An Open-Source Web-Based Approach to Industrial Supervision and Data Acquisition in the Context of Industry 4.0 ^†

Abstract

This paper addresses the need for accessible and interoperable supervision solutions within the Industry 4.0 paradigm, particularly for small-scale or resource-constrained environments. The proposed system integrates a web-based architecture using opensource technologies to enable real-time industrial monitoring and data acquisition. A hybrid setup was developed, combining a virtual glass manufacturing process in Factory IO with a physical three-phase induction motor controlled by a Modicon M580 PLC. The system architecture includes a local HMI developed in Control Expert and a remote interface built with React and Node.js, both synchronized through a MySQL 8.0 database populated via Python 3.13 using the Modbus TCP/IP protocol. Experimental results demonstrate consistent data synchronization, reliable multi-platform integration, and an average end-to-end latency of 156 ms, validating the feasibility of the approach for IIoTbased applications. The solution demonstrates how general-purpose web technologies can be effectively repurposed for industrial use, offering a cost-effective and scalable alternative to traditional SCADA systems. The proposed architecture is easily replicable, adaptable to various process configurations, and suitable for academic, prototyping, and SME environments.

1. Introduction

Industry 4.0 represents the integration of digital technologies into manufacturing systems to optimize operations, enhance safety, and reduce resource consumption through automation and real-time data exchange [1,2]. At its foundation, this paradigm promotes smart environments where cyber-physical production systems (CPPS) enable autonomous decision-making and adaptive control [3]. While large industries often lead this transition, the participation of small and medium-sized enterprises (SMEs) is crucial due to their adaptability and potential for innovation [4,5].

A key enabler of Industry 4.0 is the ability to monitor and supervise industrial processes in real time through remote connectivity. The convergence of smart sensors, wireless communication, and cloud platforms facilitates predictive maintenance and decentralized decision-making [6,7]. These developments are supported by digital twins, AI-based event detection, and distributed architectures [8,9]. The COVID-19 pandemic further emphasized the need for resilient remote supervision systems based on SCADA and IIoT infrastructures [10].

Modern SCADA platforms are evolving toward open, interoperable IIoT-based solutions [11,12], aligned with trends in smart grids and decentralized control [13]. However, challenges such as cybersecurity risks and high implementation costs continue to be major barriers, particularly for SMEs [14,15,16]. Although open-source technologies offer a promising path, many current solutions lack robustness or compatibility with industrialgrade hardware [17,18,19]. For instance, the proposals in [20,21] demonstrate feasibility but are often limited to academic or small-scale settings.

In this context, the present work introduces an alternative that combines accessibility with industrial applicability. It presents an open-source, web-based architecture for industrial supervision and data acquisition. The system integrates React (web interface), Python (data acquisition), and MySQL (storage) as a low-cost solution to traditional SCADA systems. A virtual glass manufacturing process in Factory IO was integrated with a physical induction motor controlled by a Modicon M580 PLC. Additionally, a local HMI was developed in Control Expert to validate interoperability with existing systems. The architecture was evaluated based on responsiveness, data latency, and reliability in hybrid environments.

The remainder of this paper is organized as follows: Section 2 describes the system architecture, control implementation, and data visualization tools. Section 3 presents performance results and discussion. Section 4 summarizes the conclusions and future work.

2. Methodology

2.1. System Architecture and Overview

The system adopts a hybrid IIoT architecture that couples a Factory IO glass-process simulation—structured into reception, production, and dispatch—with a physical three-phase induction motor (cooling zone) driven by a Modicon M580 PLC programmed in Control Expert. As shown in Figure 1, the architecture comprises five layers—(i) virtual process, (ii) control, (iii) communication and acquisition, (iv) data storage, and (v) visualization—enabling modular integration of simulation, control logic, data acquisition, storage, and monitoring.

The control layer synchronizes virtual and physical components via IEC programming. A Python service acquires PLC variables over Modbus TCP/IP and writes them to a MySQL schema organized by process stage; although Modbus TCP/IP was chosen for compatibility, the design remains adaptable to other protocols. Supervision is provided through two synchronized HMIs: a local Control Expert interface reading PLC memory and a remote React/Node.js interface querying the database, ensuring consistent cross-platform monitoring.

IIoT-Based Communication and Data Flow

The communication layer ensures real-time data exchange between the PLC, database, and HMIs, as illustrated in Figure 2. Process variables are published via Modbus TCP/IP and periodically acquired by the Python script, which inserts the data into the MySQL database. The local HMI reads directly from the PLC, while the remote interface accesses stored data through Node.js REST APIs, enabling platform-independent supervision.

2.2. Process Simulation and Control Integration

The glass manufacturing process is simulated in Factory IO and organized into three stages:

• Reception: Conveyors transport raw materials—silica sand, soda ash, and recycled glass—into storage tanks.
• Production: A virtual furnace and mold station process the materials. A physical three-phase induction motor is integrated into the cooling zone, adding realism and enabling hybrid control evaluation.
• Dispatch: Robotic arms handle the packaging and preparation of finished products for delivery.

Each stage includes sensors and actuators that support the implementation and validation of control logic, as well as seamless interaction with both local and remote HMIs, as shown in Figure 3.

