Smart Retrofitter

1. Product Description

1.1. Solution Overview

The AIDEAS Smart Retrofit (AI-SR) is a toolkit designed to modernize outdated machinery, aligning it with Industry 4.0 standards. This solution combines a hardware component with a software element and includes a communication layer that bridges the hardware and software levels.

The hardware component enables "sensorization" of the machine, facilitating data collection and system connectivity. The communication layer ensures the interaction between the hardware and software, allowing the software component to analyze the collected data using AI algorithms.

This integrated approach enables businesses to extract valuable insights from machine behavior, optimizing performance and extending the lifespan of legacy systems.

The main problem that can be solved by using this solution is the identification of deviations from normal behaviour, analysing any repetitive operation the machine (or one of its components) performs periodically. With this approach the user will be able to, for example:

• Recovering an outdated machine that lacks proper sensorization and/or connectivity but is still operational, and upgrading it to a machine that can be monitored and connected, thus reintroducing it into an Industry 4.0 environment.
• Monitor energy consumption.
• Identify non standards behaviour.
• Perform simulative predictions based on new conditions (data manually insert or CSV data).
• Access to the historical information (measured and predicted data) related to the energy consumption.

1.2. Features

This AIDEAS Smart Retrofit key features:

• Hardware and Communication Architecture: A tailored hardware and communication architecture is designed based on a specific AS-IS analysis of each machine. Despite being customized for individual setups, it follows a standardized structure that ensures the data acquisition and information flow. This unified framework not only enables monitoring and analysis but also facilitates direct control of the machine, bridging the gap between legacy systems and modern Industry 4.0 standards.
• Data Acquisition and Visualization: Continuously acquire and display (real-time) machine data, providing instant insights into performance.
• Scenario Simulation and Evaluation: Simulate and evaluate new working conditions through manual data entry or imported CSV files, enabling proactive decision-making.
• Data Archiving and Historical Insights: Save all operational data to a database, allowing users to access, analyze, and visualize historical records for improved long-term performance.
• Integration with the AI-MP: sending processed data and explainable AI information to the AI-MP for advanced analysis and cross-referencing information about the machine.

1.3. Prerequisites

• Technical Specifications

The AI-SR is a full-stack client-server application, with a backend developed in Python (Flask) and a frontend in React, communicating via REST API using Flask-CORS. The backend handles data acquisition, AI-based analysis, and historical storage, while the frontend provides real-time monitoring, simulations, and data visualization. All components are containerized with Docker to ensure easy deployment and cross-platform compatibility. The architecture supports integration with the AI-MP platform.

• Technical Development

This AIDEAS Solution has the following development requirements:

• Development Language: Python.
• Libraries: Numpy, Pandas, Scikit-Learn,Flask, SciPy.
• Container and Orchestration: Docker, Kubernetes.
• User Interface: React, PrimeReact.
• Application Interfaces: RestAPI.

• Hardware Requirements

Hardware - Smart Retrofit:

Component	Hardware and Software Specifications
PLC	Compact CPU module - max of 12 I/O modules
Digital I/O module	Digital inputs: +24VDC/3.7 mA; Input delay: 5 ms; Digital short-circuit proof outputs: +24 V DC/0.5 A.
Analog I/O module	Analog/potentiometer inputs (±10 V DC/16 bits or 0-100 %/16 bits); Analog outputs: ±10V DC/12 bits; output: +10 V DC/5-8.3 mA.
Alternative I/O module	Instead of two modules, it is possible to use a MIX module with a lower number of inputs/outputs, both type analogue and digital.
Industrial PC	EC900 no LTE, Linux; RAM 4GB, FLASH 16 GB
Power Supply	120W, 24V, 5A (DIN RAIL)
Display	Touch-screen/monitor (according to configuration)
Connection - RS485	No external +24 V DC supply required
Router	Wi-Fi Router
Wi-Fi	USB Wi-Fi antenna (or mini USB antenna WiFi)
Software - Algorithm	AI, Machine Learning, Deep Learning …
Software - UI	Intuitive User Interface (UI)

Hardware Resources for the solution: AI-SR can run on any platform that supports Docker containers.

