Design of a Production Efficiency Evaluation System Using an OEE Approach and IoT-Based Energy Indicators with Node-RED in the FMCG Manufacturing Industry

2.1 Production engineering

Automation and robotics have become the backbone of modern production, driving greater efficiency, precision, and adaptability in industrial processes. In the era of Industry 4.0, the integration of IoT and artificial intelligence has paved the way for the emergence of smart factories. Beyond manufacturing, these technologies are now increasingly applied in sectors such as healthcare, logistics, agriculture, and services, thereby enhancing overall productivity and strengthening industrial competitiveness [1]. The advancement of Industry 4.0, particularly through the integration of the IoT and artificial intelligence, has accelerated the transformation of factories into smart factories. These facilities can make faster decisions, minimize waste, and optimize resource use [9]. Automation in industrial construction has introduced substantial improvements in both manufacturing and assembly stages. Technologies such as the IoT, artificial intelligence, digital twins, and robotics enhance efficiency, safety, and the consistency of production quality. Their implementation also facilitates data interoperability, production scheduling optimization, and traceability throughout the manufacturing process, thereby reducing errors, lowering costs, and accelerating project completion times [7]. To enhance performance while simultaneously reducing emissions, it is essential to optimize processes, advance electrification, adopt alternative fuels, and foster innovations in computing and automation engineering. Within the framework of Industry 4.0, digitalization has proven effective in improving energy efficiency [10]. This study emphasizes the role of automation and IoT integration in enhancing production efficiency. The research specifically focuses on energy efficiency within the Industry 4.0 framework. To achieve this, IoT utilization indicators, particularly those from Node-RED, will be used to monitor and optimize energy consumption. This focus is based on the proven potential of digitalization, automation, and data integration technologies to accelerate processes, minimize waste, and reduce production costs.

2.2 Overall Equipment Effectiveness

OEE is recognized as a key indicator within TPM, serving as a tool to identify areas for improvement through the evaluation of three essential dimensions: availability, performance, and quality [2]. In addition, OEE can also be defined as the ratio of actual productive time to the total planned production time (PPT), providing a clear measure of how effectively equipment is utilized [11]. OEE serves as a measure of equipment efficiency, helping to identify sources of loss and guiding improvements within manufacturing systems [3]. Energy consumption analysis can also be applied to estimate OEE values [12]. By leveraging OEE, companies can uncover inefficiencies and design well-structured strategies to enhance productivity [13]. The implementation of OEE provides a comprehensive overview of equipment conditions, enabling companies to prioritize improvements and enhance operational performance [14]. By applying OEE, organizations can pinpoint sources of inefficiency, monitor equipment performance, and formulate targeted improvement plans aimed at increasing overall effectiveness [15]. This study focuses on two indicators: Availability and performance. The quality indicator is excluded because the research examines transfers of existing materials between plants, during which no material is added, and no processes alter the material's quality.

2.3 Automation

Automation in manufacturing aims to reduce repetitive manual work and support rapid adaptation during system changes. In this context, integrating HMI with PLC enables centralized operation and real-time monitoring of equipment [16]. Integrated process automation connects field devices, PLC controllers, and SCADA to provide centralized access to control functions and operational data. At the same time, SCADA supports real-time monitoring through mimics, trends, and alarms [17]. PLCs are widely used as the core platform for industrial automation, executing control functions at the control layer of the automation pyramid and interfacing directly with field devices.  At the same time, PLCs connect to supervisory systems such as HMI and SCADA, enabling centralized monitoring of process and equipment status as well as operator interaction at the supervisory level [18]. The control system for handling and distributing solid raw materials is based on 1994-era technology. It is run by an obsolete PLC and provides no capability to record or trace resource usage. To address these limitations, the PLC is being migrated to an Allen-Bradley platform.

