Interpretive Practices of PLC Based Automation in Industrial Production Systems under Dynamic Operational Conditions

Figure Ilustration AI

FORMOSA NEWS - Makasasar - Human Touch in Automation: Why Operator Interpretation Dictates Success in Industry 4.0. Industrial automation is often envisioned as a flawless network of independent robotics, but new evidence reveals that technological efficiency still depends heavily on human interpretation. A groundbreaking engineering study published in early 2026 demonstrates that even the most sophisticated smart factories suffer severe performance losses if human operators cannot quickly understand machine data. The research was conducted by automation experts Ferdianto Tangdililing and Stefany Yunita Baralangi from Universitas Atma Jaya Makassar. Published in the Formosa Journal of Computer and Information Science (FJCIS), the study evaluated real-time data flows and human-machine interactions within an active automotive manufacturing plant. The findings reveal a critical vulnerability in modern production lines: when operational conditions change unexpectedly, mismatches between automated system data and human interpretation cause machine downtime to skyrocket by 47%. This research proves that the future of smart manufacturing relies not just on advanced hardware, but on designing intuitive digital systems that reduce cognitive strain on the human workforce.

The Hidden Data Bottleneck in Modern Factories
The global push toward the Industry 4.0 paradigm has forced factories to connect physical shop-floor machinery with digital enterprise networks. At the center of this transition are Programmable Logic Controllers (PLCs), which act as the digital brains executing deterministic control logic for industrial hardware. In modern configurations, these PLCs are integrated with Supervisory Control and Data Acquisition (SCADA) systems, Human-Machine Interfaces (HMIs), and Internet of Things (IoT) sensor networks. While this interconnected infrastructure generates vast amounts of real-time operational data, Universitas Atma Jaya Makassar researchers Ferdianto Tangdililing and Stefany Yunita Baralangi point out that abundant data does not automatically yield superior output. Instead, modern production lines face severe volatility from shifting market demands, frequent equipment transitions, and sudden process anomalies. When disruptions happen, the factory floor becomes an information bottleneck. The core industrial challenge has shifted from generating data to helping human operators interpret that data fast enough to make accurate, time-critical decisions.

Analyzing Robotic Welding Lines in Karawang

To investigate these real-world dynamics, the research team from Universitas Atma Jaya Makassar deployed a mixed-method case study design at a prominent mid-scale automotive manufacturing facility located in the Karawang industrial area of Indonesia. The researchers observed a robotic spot-welding station over a rigorous four-week period, capturing data across three consecutive work shifts. The specific station under review utilized a high-performance Siemens S7-1500 PLC paired with a WinCC-based SCADA system and interactive HMI monitors. The methodology combined quantitative data logs directly extracted from the PLC and SCADA databases such as process cycle times, welding actuator currents, and alarm logs with qualitative insights gathered through field observations and semi-structured interviews with plant engineers and operators. By analyzing how the station functioned under standard operations versus unpredictable shifts, the authors constructed an integrated system evaluation model to measure industrial adaptability.

Key Findings: The Costly Gap Between Machine Logic and Human Response
The empirical data compiled by Ferdianto Tangdililing and Stefany Yunita Baralangi exposed a severe performance divide when the manufacturing system moved from stable conditions into dynamic, variable environments:

• Overall Equipment Effectiveness (OEE) Drop: Under normal routines, the robotic welding station maintained a stable OEE of 88.2% and an availability rate of 94.6%. However, during dynamic model transitions, the OEE fell sharply to 81.5%, and availability degraded to 89.8%.
  
• Cycle Time Delays: The average time required to process a single production unit slowed down by 9.0%, lengthening from a steady 42.3 seconds to an inefficient 46.1 seconds per unit.
  
• Spike in Operational Downtime: The most alarming metric was a 47.0% increase in production downtime, which jumped from 18.5 minutes per shift up to 27.2 minutes per shift during volatile periods.
  

Crucially, the Universitas Atma Jaya Makassar study determined that these losses were not caused by sudden mechanical breakdowns. Instead, 62% of the total recorded downtime was directly tied to human interpretation delays. The researchers isolated three distinct systemic gaps responsible for this lag:

• The Visual Gap: Cluttered, non-intuitive HMI screen designs failed to present critical information clearly, automatically adding 3 to 6 seconds of hesitation to every operator reaction.
  
• The Information Gap: System alarms were non-specific. For example, when a critical welding fault occurred, the monitor displayed a generic "F-214 welding current deviation" code. It failed to identify the root cause, forcing operators to spend valuable minutes manually inspecting pneumatic pressures, actuators, and electrode connections.
  
• The Temporal Gap: The accumulated delay from visual confusion and ambiguous data created a heavy cognitive load, adding an extra 5 to 9 minutes of diagnosis downtime per incident.
  

The study also highlighted a widening skills gap. Senior operators leveraged past experience to diagnose faults in roughly 6.8 seconds with 91.5% accuracy. In contrast, junior operators with less than two years of experience suffered from severe informational overload, taking 12,4 seconds to respond with a much lower diagnostic accuracy of 76.2%.

Real-World Impact and Industrial Implications

The insights generated by the Universitas Atma Jaya Makassar researchers provide an essential roadmap for industrial sectors investing in smart infrastructure. The study confirms that connecting PLCs to a centralized SCADA system yields clear historical advantages cutting data access times by 80.7% (from 30 seconds down to 5.8 seconds) and accelerating overall fault identification by 55.3%. Yet, tech integration alone is insufficient if the human element is ignored. For businesses and industrial policymakers, this means that future capital expenditures must balance machine capabilities with user-centered interface design. To mitigate human error and reduce costly downtime, manufacturing enterprises should prioritize the implementation of adaptive HMIs that filter out non-critical alerts to prevent operator sensory overload. Furthermore, engineering teams should integrate Artificial Intelligence (AI) and machine learning directly into PLC-SCADA architectures to shift from reactive troubleshooting to predictive, context-aware diagnostics.

Author Profiles
Ferdianto Tangdililing, S.T., M.T. is an industrial automation engineer and faculty member at Universitas Atma Jaya Makassar. His primary research focuses on industrial control systems, Programmable Logic Controller (PLC) architecture, and the integration of SCADA infrastructure in manufacturing plants.
Stefany Yunita Baralangi, S.T., M.Eng. is an engineering scholar at Universitas Atma Jaya Makassar. Her expertise lies in Human-Machine Interaction (HMI), digital factory interfaces, operational data analytics, and optimizing cognitive workflows for industrial workforces.

Sumber Penelitian
Ferdianto Tangdililing & Stefany Yunita Baralangi, 2026: Interpretive Practices of PLC Based Automation in Industrial Production Systems under Dynamic Operational Conditions 2026. Formosa Journal of Computer and Information Science (FJCIS). Vol. 5, No. 1, Halaman 63-82
DOI: https://doi.org/10.55927/fjcis.v5i1.16489
URL: https://journal.formosapublisher.org/index.php/fjcis