Lean Six Sigma is one of the most well-known and most used methods for transforming organizations into data-driven and continuously improving companies. However, one of the challenges Operational Excellence executives face is getting their entire organization to be actively involved in continuous improvement programs.
Most companies have multiple ongoing projects, but generally there are only a handful of trained team members who are proficient enough in the required statistical methods for improvement projects. With few experts in this field, bottlenecks quickly arise. Additionally, many projects aim at addressing production performance issues, which require input from busy process and asset experts. As continuous improvement experts need to rely on subject matter experts (SMEs) for elaboration on process intricacies, many projects go unfinished or unaddressed due to these SME’s limited availability for deep dive analysis.
Additionally, the organization and all personnel involved can experience detrimental impacts such as:
- Underutilization of the local process expertise (at the plant level).
- Missed improvement opportunities.
- Long project cycles.
- Lack of smooth symbiosis between plant and central level stakeholders.
- Potential financial losses for the organization.
These challenges can be addressed when pre-work is completed by SMEs who can analyze the production data themselves and thus contribute to the central improvement projects or even reduce the number of projects by finding improvements on their own. As SMEs deal with more of the low-hanging-fruit opportunities, the central group and their resources are free to work on more complex projects in collaboration with the SMEs. This approach accelerates improving operational performance at scale.
Allowing process and asset specialists to contribute to these projects dramatically increases the operational improvements needed to meet the expected organizational goals. By fully leveraging captured time-series data and more efficiently utilizing the operation expertise of process experts, companies establish a cornerstone to maximizing organizational continuous improvement.
Providing a common and accessible approach for analyzing process data
The structure of the DMAIC (define, measure, analyze, improve, control) cycle is well-suited for data-driven analysis, but previously the tooling was not up for the challenge. What is needed is a way to provide a common and accessible approach for analyzing process data between central and plant level team members that significantly lowers the threshold for starting improvement projects. The data gathering, modeling and analyzing may be a hurdle for many asset and process engineers when using the DMAIC method for continuous improvement, but new approaches such as self-service industrial analytics are now available to overcome these challenges. This tooling accelerates discovery through process data analytics by enabling non-data science experts throughout the organization to capitalize on their domain knowledge.
Self-service industrial analytics empowers process experts, so they can perform analytics themselves to monitor, analyze and predict their processes. It combines the necessary elements to visualize and gain insights into the relationships hidden in time-series data from process historians. They will be able to overlay similarly matched historical patterns and enrich the data with contextualized process information. Additionally, process experts can use self-service analytics to fingerprint best operating zones. These operating zones are especially useful in batch processes, startups/shutdowns, grade transitions and monitoring bivariate relationships such as tower differential pressures, pump curves or valve wearing. Process experts can also set alarms when production deviates from these areas and send out alerts to personnel in time to take corrective action to prevent issues.
With this accessible and user-friendly approach to analyzing plant data and monitoring plant operations, every team member knowingly or sometimes even unknowingly contributes to corporate KPIs through continuous improvement projects. More importantly, there is a high utilization of all skill sets allowing for the establishment of common workflows with a consistent set of tools suited for all the process experts. One example of how a global process manufacturer used self-service analytics to realize continuous improvement is discussed below.
How DSM reduced completion time for Six Sigma projects
DSM Dyneema is the inventor and manufacturer of ultra-high molecular weight polyethylene (UHMWPE), branded as Dyneema, the world's strongest fiber. One of the corporate programs at DSM Dyneema focuses on the use of Six Sigma and the DMAIC cycle as the go-to approach for improvement projects. The company set a target to increase the number of Six Sigma projects, with average Six Sigma project duration amounting to 5-6 months.
By mapping the self-service industrial analytics functionalities to the different phases of the DMAIC cycle, it became clear that the tool can be applied to most of the phases and could considerably reduce the time needed to complete each phase (Figure 1). In a first test case, the Define and Measure phases were completed with self-service industrial analytics in two afternoons, instead of two weeks.
To better understand the value and contribution of self-service analytics to continuous improvement projects, the following use cases will be discussed.