2.4 Internet of Things

In the context of Industry 4.0 and the Industrial Internet of Things (IIoT), a functional architecture is required to coordinate and integrate both hardware and software, ensuring that the latest digital technologies can be fully leveraged in the industrial sector [6]. IoT plays a central role by connecting machines, sensors, and control systems to enable smarter production processes. Through such connectivity, manufacturing industries can perform real-time monitoring, implement predictive maintenance, and make data-driven decisions [19]. The application of IoT in industrial energy management also allows continuous monitoring and regulation of energy consumption, while providing data-based insights to optimize efficiency and reduce waste [4]. In the case of raw material transfer within plants, however, IoT has not yet been implemented since the existing legacy control systems are still under development and lack internet connectivity. Once adopted, IoT will enable the integration of hardware and software components and the transmission of sensor data, which in turn will support the acquisition of OEE-related information.

2.5 Node-RED

Node-RED functions as a flow-based programming platform for integrating hardware devices, communication protocols such as Modbus and MQTT, and cloud services.[4] It enables real-time data integration and process automation, providing seamless connectivity between industrial systems and cloud-based platforms. When combined with IoT, Node-RED within Smart Factory architectures also creates opportunities for predictive analytics [5]. In this context, Node-RED will serve as a bridge for the new system, allowing sensor data to be transmitted directly into the platform. This setup enables automatic collection of sensor readings, operating time, and energy usage—without human intervention—resulting in OEE values that are more accurate, consistent, and accessible at any time.

2.6 Energy efficiency and sustainability

Energy efficiency has become a key component of Industry 4.0. By integrating IoT technologies, energy usage can be monitored and controlled in real time, with sensor data processed through intelligent analytics and IoT-based monitoring systems. This allows industries to reduce energy consumption while maintaining production performance. One widely adopted approach involves energy monitoring using an open architecture integrated with OEE systems, enabling the detection of inefficiencies and the activation of automated energy-saving measures [20]. Improving energy efficiency has also emerged as a major priority in the manufacturing sector. With growing concerns over environmental sustainability, optimizing energy consumption not only reduces operational costs but also mitigates the ecological impact of production processes. By employing advanced energy monitoring technologies, industries can supervise energy usage in real time, allowing corrective actions and saving strategies to be implemented quickly to keep systems performing at their best [21]. Currently, the raw material transfer plant lacks an interlock system to detect and respond when the machine is idle, resulting in considerable energy waste. To address this issue, this study proposes installing a pressure transmitter sensor that serves as an interlock, automatically shutting down the raw material transfer system during idle periods.

Electricity demand in our plant is met through the Java–Madura–Bali (JAMALI) interconnected power system rather than by a single local generating facility. National power-sector planning indicates that supply to the DKI Jakarta load center is delivered via 500 kV extra-high-voltage and 150/70 kV high-voltage transmission corridors. Accordingly, even plants located close to the load area inject their output into the transmission network, where it becomes part of the system-wide power flows, before being routed through primary substations and stepped down for distribution to end users. Building on this grid-based arrangement, the greenhouse-gas (GHG) emissions associated with electricity consumption were estimated using the official JAMALI grid emission factors published by the Directorate General of Electricity under the Ministry of Energy and Mineral Resources (ESDM). The latest publicly available JAMALI values in that series (reference year 2019) report an Operating Margin (OM) of 0.83 tCO₂/MWh equivalent to 0.83 kg CO₂/kWh) [22], which were adopted in this study due to the absence of a newer official release. In addition, national electricity statistics show that coal remains the dominant primary energy source in PT PLN (Persero)’s generation mix (66.43% in 2024), supporting the interpretation that JAMALI’s grid emission intensity is largely influenced by coal-fired generation.

