When automation is recognized as a core competitive advantage for a company, significantly enhancing plant safety and profitability, the next critical step is translating this recognition into actionable strategies. Many advanced control projects, due to their substantial investment, require feasibility studies to identify pain points and potential benefits. These studies must start with the process itself and be tailored to specific conditions, demanding a high level of expertise. Often, process engineers lack familiarity with new technologies, and control experts lack deep knowledge of the plant, leading to unclear project goals. When issues arise during execution, this can result in disputes, with outcomes falling short of expectations. Additionally, insufficient resources for maintenance and improvement can lead to advanced control projects yielding minimal returns on significant investments, or even being abandoned. No matter how advanced the tools, if the company cannot derive tangible benefits, sustaining investment becomes challenging. So, how can process and control teams achieve cognitive alignment in automation enhancement projects?
I propose that operational events are key to achieving this alignment. When the process details are not fully understood, operational analysis simplifies the complex challenge of automation enhancement, turning abstract problems into concrete ones. High-frequency operations are prime candidates for automation—the more frequent the operation, the more feasible and beneficial automation becomes. Therefore, whether for process engineers, control engineers, continuous plant improvement, or third-party automation projects, I recommend focusing on operational analysis with the goal of reducing manual intervention. High-frequency operations represent process pain points and simplify control challenges because their cause-and-effect relationships are clear, making them “low-hanging fruit” for resolution.
Operational analysis concretizes process demands, and control experts no longer need to understand the entire process flow, eliminating confusion from lack of project experience. By breaking a large problem into smaller, variable-specific issues, operational analysis reduces the challenge from scaling a mountain to climbing a hill, fostering consensus. The approach is to discover problems through operational analysis and solve them in the simplest, most optimal way. This methodology is versatile and applicable across different processes and industries. While anyone can learn this methodology, problem-solving outcomes vary depending on the individual. Initially, engineers may feel overwhelmed by the analysis, let alone designing solutions. Different experts may propose significantly different solutions. Although solving problems can be challenging, identifying opportunities for plant automation and knowledge automation through operational analysis is achievable.
Operational analysis is not about evaluating operators but about supporting them and enabling knowledge automation. If operations require evaluation, it should target engineers: Why hasn’t knowledge automation been achieved? Why is manual intervention still necessary? The purpose of operational analysis is to solve problems, not merely explain them. Some operations are easy to reduce, while others are more difficult. An engineering mindset is crucial—striving to solve problems under constrained conditions. Large-scale overhauls or aggressive approaches often lead to poor cost-benefit ratios, deterring decision-makers and deviating from engineering principles. Creative problem-solving is the challenge. In soccer, some teams win world championships, others can’t advance beyond their region, and some are relegated to commentary. The Lambda tuning method has largely resolved PID tuning challenges. Similarly, with operational analysis, identifying problems isn’t difficult once the mindset is adopted.
I encourage everyone to analyze operational events. If the data can be exported to Excel, pivot tables can easily organize and sort it. Each company’s situation varies—some may not know how to start with operational analysis to identify issues, while others turn analysis into mere explanations. Identifying problems isn’t hard, but analyzing and solving them is more complex; otherwise, these issues wouldn’t persist in plants for years. Problem-solving is an art, requiring case-specific analysis. For example, in drum level control, cascade control can decouple flow if a flow measurement is available, or a variable frequency pump can decouple via constant pressure supply. In a sulfuric acid plant, incinerator temperature control is critical—stabilizing boundary conditions and minimizing disturbances significantly improves temperature stability, a solution that might seem counterintuitive (“stable without control”). For tail gas scrubber SO2 control, neither manual nor APC control achieved stability, but leveraging existing DCS functionality stabilized it. Once stable, manual intervention dropped to zero, allowing process parameter optimization.
The simplest and most optimal solutions should be standardized, minimizing custom code, modules, and resource usage. They should avoid linear thinking, understand design intent and operational logic, leverage existing conditions, and be easily explainable to operators. An optimal solution is one that looks straightforward but is challenging to design. If a solution occasionally requires operator intervention, it’s not truly optimal.
Methodology Summary:
- Discover Problems: Through operational and alarm analysis.
- Analyze Problems: Understand intent and operational logic.
- Solve Problems: In the simplest, most optimal way.