Steps to Consider Before Launching Your Predictive Maintenance Program
Maintenance is an integral part of life in order to keep things in working order. It is especially important in an industrial setting where equipment and machines work tirelessly. Regular preventative maintenance has been a semi-science balancing objective and subjective considerations when determining service requirements. For example, a certain part should be replaced after X number of rotations or hours in use. However, this methodology leaves the door open for unplanned downtime when an issue occurs. These unplanned events cost companies more time and money than scheduled maintenance.
With the transition from basic computerization to Industry 4.0, data becomes available to drive predictive maintenance. Yet, a few steps must be taken to prepare your operations for this new process. The benefits are certainly worth the effort and smart factories are taking full advantage.
What is Predictive Maintenance
First, let’s define what predictive maintenance is. According to the VDMA, the largest network organization in European mechanical engineering:
“Predictive maintenance (PM) is one of the key innovations brought forth by Industry 4.0. Based on continuous measurement and analysis of machine data, PM makes it possible to forecast necessary maintenance activities or upcoming machine errors. Critical operating parameters can serve as decision measures to optimize the timing of maintenance and define operating statuses. PM underlies a long existing concept called “condition monitoring”, which collects real-time information about the operating status of equipment and their component being monitored.”
Innovation in the area of smart sensors, real-time data transmission and powerful cloud platforms enables the processing of big data and analyzing of it. VDMA notes that with fault pattern analysis “using stochastic algorithms, it is now possible to identify, simulate and interpret patterns in real-time sensor data.”
With these enhanced or new components in place, predictions about service life can be calculated more accurately. This information can improve every aspect of maintenance and service for equipment operators and service partners as well as OEMs who use the data to update their products, or design new ones.
The benefits of determining the health condition of equipment and predicting when maintenance should be performed are operational and strategic. The implementation of predictive maintenance solutions can lead to:
- cost savings
- better resource management from inventory to staff time
- higher visibility
- less downtime or the increased availability of the assets
Focusing on a simple example for predictive maintenance, consider the temperatures of an essential equipment for several manufacturing processes. If this equipment unexpectedly overheats, continuing to run operations could cause unexpected delays, leading to additional costs and customer disappointment. This example issue can be avoided by gathering temperature readings and other sensor values of this equipment, which have a critical impact on the whole asset. This real-time information, in combination with the maintenance history of the equipment, can then be used to develop a model that can predict the next breakdown of an asset or to detect a significant anomaly regarding the current condition of the asset.
The Way to Predictive Maintenance
There are some prerequisite steps that should be taken before launching your predictive maintenance program. Generally, they can be categorized into the following areas:
- IoT Connectivity
- Maintenance History
- Maintenance Strategy
- Master Data
- Predictive Analysis
IoT connectivity is essential for creating, capturing and transmitting the relevant data needed for analysis. Smart PLCs, sensors and more offer real-time conditions of equipment, inventory and operations. The information can be sent via Bluetooth, LAN or Wi-Fi to the cloud for access and reporting. Your operations may have completed this step as part of your continual upgrade and replacement strategy. However, conducting a connectivity assessment will reveal any potential gaps in coverage.
Maintenance history provides the records that are a framework for the asset lifecycle management foundation. It should include information on the equipment as well as defects, downtime, regulation infractions, safety incidents and more. Maintenance strategy helps you determine which assets or group of assets are critically important and what action is best suited to mitigate risks to their operation, such as run to failure or predictive maintenance. Ideally, the benefits of the selected solution should outweigh the investment.
Master data modelling of your equipment and its structures ensures quality information that can be used as an input for machine learning tasks. Predictive analytics are an essential part of the algorithms included in machine learning. They help construct models based on historical training data so rules can be inferred. Each newly arrived data can be used to be applied on the trained model to achieve live predictions of your assets. Not only recent predictions are possible, but the trained model can be adjusted based on the new data to increase its predictive capabilities in the future.
There are other considerations to include that are part of any significant transition. One is getting buy-in from major stakeholders as well as line personnel through discussions and education. Training should be an integral part of the implementation progress to provide confidence in the changes as well as learning new processes.
SAP Predictive Maintenance and Service
SAP understands how manufacturing works and the challenges your business faces. Therefore, they developed the SAP Predictive Maintenance and Service (PDMS) solution as part of their SAP Intelligent Asset Management suite. It is a highly integrated suite of products supporting next generation maintenance and service operations. This solution replaces time-based maintenance of industrial assets with predictive and prescriptive maintenance. This type of maintenance relies on information from physical sensors paired with corresponding algorithms and machine learning models.
SAP PDMS enables you to:
- Collect and analyze sensor data from physical assets to predict operational failure
- Initiate preventive countermeasures that trigger service or maintenance orders
- Optimize asset performance with a closed-loop maintenance and service process
- Create digital twins of industrial assets based on a real-time and predictive analysis
Do you want to learn more about implementing your own program? Download our Guide “Predictive Maintenance: 5 Steps to Consider Before Launching” to understand how to implement Predictive Maintenance for your department or company. If you have any questions, contact our IAM experts today.