Carbon dioxide (CO₂) is the primary greenhouse gas emitted from the combustion of liquid natural gas in power generation processes. Coal power plants continue to be the largest contributors to total carbon dioxide emissions in many countries, maintaining a dominant role in global greenhouse gas output over the past several decades [23]. The combustion of coal for electricity generation releases substantial amounts of carbon dioxide into the atmosphere, making it a major contributor to global warming and climate change [24]. On a global scale, carbon dioxide emissions from coal-fired power plants remain a significant driver of global warming, despite the growing efforts of many nations to implement emission reduction policies and shift toward renewable energy alternatives [25]. Carbon dioxide is a direct byproduct of the carbon combustion process in coal, with each ton of coal burned generating over two tons of CO₂ emissions released into the atmosphere [26]. The electricity supplied to our plant is generated from coal-fired power stations, which produce significant carbon emissions and are located in densely populated urban areas. In this context, we aim to improve the efficiency of our electricity consumption not only to reduce operational costs but also to support long-term sustainability objectives.

4.1 System implementation results (IoT–Node-RED OEE & energy monitoring)

The monitoring system was deployed at the solid raw material transfer plant to overcome the main constraint of the legacy control system (installed in 1994), which provided limited digital data logging and prevented near-real-time assessment of equipment efficiency and energy behavior.

Three signals were acquired to characterize operating state and energy demand: (i) a CT measuring motor current as a proxy for electrical load, (ii) a pressure transmitter representing process condition to support idle/no-flow detection, and (iii) a run-status signal to determine machine on/off state and accumulate runtime. The combined sensing scheme enabled discrimination among running, idle, and stopped conditions, which is required for availability and performance assessment and for state-based energy analysis.

All signals were streamed to Node-RED as a middleware layer. Within Node-RED, the signals were timestamped, filtered, and converted into derived variables, including runtime accumulation, downtime duration, estimated power, energy (kWh), and event flags (running/idle/stopped). The computed indicators were published to a FactoryTalk View dashboard for near-real-time visualization, allowing operations and maintenance personnel to monitor availability, performance metrics, and energy trends without manual data entry.

Overall, the implementation delivered (a) continuous runtime and downtime records, (b) real-time computation of availability and performance (quality excluded due to transfer-only scope), and (c) energy indicators distinguishing consumption during productive running versus idle periods. These outputs provided operational transparency not achievable under the legacy arrangement and supported improvement actions.

4.2 Overall Equipment Effectiveness result (Availability and performance)

Because the studied process involves inter-plant transfer of already-produced materials, the quality factor was excluded from the OEE calculation, as the observed system is not a product-forming process with measurable good and defective units. The case study focuses on the blower-based material transfer process, in which the main performance losses arise from operating time, transfer delay, and idle energy consumption. Therefore, the evaluation emphasizes availability and performance as the most relevant indicators for assessing operational efficiency. Including a quality factor with an assumed value of 100% would not add analytical value and could create the misleading impression that product quality was directly measured in this study.

To maintain evaluation validity, the modified OEE used in this study is explicitly defined as an availability–performance-based OEE. This metric is not intended to replace the full standard OEE formulation, but to adapt the OEE concept to a utility or transfer-process context where quality loss is not directly observable. Therefore, the results should be interpreted as measures of equipment utilization and transfer performance rather than of complete manufacturing OEE. For comparison with standard OEE benchmarks, this limitation is acknowledged, and the modified OEE values are compared only in terms of availability and performance improvement, not as a full three-factor OEE including quality.

4.2.1 Availability

Availability is the proportion of planned production time during which the transfer system operates. Planned production time (PPT) is defined as the scheduled operating time within the observation window; operating time (OT) is the total accumulated runtime; and downtime (DT) is the difference between PPT and OT.

During the observation window before using the interlock system (22 November 2025 – 6 December 2025) and the observation window after the interlock system (9 November 2025 – 23 December 2025).

Figure 10. Availability before and after using the interlock

As shown in Figure 10, the availability before the interlock system fluctuated considerably, with several significant decreases during the observation period. The average availability before implementation was 92.10%. After the interlock system was implemented, availability became more stable and generally remained higher, averaging 95.42%.

