CPE 613 Lecture 34: Plantwide Control Strategies and Variable Selection

General Topic Scope: This technical guide covers the design and selection of variables in plantwide control systems, specifically detailing: * Degrees of Freedom Analysis. * The selection criteria for Controlled, Manipulated, and Measured variables. * Case study application: Stirred-tank heating systems. * Specific challenges in distillation column holdup control.

Control systems are designed to meet three hierarchical pillars of operation:

1. Safety: * Safety is the highest priority in both process and product design. * Acceptable Risk: This is not a static value but is dependent upon the specific situation being considered.
2. Profitability: * The primary economic goal is to maximize profits. * Sub-objectives for Profitability: * Meeting specific product purity goals. * Meeting production rate goals. * Minimizing utility costs. * Minimizing labor costs. * Maintenance: Operating in a stable manner is required to minimize long-term maintenance costs.
3. Environmental Performance: * Regulatory Compliance: Meeting environmental regulations is mandatory. * Impact Minimization: Ideally, the system should minimize environmental impact beyond simple regulation. * Assessment Criteria: Designers must consider both the total amount of material used/produced and the specific impact (toxicity) of that material. * Energy Efficiency: Minimizing energy use is a direct method of decreasing environmental impact. * Life-cycle Perspective: Engineers must consider the entire life-cycle of the product being created. * Conflicting Objectives: The major engineering challenge arises when safety, profitability, and environmental goals conflict.

Variable Selection: Select controlled, manipulated, and measured variables. It is critical not to overspecify the system.
Set-point Determination: Determine set-point values based on the optimization of the plant to reach the overarching objectives (safety, profit, environment).
Control Strategy and Algorithm Selection: * Choose between feedback, feedforward, or other strategies. * Select the algorithm (e.g., PID or others).
Controller Specification: Specify the numerical settings for the controllers, including: * Gains ( $K_c$ ). * Integral times ( $au_I$ ). * Derivative times ( $au_D$ ).

Constraint: A plant can only control a set number of variables, and controlling each specific variable incurs an expense.
Criteria for Variable Selection: * Self-regulation: Any variable that is not self-regulating must be controlled due to safety concerns. * Structural Integrity: Control variables that can cause equipment failure if not kept within strict bounds (Example: tank pressure). * Quality Metrics: Control variables that relate directly to product quality. * Downstream Compatibility: Control variables that have a major impact on units located downstream.

Degrees of Freedom Analysis: Computing the number of degrees of freedom allows an engineer to determine exactly how many variables can be directly manipulated.
Criteria for Choosing Manipulated Variables: * Efficacy and Speed: Select variables that have a large, rapid, and direct effect on the controlled variables. * Ease of Control: Select variables that are simple to manipulate. * Disturbance Management: Do not select variables that will cause disturbances to be recycled back into the plant. Specific examples to avoid include: * Inlet streams. * Recycle streams.

Determining which variables to measure is complex because concentrations and other values are often determined indirectly rather than directly.

Selection Guidelines: * Accuracy: Choose variables that can be measured with high precision and accuracy. * Sensitivity: Choose variables with sufficient sensitivity, meaning they correlate strongly with the target controlled variable. * Time Lag: Choose variables that can be measured without large time delays. * Sensor Optimization: Time delays can be mitigated by choosing proper sensor locations.

Mathematical Modeling: The system is analyzed by writing an energy balance: * $V \rho C_p \frac{dT}{dt} = w C_p (T_i - T) + Q$
Variable Definitions and Assumptions: * The equation is usually written in terms of steam pressure. * Assumption: A known, linear relationship exists between steam pressure and the steam temperature ( $T_s$ ).
Variable Role Identification: * Controlled Variable: Temperature ( $T$ ). * Manipulated Variable: Steam pressure (which dictates the heat flow $Q$ ). * Disturbance Variables: Inlet flow rate ( $w$ ) and Inlet temperature ( $T_i$ ).

Question: When we set up the control of a distillation column (with a partial condenser and a reboiler), there are two variables which are not self-regulating, and therefore it is very important to control those. What are they?
Answer: The liquid holdup in the column and the liquid holdup in the reboiler.
Detailed Reasoning: These specific variables are not self-regulating. If they are not actively controlled, the column or the reboiler will eventually overfill, leading to process failure.

Engineer Involvement: Chemical engineers must be heavily involved in control system design because they possess the core understanding of safety hazards and processing issues.
System Interactions: In large-scale plantwide systems, the interactions between different variables must be evaluated. Failure to account for interactions can lead to controllers "competing" with one another.
Precision in Planning: Careful selection of controlled, manipulated, and measured variables is the foundation of a successful control strategy.