Process control engineers increasingly work alongside Generative AI tools that write code, derive equations, analyze data, and draft reports. This course teaches the fundamentals of process dynamics and control while training you to use AI to augment your learning and engineering judgment, not replace it. Syntax and coding mechanics matter less than they once did; what matters more is whether you can pose the right problem, judge whether an answer is correct, and take responsibility for the result.

Why Generative AI in this Course

A few years ago, most of the effort in a dynamics and control course went into writing and debugging Python code. AI assistants now produce working simulation and control code in seconds. That shift does not make this course easier; it raises the bar. When code is cheap, the scarce skills are:

• Problem formulation: translating a physical system into balance equations, assumptions, and specifications an AI (or a colleague) can act on.
• Verification: checking units, limits, conservation, stability, and physical plausibility of any result, whether it came from your own derivation or an AI.
• Engineering judgment: deciding what is good enough, what is unsafe, and what the numbers mean for a real process.

Employers expect graduates who can direct these tools productively. This course gives you structured practice: every assignment includes prompts that test your understanding, and every report you turn in must show that you guided the AI with correct visualizations, justifications, and assumptions.

AI as a Learning Partner, Not an Autopilot

Use this simple contract throughout the course:

• Learn with it, don't hide behind it. Asking an AI to explain, quiz you, or critique your work builds understanding. Pasting the assignment and submitting the output does not, and it shows immediately on exams and in project work.
• Verify everything. AI output is a first draft from a confident junior collaborator. Check equations by substituting limiting cases, check code by testing against known solutions, and check claims against the course material.
• Own the result. You are the engineer of record for anything you submit. "The AI said so" is never a justification.
• Disclose use. Reports should state where AI assisted (drafting, coding, plotting) and what you did to validate the result.

Course Roadmap

The course has four major sections. Follow the links to preview each area:

• Dynamic Modeling - derive balance equations, simulate differential equations, linearize, and fit first- and second-order models with dead time to data.
• Sensors, Actuators, and Data Acquisition - measure, transmit, and act on process signals with sensors, valves, and controllers.
• Control Design - design and tune P, PI, and PID feedback controllers, then extend to cascade and feedforward strategies.
• Control Project with Microcontrollers and Electrical Systems - build a working control system with microcontrollers, sensors, and actuators, supported by electrical engineering fundamentals for chemical engineers.

Hands-on labs use the Temperature Control Lab (TCLab), and browser-based simulation and control apps reduce the coding burden so you can focus on the physics and control concepts.

Course TA Skill for Your AI Assistant

The course provides a pre-built AI Teaching Assistant that turns your AI tool into a course-aware TA. It knows the schedule, every assignment and TCLab activity, the simulation apps, the course notation (FOPDT fitting, IMC tuning, PID forms, the TCLab energy balance), and the course AI policy: it tutors, quizzes, and debugs with you - it does not do graded work for you.

Download it with installation instructions for each tool from the TA Skill archive on GitHub:

• Claude: install pdc-course-ta.skill (Settings, Capabilities, Skills), or unzip it into ~/.claude/skills for Claude Code.
• ChatGPT: create a Custom GPT or Project with pdc-course-ta-instructions.md as the instructions and pdc-course-ta-knowledge.md as a knowledge file. For Codex, copy AGENTS.md and the knowledge file into your working folder.
• Google Gemini: create a Gem with the same instructions and knowledge files, or copy GEMINI.md into your working folder for the Gemini CLI.
• Any other tool: paste the instructions and knowledge files as the first message of a conversation.

Install it during the first assignment and use it all semester for concept help, quiz-me practice, TCLab and code debugging, report coaching, and project support. The TA follows the Responsible Use policy below, so expect it to coach rather than solve, and to refuse to invent data for labs you have not run.

How to Write Effective Prompts

A good prompt is a small engineering specification. Include four parts:

1. Context: who you are and what system you are working on ("I am a chemical engineering student modeling a stirred-tank heater...").
2. Task: exactly what you want ("derive the transient energy balance and state the assumptions").
3. Constraints: scope, level, format ("use SI units, keep it at the level of a senior ChE course, no code yet").
4. Verification: how the answer will be checked ("then give me 3 questions to test whether I understood the derivation").

The Machine Learning for Engineers course develops these ideas further with structured prompt builders for workplans, coding, visualization, and reports. Use them as templates when you reach the course project.

Five Prompt Patterns for Learning

These patterns appear throughout the course assignments. Practice them now:

• Tutor: "Teach me {topic} step by step. After each step, ask me a question and wait for my answer before continuing."
• Quiz Master: "Ask me 5 conceptual questions about {topic}, one at a time. Tell me if I am right and why. Keep score and end with a list of my misconceptions."
• Flawed Explanation (Critic): "Give me an explanation of {topic} that contains one subtle conceptual error. I will try to find it. Then reveal the error and explain."
• Analogy Builder: "Explain {topic} with an everyday analogy (shower temperature, cruise control, filling a sink), then tell me where the analogy breaks down."
• Judgment Check: "Here is my answer/model/plot: {paste}. Critique it as a senior process control engineer would: check units, assumptions, limiting cases, and physical plausibility."

Verification and Engineering Judgment

Every equation, simulation, and controller in this course can be checked without an authority figure:

• Units: every term in a balance equation must have the same units.
• Limiting cases: set time to zero or infinity, set inputs to zero, and confirm the response makes physical sense (e.g., a step response $$\Delta y = K_p \Delta u$$ at steady state for a first-order system).
• Conservation: mass and energy must balance at steady state.
• Known solutions: compare numerical simulation to analytic solutions for simple cases.
• Reality: TCLab data is the final referee - a model or controller that fails on hardware is wrong, no matter how confident the AI was.

Responsible Use

Follow the course academic integrity policy: AI use is encouraged for learning, coding assistance, visualization, and report drafting on assignments and the project, with disclosure. AI use is not permitted on quizzes taken for a grade or on exams. Do not share personal data or unpublished research data with AI tools. When in doubt, ask the instructor.