will need to invest in and implement a state-of-the-art software solution, like ICRON, which contains powerful optimization, scheduling, and analytics engines as well as machine learning capabilities.
Step #5: Make optimized decisions that ensure predictive maintenance excellence.
Beyond merely being able to create optimized maintenance plans and schedules, the ultimate goal of implementing an algorithmic planning and optimization system is to empower planners and key stakeholders in your company to make optimized decisions about which maintenance tasks and when to do them.
With such a system, you can give your planners and other key stakeholders the power to predict the frequency and duration of future maintenance, schedule maintenance tasks optimally (taking into account the asset’s past performance and present condition, your company’s operational and maintenance schedules, and the availability of resources), and make the best possible decisions on how to utilize your assets to minimize downtime and maintenance costs.
Furthermore, when unexpected asset breakdowns do occur (as they inevitably will) and corrective maintenance is necessary, your planners and other key stakeholders will be able to instantly, optimally revise your plans and make the right decisions to mitigate the potential impact of these disruptions on your operations.
By following these five steps, your company can achieve predictive maintenance excellence.
Of course, each company’s journey to reach this state of predictive maintenance-based planning and operations will be different (and some may be more challenging and complicated than others), but if you are willing to invest in the right technologies and are committed to changing your company’s mindset, processes, and practices, your company will ultimately achieve this goal.
With optimized planning and decision making, your company can discover the “magic” of predictive maintenance. With predictive maintenance, you can get the most out of your assets – whether they be ships or high-tech manufacturing equipment or jet engines or mining drills – and dramatically improve your company’s overall productivity and profitability.