Case study 1: European wastewater treatment plant more effectively monitors pump operation states
Pumps are critical assets for the process industry. Keeping these assets in operation is necessary to keeping processes running. A breakdown of an important pump can lead to serious problems, so to avoid an unplanned shutdown, process experts must regularly maintain them. To maintain pumps more efficiently, process manufacturers should have access to a user-friendly, reproducible and reliable approach for monitoring the operation state of important assets. A best-practice is to leverage captured sensor-generated time-series data. This data includes information related to pressure difference, vibration, flow rates and shock pulse measurements (SPM) and is available for historical analysis.
In a recent use case from a European wastewater treatment plant, self-service analytics was used to create an indicator to trigger a monitor. The pump in question is supposed to deliver collected wastewater to a distillation column. The wastewater is stored in various storage tanks, which serves as buffer inventory. The distillation column processes the wastewater and removes volatile components, such as alcohols or other solvents. Process experts can look at pressure difference as a good indicator of the pump operation state and as the basis for further analysis and monitoring.
Since the pressure difference had a lot of noise (yellow trend in Figure 2), process experts used an aggregation to smooth out the trend, as seen in the annotation in Figure 2. The other red tag displays the flow rate, which is roughly kept constant during operation of the column. The pressure difference rises throughout its lifecycle until the next maintenance cycle, which can be visually seen as the sudden drops in the flow rates and pressure differences. By knowing this behavior and by using self-service analytics, the expert can easily and autonomously create a monitor to alert staff of the upcoming maintenance needs, thus preventing unplanned failures.
Case study 2: Huntsman sets soft sensors to monitor product quality
Huntsman Corporation is a family-led publicly traded global manufacturer and marketer of differentiated and specialty chemicals. Its digitalization journey started some years back, enabling the company to secure quick and valuable wins by leveraging time-series data. As operational data was kept in separated silos, this added an extra challenge to improve operational excellence and was successfully overcome by using self-service industrial analytics. This platform was selected to help the process and asset experts to continuously improve their processes and contribute to corporate operational goals.
For years, a Huntsman continuous isocyanate plant had been collecting daily process and offline-created lab analysis data that was stored in the historian database. Early in 2016, the company’s teams used self-service analytics to build soft sensors based on operating conditions to predict product quality for certain isocyanates. Additionally, the process experts used these sensors to make micro-adjustments to process setpoints to pro-actively minimize impurity levels. For example, one of the monitors predicted off-spec hydrolysable chloride levels in the final product. By adjusting vacuum pressure conditions, product quality was ensured. In addition, monitors were set up to send out early warnings alerting operators not to load trucks, thus preventing off-spec material from going out to a customer.
By being empowered by analytics capabilities, Huntsman process experts established 24/7 quality control compared to quality control with lab analyses that was only available for regular weekday work hours. With trucks being sent out 7 days per week, the soft sensors eliminated 75% of the expensive off-spec transportation cases which occurred on the weekends. As a result, a significant positive impact on lead time was achieved as unnecessary wait hours for in-spec product were eliminated, with the average lead time being reduced by several hours. Finally, the extra insights into product quality also reduced the demand on lab resources as the number of uncertain situations for this specific product was reduced by as much as 10%.
Conclusion
Whether a company is seeking to reduce maintenance costs, improve asset reliability, improve plant safety, reduce energy consumption and carbon footprint, or keep better process records, the application of self-service industrial analytics within any continuous improvement project will help organizations achieve operational excellence. These companies can therefore reach their corporate goals faster and more efficiently. Ultimately, self-service industrial analytics utilizes the most valuable resources of any industrial manufacturer: the process and asset experts and their process knowledge.
Nick Petrosyan is a chemical engineer whose passion is solving problems through collaboration and data driven decision making. As a Customer Success Manager at TrendMiner, Nick draws on his extensive experience in manufacturing and data analytics to lead customers through use case resolution. Nick holds a Bachelors of Science in Chemical Engineering from the University at Buffalo.
TrendMiner
