Manufacturers have reached a crossroads. Increased connectivity and machine data have made traditional operations and maintenance practices outdated, costly and even high-risk. As a result, companies are looking for new ways to harness and analyze this data. Over the next five years, PwC predicts manufacturer’s adoption of machine learning and analytics to improve predictive maintenance will increase 38 percent.
While letting go of deeply ingrained practices is always challenging, manufacturers are not alone in navigating today’s digital transformation and identifying new ways to operate. In the consumer space, many car owners, for example, still follow the ‘golden rule’ for an oil change: schedule an appointment every 3,000 miles or three months. This age-old advice, however, is just that – old.
In the U.S., 87 percent of vehicle owners pay for oil changes, costing Americans billions per year. Drivers incur unnecessary costs when adhering to the 3,000 mile myth as it doesn’t account for various conditions that impact oil quality and availability, such as automobile design, driving environment, tire pressure, and average trip duration and weight carried.
More drivers are embracing newer vehicles with engines designed and built to exact standards that require specific oil blends at varied intervals. To determine whether a car needs an oil change after 3,000 miles or after 7,500 miles, service technicians can now use advanced machine sensors, historical maintenance data and service reports to yield a more accurate timeline and advanced warning.
When both service history and machine data are combined, maintenance and performance become more effective. This integration of advanced technology and human insight in manufacturing results from combining asset performance management (APM) and field service management (FSM) for enhanced predictive maintenance. Predictive maintenance has been top of mind for manufacturers since the emergence of the Industrial Internet of Things (IIoT), but it has often been executed in silos: the collection and analysis of asset data or APM, and work order execution of maintenance by technicians or FSM. For manufacturers, the integration of the two is key to modernizing maintenance practices and optimizing operations for cost savings.
Together, FSM and APM unite three important advancements: service in an IIoT infrastructure, connected machinery with real-time data feeds and condition alerts, and advanced predictive models that leverage machine learning to calculate where and when assets fail and maintenance should occur.
Service in the IIoT Infrastructure
As manufacturers know all too well, they can’t afford to have technicians inspect equipment each day to prevent failures. Not only is maintenance and machine downtime expensive, these inspections are also still very manual and time consuming. Until recently, however, inspections have been a top source for historical equipment performance data.
Now that machines are being equipped with advanced sensors and control systems as part of the IIoT infrastructure, technicians have the ability to automate data collection and run analytics to assess equipment in a much less labor-intensive manner – relying on mobile devices, sophisticated software algorithms and virtual replicas of assets to make informed decisions.
Data analytics is becoming a critical component of service practices to increase productivity and reduce costs related to failures. APM on its own brings equipment data and operations readings into a predictive failure model. Similarly, FSM captures the execution data recorded by technicians and accumulates KPI measurements. The true potential for analytics that Manufacturers ultimately want requires robust APM and FSM programs that work together through unified data collection and transparent access and communication.
One customer in the medical technology industry, that manufactures and sells medical devices, implemented eight different versions of an enterprise resource planning (ERP) solution while lacking data standardization and benchmarks to measure performance against. Without consistent and clear performance data, the company had difficulty identifying and tracking failures. This created increased risk in cases where an infection emerged in a hospital, because the company’s technology was relied on to test, confirm and alert the hospital to infections. If the microbiology machine didn’t function properly, the entire healthcare ecosystem would be put at risk. After implementing more effective field service management software and integrating this data into service operations, the company was able to develop better performance indicators and reduce service costs across products by 30 percent – a $13 million reduction in one year. The IIoT framework that manufacturers now function within enables the integration of APM and FSM for real-time monitoring, smart decisions and efficient service.
Real-time Data Feeds and Condition Alerts
The emergence of real-time data feeds and more intelligent condition alerts have provided major enhancements to predictive maintenance. These advancements help directly integrate the APM and FSM technologies in a closed-loop process where data is available in APM systems and shared to service technicians with detailed assessment on asset conditions and maintenance needs. Alerts from an industrial vehicle, for example, signaling low pressure with high exhaust temperatures may align with prior obstructions in the airflow system. Service technicians will pull the data from the alerts with the history of prior issues and prepare for the required maintenance in advance to deliver more efficient maintenance.
Beyond the service teams, company executives can use failure reports to identify issues that require more strategic machine replacement or better align service scheduling across products. One global manufacturer of cancer treatment equipment had a tendency to rely on local service team guesswork and assumptions for maintenance schedules that resulted in dramatically varied service rates worldwide and higher costs. Machine failure required seven days of downtime, delays for high-risk patients in need of treatment and lost medical costs for hospitals. Once the company began collecting data and harnessing the IIoT, the company analyzed failure patterns and identified that one sensor provided an indication for failure 13 days in advance. The company then used this data to optimize service visits, significantly reducing the service schedule disparities worldwide and saving money on maintenance.
Advanced Predictive Models
As computing capabilities, machine learning and even artificial intelligence become fully integrated into APM, predictive models for maintenance will become more exact. According to McKinsey, enhanced predictive maintenance allows organizations to avoid machine failure by combining data from IoT sensors, maintenance logs and external sources – ultimately increasing asset productivity up to 20 percent and reducing overall maintenance costs by 10 percent.
Combining data and maintenance logs through the integration of APM and FSM makes predictive maintenance infinitely smarter, because when factoring into the model what actions were taken by technicians that yielded the quickest return to operation, any future maintenance work orders will incorporate these recommendations.
A high-speed printer manufacturer in the U.S. sells its products to top food and beverage manufacturers for printing expiration dates, which is highly regulated and controlled for safety. The company was often called to service printers at manufacturing plants only after failure occurred, causing product loss and delaying further food and beverage production. By implementing APM and FSM, the company was able to rely on advanced predictive models for a more accurate maintenance schedule prior to failure. The company also developed a better understanding of optimal machine performance and lifespan across its customer base.
Manufacturers implementing IIoT solutions have primarily focused on connecting assets and ingesting data. As a result, they often fall short in assessing the risk of failure and the maintenance cost. This is why the combination of APM and FSM is a perfect marriage – together they close the automation gap in service. Maintenance recommendations can automatically generate work orders for service technicians, and asset service history from the field is fed back into the APM system, creating a feedback loop. The APM system can factor in service history information, including prior failure incident details, part replacements and breakdown rates, to the performance and usage data it’s already monitoring. Ultimately, this integration optimizes maintenance schedules and puts maintenance activities in context. This closed loop will enable true predictive maintenance. With more precise data analytics and recommendations, inspections and maintenance activities will be scheduled dynamically so that manufacturers can optimize asset performance and focus more on achieving business outcomes.
Scott Berg is CEO of ServiceMax, an operating unit of GE Digital.