• Written by Sonali Subhadarsini, Product Manager

This week, the shift supervisor at a chemical manufacturing plant needs to decide whether to adjust the production schedule to meet a surge in demand, and they currently base this decision on a weekly report that summarizes the previous week's yield and equipment performance, but this report often arrives too late to inform timely decisions, and even then, it may not provide the level of detail needed to make an informed choice, such as the specific batches that were affected by equipment downtime or the quantity of raw materials that were wasted due to quality control issues.

The operations questions that currently wait on a report

The shift supervisor is not alone in waiting for reports to answer operational questions, as many teams in the chemicals industry rely on periodic summaries of data to inform their decisions, such as the category manager who needs to know which suppliers are consistently delivering high-quality raw materials on time, or the maintenance team that wants to identify which equipment is most prone to breakdowns and prioritize their maintenance schedule accordingly, and while these reports can provide valuable insights, they often do not arrive quickly enough to address the fast-paced and dynamic nature of chemical manufacturing, where small changes in process conditions can have significant effects on yield and safety.

For example, a shift supervisor checks the maintenance log to see if any equipment is due for scheduled maintenance, but the log may not be up to date, or may not provide enough detail about the type of maintenance that is required, and the supervisor may need to wait for a report from the maintenance team to get a clearer picture of the equipment's status, which can delay decision-making and potentially impact production schedules.

The data behind them: process and exception records, project status, budget consumption, supplier performance

The data that informs these operational decisions is often scattered across different systems and departments, including process control systems, enterprise resource planning software, and project management tools, making it difficult to get a comprehensive view of the current state of operations, and even more challenging to identify trends and patterns that could inform future decisions, such as the impact of changes in feedstock quality on yield, or the effect of equipment maintenance on downtime and production capacity.

The category manager, for instance, may need to compare the performance of different suppliers, including their delivery times, product quality, and pricing, to determine which ones to use for upcoming orders, but this information may be stored in separate databases or spreadsheets, requiring manual effort to compile and analyze, and potentially leading to errors or inconsistencies in the decision-making process, and most people get wrong the idea that this data can be easily centralized and analyzed, when in reality, it requires significant effort and resources to integrate and make sense of the data.

Sector constraints: yield drift and feedstock variability

Chemical manufacturing is a complex and highly variable process, with many factors that can affect yield and safety, including changes in feedstock quality, equipment performance, and process conditions, and teams must navigate these constraints to optimize production and minimize waste, but this can be a challenging task, especially when faced with tight yield and safety envelopes, where small deviations from optimal conditions can have significant effects on product quality and safety.

For example, a change in the quality of a raw material can affect the yield of a batch, and if not detected and addressed quickly, can result in significant waste and potential safety risks, and the shift supervisor may need to adjust the production schedule or process conditions to compensate for the change in feedstock quality, but this requires timely and accurate data on the current state of operations, which may not always be available.

What improvement looks like: on-time delivery, exception rate, expediting cost

Improving operations in chemical manufacturing can have significant benefits, including increased on-time delivery, reduced exception rates, and lower expediting costs, and teams can achieve these improvements by having better visibility into their operations, and being able to make more informed decisions based on timely and accurate data, such as using real-time data analytics to identify trends and patterns in production, and anticipate potential issues before they occur.

The category manager, for instance, may be able to reduce expediting costs by identifying which suppliers are most reliable and prioritizing orders with them, and the shift supervisor may be able to reduce exception rates by detecting potential issues with equipment or process conditions before they result in downtime or waste, and by taking proactive steps to address these issues, such as scheduling maintenance or adjusting process conditions, and for more information on how to achieve these improvements, teams can contact us to discuss their specific needs and challenges.

What this does not do, or when it is the wrong choice

While AI for operations in chemicals can provide significant benefits, it is not a solution for all operational challenges, and teams should carefully consider their specific needs and constraints before implementing an AI solution, such as whether they have the necessary data infrastructure and resources to support the solution, and whether the solution is compatible with their existing systems and processes, and in some cases, other approaches, such as process optimization or training and development programs, may be more effective or efficient, and teams should weigh the potential benefits and costs of an AI solution against these alternative approaches.

For example, if a team is struggling with equipment maintenance, an AI solution may not be the best choice if the underlying issue is a lack of trained maintenance personnel, or if the equipment is outdated and in need of replacement, and in these cases, other approaches, such as training programs or equipment upgrades, may be more effective in addressing the root cause of the issue, and teams should consider these factors when deciding whether to implement an AI solution, and what specific benefits and outcomes they can expect to achieve.