• Software Requirements

• Docker Desktop (Windows, Mac, or Linux)

• Python

• npm (for frontend deployment)

• External dependencies

• MongoDB (preseted): for structured storage of historical data and acquired data readings.

2. Installation

2.1. Environment Preparation

Ensure that all dependencies, including Docker, Python, and npm, are installed. Clone the repository from the official GitLab project and configure the backend and frontend environments as needed.

2.2. Step-by-Step Installation Process

• Local Installation: Requires configuring backend and frontend, installing dependencies, and launching services manually.

• Docker Installation: Uses a docker-compose.yml file to deploy the application.

• Kubernetes Installation: Not fully developped - under development.

AI-SR installation

The application can be installed locally in a and launched following the next steps:

1. Install Python and npm.
2. Install backend dependencies: Open a new terminal in the folder where the "dev.txt" file is located, "./subsystems/backend/requirements", and run the following command:

    pip install -r requirements.txt

3. The component may be run on any platform that support Docker images.
Docker Desktop for Windows
Docker Desktop for Mac
4. Download the zip folder or clone the repository from this link: Gitlab Project.
5. Configure the Docker environment:
1. Open the "docker-compose.yml" file from ".orchestration/docker".
2. If needed, modified the ports of the desired service.
3. Save and close the file.
6. Launch the application.
1. Open a new terminal in the folder "./orchestration/docker" the "docker-compose.yml" file is located and run the following command:

docker-compose up --build

2. The application will be running as soon as the process finishes and the user interface will be accesible in the defined URL, http://localhost:3010, if the IP and port have not been changed. It is recommendable to use Google Chrome.

3. Initial Configuration

3.1. First Steps

• Login

To access the solution, it is necessary to log in by clicking on the section in the upper right corner of the interface. Then entering your credentials.
If these credentials are authorized to access the solution, the panel will appear in a new version.
You are logged in!

• Interface Navigation

The solution is composed by three different cards:

View the variable: provides access to the variable display screen where variables can be monitored and possible non-standard behaviour observed

Stored data: allow to visualize historical behaviour trends

Simulator: allow to test new configurations without the necessity to try them on the machine

Otherwise, the configuration is very simple and intuitive even for a less experienced user.

• View the variable

Accessing the first card View the Variable you can go and click on Start Visualization to start displaying real-time data and their standard behaviour regions. If a variable follows a non-standard trend, the dot marker will change from green to red. On the same screen you can select the button for:

• Stored data

Accessing Stored data it is possible to select a recorded test of a specific day for a specific variable. Selecting all the information on the dropdown sections, the trends are displayed in a graph.

• Simulator

Accessing Simulator it is possible to predict the value or the trend of a specific varible under new conditions. The conditions can be uploaded manually or through a CSV file.

OR

3.2. Main Workflows

• Workflow Description (Step-by-Step)

1. Hardware Installation and Configuration

   • Install the Smart Retrofit (AI-SR) hardware on the target machine.

   • Ensure that sensors are correctly placed and calibrated for accurate data collection.

   • Establish communication between the hardware and the AI-SR software.

2. Software Installation and Initial Setup

   • Install the backend and frontend following the installation steps provided in Section 2.

   • Configure Docker and environment variables (e.g., MongoDB connection string and backend base URL).

   • Deploy the application and ensure connectivity between the hardware, communication layer, and software.

3. Real-Time Monitoring

   • Access the View the Variable section to start real-time monitoring.

   • Observe machine variables and track standard vs. non-standard behavior.

   • If a variable deviates from expected behavior, the dot marker changes from green to red, indicating an anomaly.

4. Historical Data Analysis

   • Navigate to the Stored Data section to access previously recorded machine behavior.

   • Select a specific date and variable to visualize historical trends.

   • Analyze patterns over time to detect anomalies, inefficiencies, or performance variations.