The control layer, implemented in Control Expert using IEC-standard programming, manages sequencing, sensor inputs, and actuator outputs for the entire system. The real motor in the cooling zone is driven via analog signals and monitored through encoder and current sensors, ensuring precise synchronization between the simulated process and the physical component.

2.3. Data Acquisition and Storage

A Python script, developed with pymodbus, reads PLC registers via Modbus TCP/IP, timestamps the values, and stores them in a MySQL database (Figure 4). Batch reading is used to reduce latency and improve performance.

The database is structured into five tables—Reception, Production, Dispatch, Motor Variables, and Remote Control—organized by process stage and optimized for time-series queries (Figure 5).

2.4. Human-Machine Interfaces (HMI)

Two synchronized HMIs enable real-time supervision:

• Local HMI: Developed in Control Expert, it provides direct access to PLC variables with animated views of each process stage and motor operation.
• Remote HMI: Implemented using React and Node.js, it mirrors the local interface and retrieves data from MySQL via REST APIs. Features include status indicators, real-time charts, and historical trends (Figure 6).

2.5. Comparison with Commercial Alternatives

Several commercial and open-source platforms are available for industrial supervision and IIoT integration. A qualitative comparison with representative solutions is presented below:

• Ignition (Inductive Automation): A highly flexible commercial platform supporting OPC UA, Modbus TCP, and MQTT. While it offers full web integration and modular extensibility, licensing starts at approximately $1500, which may be restrictive for small deployments.
• AVEVA Citect SCADA: A robust commercial solution with WYSIWYG graphical tools and support for proprietary industrial protocols. It targets large-scale installations and involves significant licensing costs.
• ThingsBoard: A freemium/commercial platform focused on IoT applications, compatible with MQTT, Modbus TCP, and HTTP. It provides a customizable web interface, although the free version is limited to 30 devices.
• Eclipse NeoSCADA: An open-source Java-based project supporting OPC and Modbus protocols. While flexible, it requires advanced configuration and technical expertise for deployment.
• Karabo: A powerful open-source platform developed for scientific infrastructures. It supports AMQP and P2P TCP protocols but is primarily designed for research environments rather than industrial production lines.
• Proposed solution: This work introduces a modular, open-source architecture combining Python, React, and MySQL. It enables synchronized local and remote supervision using Modbus TCP/IP but is adaptable to any industrial communication protocol. It achieves sub-200 ms latencies and is particularly suited for academic, prototyping, and SME scenarios without the need for commercial licenses.

Compared to commercial and scientific platforms, the proposed architecture offers a cost-effective and adaptable alternative. Its use of general-purpose technologies and protocol-agnostic design aligns well with Industry 4.0 goals of flexibility, interoperability, and democratization of access to digital tools.

3. Results and Discussion

3.1. System Integration and Real-Time Operation

The proposed IIoT architecture integrates simulation, control, data acquisition, and visualization into a cohesive system. Factory IO operates in real time with the M580 PLC via Modbus TCP/IP, coordinating virtual elements and the physical induction motor (Figure 7).

Both interfaces delivered synchronized visualization. The local HMI in Control Expert updated instantly from PLC memory (Figure 8), while the remote HMI—developed in React and connected via Node.js and MySQL—maintained a sub-second delay (Figure 9).

3.2. Performance Evaluation and Synchronization Analysis

System performance was evaluated based on synchronization, interface responsiveness, data acquisition latency, and database efficiency, confirming the feasibility of the proposed architecture for IIoT-enabled supervision.

Tests showed consistent synchronization across the PLC, database, and both HMIs. Events in the virtual process were logged within milliseconds and displayed in real time on local and remote interfaces without desynchronization.

The local HMI provided immediate PLC feedback, while the remote interface showed real-time and historical data with sub-second delay, including motor speed, reference values, and current readings (Figure 10).

Latency measurements between the PLC and remote HMI averaged 156 ms, with delays during extended operation ranging from 250 to 300 ms. The system sustained an update rate of 15–20 variables per second, as shown in Figure 11.

Database writes below 50 ms and reads under 30 ms confirmed support for highfrequency monitoring. With latency values approximating a Gaussian distribution, the system demonstrated robust integration, low latency, and synchronized visualization, validating its applicability in real-time industrial monitoring.

4. Conclusions

This work presented the design and implementation of an open-source, web-based architecture for industrial supervision and data acquisition under the Industry 4.0 framework, integrating virtual and physical components to enable real-time monitoring through synchronized local and remote interfaces.

A glass manufacturing process was modeled in Factory IO and linked to a physical induction motor controlled by a PLC. Control logic and a local HMI were developed in Control Expert, the remote interface in React/Node.js, and data acquisition via Python over Modbus TCP/IP stored in MySQL.

Experimental results showed stable performance with 156 ms average response times and reliable synchronization. Adapting general-purpose web and database tools to industry, the solution provides a cost-effective alternative to conventional SCADA.

Future work includes validation in real industrial settings, use of time-series databases, and integration of machine learning for predictive analytics to enable anomaly detection, equipment degradation estimation, and process optimization. Interface enhancements will also improve usability on mobile and remote devices.

Given the use of open-source components and Modbus TCP/IP, cybersecurity is essential, requiring network segmentation, restricted access, encrypted web traffic, and secure database management. Though aimed at academic and SME contexts, these measures are vital to safeguard IIoT environments.