[Plant Simulation Study #20] – Let’s Operate Virtual Factory by ChatGPT Plant Simulation

🧪 [Plant Simulation Study #20] – Let’s Operate Virtual Factory by ChatGPT Plant Simulation

🎬 Overview

In this study, I demonstrate how to connect Siemens Plant Simulation with OpenAI’s ChatGPT API, enabling an AI to make real-time decisions and control the flow of a virtual factory.

Using Python as a bridge, the simulation continuously sends live data to ChatGPT, receives its response, and dynamically changes simulation variables such as conveyor directions.
This structure shows how AI-powered digital twins can simulate autonomous decision-making systems in a smart factory environment.

 

YouTube : https://youtu.be/mDQOXU9y6mk

 

---

💻 Integration Architecture & Key Components

ComponentDescription

Tech Stack	Plant Simulation + Python + OpenAI ChatGPT API
Main Goal	AI-driven exit strategy logic for production flow
Communication	Data sent to ChatGPT every 5 seconds → Response received → Variables applied
Real-Time Loop	Python script executes on interval trigger

🔧 Python Core Logic (Simplified)

 

import openai
import pandas as pd

obj = root.EXIT_Seq_SET.Value
set_A = root.Con1_nummu.Value
set_B = root.Con2_nummu.Value

# OpenAI API 키 설정
api_key = "PUT API HERE"  # 여기에 실제 OpenAI API 키를 입력하세요.
openai.api_key = api_key

# ChatGPT API 호출 함수
def chat_with_gpt(prompt, model="gpt-4", temperature=0):
    response = openai.ChatCompletion.create(
        model=model,
        messages=[
            {"role": "system", "content": "You are a decision-making assistant."},
            {"role": "user", "content": prompt}
        ],
        temperature=temperature
    )
    return response.choices[0].message['content'].strip()

# 프롬프트 설정
user_input = f"""
Con1 위 제품 개수 :
{set_A}
Con2 위 제품 개수 :
{set_B}

0. 아래 조건들을 확인하고 Con1 과 Con2 중 한 곳을 출력해라.
1. if {set_A} > {set_B} , Con1 출력
2. if {set_A} < {set_B} , Con2 출력
2. if {set_A} = {set_B} , Con2 or Con1 50% 확률로 랜덤 출력
5. 설명 없이 정확히 둘중 하나의 이름만 적어주세요.
"""

# API 호출 및 응답을 obj 변수에 저장
try:
    reply = chat_with_gpt(user_input)
    
    print("🔹 ChatGPT Answer (원본):\n", reply)

    # 응답을 obj 변수에 직접 저장
    obj = reply  # 기대한 단일 값인 A1, A2, A3, A4, A5 중 하나를 저장
    print("🔹 done:", obj)

except Exception as e:
    print("API 호출 중 오류 발생:", str(e))

root.EXIT_Seq_SET.value = obj

• The prompt includes context and current simulation data.
• The AI’s response is parsed and applied to simulation variables (e.g., Con1, Con2) to alter the model’s flow.

---

⚙️ Simulation Flow

1. Python executes every 5 seconds
   (Run Python – 5 Sec Interval
   

   

   
   
   
2. ChatGPT processes the prompt
   • Response appears in the log window
   • Label: “Answered by ChatGPT”
3. Simulation reacts dynamically
   • Conveyor logic and decision paths are changed
   • “Line Operated by ChatGPT” appears in the model view

---

📈 Use Cases & Possibilities

✅ AI-in-the-loop simulation for smart factory development
✅ Educational content for AI-powered Digital Twin modeling
✅ Prototype for decision support systems using LLMs (Large Language Models)
✅ Expandable to KPI-based optimization, multi-agent decision-making, or RAG-enhanced scenarios

---

🧠 Final Thoughts

This demo shows a practical example of autonomous factory control using ChatGPT inside a discrete-event simulation.
No longer is simulation only about modeling logic — it’s about collaborating with AI to make decisions and operate the factory in real time.