5. Simulation of New Conditions

   • Use the Simulator to test new operating conditions without affecting the actual machine.

   • Enter new parameters manually or upload a CSV file with predefined data.

   • Observe the predicted impact of changes, enabling predictive analysis and proactive decision-making.

6. Integration with AI-MP (Optional)

   • Processed data and AI-driven insights can be sent to the AI-MP platform for deeper analysis.

   • Cross-reference machine data with broader datasets to identify systemic issues and optimization opportunities.

7. Continuous Optimization and Maintenance

   • Use insights from real-time monitoring, historical data, and simulations to optimize machine operation.

   • Enhance monitoring the energy condition of the machine during operations and extend machine lifespan by addressing potential failures early.

---

• Examples or Use Cases

• Performance monitoring: Check the energy consumption of a machine to detect possible waste.

• Detecting non-Sandard behaviour: Detect abnormal vibrations that could indicate an impending failure.

• Simulation of operating conditions: Test new working parameters before applying them to the real machine.

• Extend useful life of the machine: Thanks to AI-SR, obsolete machines can be integrated with an hardware part, avoiding premature disposal. This makes it possible to extend their life cycle and adapt them to Industry 4.0 , making them more efficient and connected to a smart production system.

4. General Queries

4.1. Installation and Configuration Contact

In the “Information” and “Contact Detailes” sections of the interface it is possible to find more informations about AI-SR and the AIDEAS project.

Users can reach support if they need help with installation or configuration via email:

• UNIVPM

• Ilaria Pietrangeli i.pietrangeli@pm.univpm.it

• Filippo E. Ciarapica f.e.ciarapica@staff.univpm.it

• Giovanni Mazzuto g.mazzuto@staff.univpm.it

• ITI

• Diego Silveira Madrid dsilveira@iti.es

• UPV

• Miguel Ángel Mateo Casali mmateo@cigip.upv.es

4.2. Licensing and Support

Users can obtain support through the following channels:

Email Support: Users can contact our support team at the emails listed in 4.1. section for assistance.

Contact AIDEAS: A Box for messages is available for users at https://aideas-project.eu/.

5. Appendices

5.1. Glossary of Terms

AIDEAS : AI Driven Industrial Equipment Product Life Cycle Boosting Agility, Sustainability and Resilience

AI : Artificial Intelligence

AI-SR : AIDEAS Smart Retrofit

SR : Smart Retrofit

5.2. API Documentation

By default, the backend server runs on port 5010 and provides the following API endpoints. These can be accessed from the frontend interface or tested using tools like Postman or Python scripts.

Resource	GET	POST
/save_to_mongodb		Supported
/simulator_pama		Supported
/simulator_pama_22		Supported
/output		Supported
/input	Supported
/output_data/<selectedFile>/<selectedDay>/<id>	Supported
/output_data		Supported
/dataMP	Supported

5.3. API Endpoints Description

/save_to_mongodb

• POST → Saves prediction and real data for multiple variables into MongoDB and the Machine Passport collection.

/simulator_pama

• POST → Simulates new operating conditions through manually entered values. It returns the selected variable prediction as result.

/simulator_pama_22

• POST → Simulates new conditions using a CSV file and a selected variable. Returns prediction results.

/output

• POST → Reads, aggregates and processes the most recent values from MongoDB and returns predictions for real-time variables.

/input

• GET → Retrieves the list of stored variables with available dates and IDs.

/output_data/<selectedFile>/<selectedDay>/<id>

• GET → Fetches stored prediction vs. real-time series for a specific variable, date, and MongoDB document ID.

/output_data

• POST → Receives and validates a set of prediction data in JSON format. Returns the same structure if valid.

/dataMP

• GET → Returns all stored machine prediction data and feature importance information from the Machine Passport.

5.8. Console Commands List

• npm install for frontend dependencies.

• pip install -r requirements.txt for backend dependencies.

• docker-compose up --build for Docker-based deployment.

• python enpoint_sr_pama.py to launch the backend server.

• npm run dev to start the frontend server.