In both operational and financial measurements, the Chemical Industry has outperformed other sectors in the last 15 years. In recent years IIoT has emerged as a promising technology for Chemical to further improve these two benchmarks. However, as global recessions and supply chain disruptions caused by Covid-19 start to take effect, the business case question has fundamentally changed. Can Industry 4.0 mitigate against the economic loss in this unstable environment?
Is the Chemical Industry Willing to Embrace Industrial Analytics?
In a survey of 222 chemical industry executives, PwC showed a progressive approach to automation. Of the survey results, three are particularly noteworthy:
- Companies anticipate re-investing 5% of revenues in a digitalization process in the next five years.
- 75% of respondents predict a “high level” of digitalization in the next five years.
- Perhaps most tellingly, 88% of respondents indicate that in five years there will be a high significance for the gathering, analysis and utilization of data for decision making.
Taken together, the PwC report indicates a commitment on the part of plant executives to adopt Industrial Analytics solutions. This gradual paradigm shift is an important first step.
Industrial Analytics for Predictive Asset Maintenance: A Short Introduction
It was not long ago that Machine Learning was primarily a topic of studies within the confines of academia. However, in the last few years there has been surge in commercial applications, especially in financial services and the industrial domain.
What has changed? First, there have been substantial price drops in the cost of storage and transfer of Big Data as well as computational power. Second, as a result of cloud technology advances, Artificial Intelligence applications are more easily accessible and affordable. Third, as indicated by many reports including the one by PwC referenced above, there is a top-down effort on the part of senior executives to strategically leverage Industry 4.0 to improve growth and productivity.
Industrial plants have thousands of sensors embedded in machine equipment. Most of the data generated from the sensors is not captured or utilized in a meaningful way. The exception is the SCADA-based monitoring of small number of high priority sensors (e.g., temperature or vibrations), which is used to audit the health of a machine.
The traditional SCADA approach monitors whether human-set control thresholds have been breached. For example, if the threshold is set at 40 degrees and the machine temperature surpasses 40 degrees, this over-heating will prompt further actions. With Machine Learning, an algorithm is taught to identify abnormal data patterns and correlations between patterns, irrespective of whether the control thresholds have been breached or not. Advanced AI enables the measurement of Time to Failure (TTF) and the application of Root Cause Analysis (RCA).
Within Machine Learning there are multiple methodologies for failure detection. Perhaps the best known is one is the Digital Twin, a type of Supervised Machine Learning. This form of Machine Learning relies on physical modeling of the machine and involves mechanical engineers and data scientists working together to build the representing machine model.
Another PdM option that is gaining a growing following is Unsupervised Machine Learning. Instead of relying on human expert knowledge, vast amounts of data are analyzed and the algorithm itself generates the machine’s digital model based on statistical features detected and extracted from the data.
The differences between Supervised and Unsupervised Machine Learning are not just methodical. Supervised Machine Learning demands more resources, which in turn affects the ROI on Industrial Analytics solutions based on this system.
How Industrial Analytics Can Improve Production Output
A major challenge for chemical plants is the cost associated with unplanned machine downtime. As companies do not publish unscheduled downtime information publicly, it is often difficult to determine asset failure for the industry as a whole. Nevertheless, there are certain indications. According to Accenture’s ICIS Chemical Business, “missed profit opportunity alone, on a major cracker shutdown in the US Gulf Coast, was $1.4m per day per worldscale cracker.” Moreover, a calculation by the Aberdeen Group indicates that 2% to 5% of production is lost in the petro-chemical sector.
The chemical industry has achieved productivity gains over the past few years, thereby raising the bar for incremental gains.
As we noted above, Industrial Analytics offer chemical plants two important, complementary pieces of information. First, when an alert is generated with an accurate Time to Failure, plant maintenance staff can schedule repair so that it least disrupts production. Second, Root Cause Failure Analysis helps decrease the probability of similar failures occurring elsewhere in the plant or in other company plants.
It’s a simple but powerful equation that is spurring adoption of Industrial Analytics across industries: lower machine downtime means higher yield rates and added revenue.
Lower Operations and Maintenance Budgets
Reducing O&M budgets is an oft–stated goal of many plants. Until recently, this goal was hard to achieve. Industrial Analytics reduces O&M spending by shifting resources away from Reactive Maintenance activities.
The closer the repair activity to the repair incident, the greater the cost of maintenance. When asset failure is identified close to occurrence, it can lead to a disruption of production and of other repair activities. Tight timelines are a source of pressure on the crew. Without adequate data about the root cause of failure, they may have to rely on trial and error.
With reactive maintenance, there is more of a chance of inefficiencies. Technicians are left idle as they wait for spare parts to be delivered or for extra resources to be brought on site. The combination of TTF and RCA enable planned repair activities, and therefore lower O&M budgets.
Another factor that is occasionally overlooked but that also impacts O&M budgets is the relatively high pricing of spare parts for the chemical industry. Redundant practices such as maintaining needless spare parts and conducting excessive Preventive Maintenance can also be reduced when there is a more exact and timely view of machine deterioration.
Upgraded Occupation Health and Safety
Unsurprisingly, OHS standards are severe for the Chemical industry. The United States has proven indications that manufacturers are adhering to these tough standards. According to the Bureau of Labor Statistics, the chemical industry has one of the best annual performance records in the US.
Nevertheless, fatal tragedies in chemical plants such as the infamous Bhopal Toxic Gas Leak have been at least partially ascribed to deficient maintenance practices. These cases serve as a harrowing reminder of what can happen when maintenance goes awry.
The implementation of Industrial Analytics improves occupational safety in different ways. The most obvious improvement is intuitive: Transitioning from Reactive to Predictive Maintenance leads to less accidents.
By using learning algorithms to replace the majority of human inspections, Industrial Analytics also reduces the risk of reliability workers getting injured as they physically assess machinery. Just as robots and drones decrease the likelihood of factory workers getting hurt on the job, deploying a Machine Learning analytic solution prevents unnecessary and potentially dangerous maintenance from taking place in a chemical plant.
Reactive and Preventive Maintenance will continue to fill important roles on the factory floor. Nevertheless, migrating towards Machine Learning-based Predictive Maintenance can provide greater safety for O&M workers.
Prolonging Machine Lifespans
On average, North America and Europe have substantially older chemical plants than their counterparts in the Middle East and Asia.
The familiar Bathtub Curve graph below demonstrates the correlation between machine failure and machine age. As Accenture notes, aging equipment contributes to the rising frequency of unscheduled downtime.
Industrial Analytics provides chemical plants with early alerts of evolving machine health issues. As a result of this advance warning, reliability professionals are able to act before the machine shuts down. Moreover, maintenance crews can get a clearer picture of the underlying sources of failure using Root Cause Failure Analysis.
If maintenance and reliability standards are improved, the lifetime of a chemical machine asset can be prolonged.
It’s important to remember that Industrial Analytics is only one part of a greater initiative to extend the life a chemical asset. Other steps to combat deterioration include using equipment as designated and performing heavy maintenance repairs.
There are certain preliminary actions that need to occur before chemical plants can begin to take full advantage of Industrial Analytics. For example, many chemical plants are not tracking or storing their sensor data – without data, there can be no Machine Learning-based solutions.
In recent years the Chemical Industry has made strides in terms of operations and finances. In light of this, there is reason to hope that industry executives will identify Industrial Analytics’ potential to boost productivity, increase worker safety, and grow revenue.
Industrial plants generate terabytes of process data. SKF Enlight AI is a SaaS Predictive Maintenance solution that uses Automated Machine Learning to identify emerging asset failure patterns within this data. It provides early warnings and sensor-level intelligence to help avert unplanned downtime and meet production goals. For more information on how SKF Enlight AI can improve performance and reliability, click here.