These results indicate that the interlock system improved the operational readiness of the transfer system by reducing downtime and maintaining more consistent operation during planned production time. Thus, the implementation of the interlock system had a positive effect on system availability.

4.2.2 Performance

Performance represents how effectively the transfer system operates relative to its ideal pace during operating time (OT). To ensure methodological consistency, the Performance component of OEE was computed using a rate-based approach, comparing the ideal transfer rate with the actual average transfer rate over the observation window. Figure 11 description condition performance before and after using the interlock system.

Figure 11. Performance before and after using the interlock

As shown in Figure 11, the performance value before the implementation of the interlock system fluctuated during the observation period, with an average value of 89.12%. Several decreases occurred on specific days, indicating that the actual transfer rate was not consistently aligned with the ideal operating rate. After the interlock system was implemented, the performance value showed a higher, more stable trend, with an average of 93.16%.

The increase in average performance indicates that the interlock system contributed to improving the operating efficiency of the transfer system. This improvement suggests that the system operated closer to its ideal transfer rate after the interlock system was implemented. Therefore, the interlock system had a positive effect on maintaining transfer stability and improving system performance during operating time.

4.2.3 Combined effectiveness without quality the resulting value indicates that improvement leverage is concentrated in (availability/performance), with quality excluded; the OEE indicator before and after using the interlock system is shown in Figure 12.

Figure 12. Overall Equipment Effectiveness (OEE) before and after using interlock

The OEE trend over 15 days shows a consistent improvement after the implementation of the interlock system. The average OEE increased from 82.10% to 88.89%, indicating enhanced equipment effectiveness. Despite minor fluctuations, the post-implementation values remain generally higher and more stable, suggesting that improvements are driven primarily by availability and performance rather than quality.

4.3 Energy efficiency results (Power, kWh, Cost, and CO₂)

4.3.1 Idle detection and interlock impact (pressure-based)

The pressure transmitter signal was used to determine the transfer system's operating state using a simple threshold rule. A 300 mbar cut-off was set to distinguish normal operation from idle conditions. When the pressure drops below 300 mbar and above 10 mbar, the system is considered idle (no-flow). When the pressure is 300 mbar or higher, the system is classified as running under normal transfer operation. This running/idle flag is then used by the interlock logic to trigger an automatic shutdown during idle periods, preventing the equipment from operating when no effective material transfer is occurring.

System triggers shutdown/interlock to prevent continued running without transfer. After enabling this interlock, Figure 13 shows the idle energy consumption before and after its activation.

Figure 13. Energy idle consumption before and after using interlock

Figure 13 compares idle energy consumption (kWh) over 15 days before and after the implementation of the interlock system. The blue line represents energy consumption before the interlock, while the orange line indicates values after implementation.

The graph shows that idle energy consumption is significantly reduced after the interlock system is applied. Before implementation, energy usage fluctuates widely, reaching higher peaks, with an average of approximately 23.33 kWh. After implementation, the consumption becomes lower and more stable, with an average reduced to about 11.28 kWh.

Although daily variations still occur, the post-interlock values consistently remain below the pre-interlock levels. This indicates that the interlock system effectively minimizes unnecessary energy consumption during idle periods by automatically shutting down when no material transfer occurs.

4.3.2 Electricity cost and CO₂ emissions

Using the tariff middle industrial consumption >200 kVA - <30.000 kVA rate of 1.035,78 IDR/kWh and the natural gas emission factor of 0.83 kgCO₂/kWh, after enabling this interlock, the savings were calculated as, cost saving 12.05 kWh × 1.035,78 IDR/kWh = IDR 12,476.00, and CO₂ reduction 12.05 kWh × 0.83 kgCO₂ = 9.99 kgCO₂/days.

These results show that the proposed approach improves efficiency not only in operational terms (availability/performance) but also in sustainability terms by reducing avoidable energy use associated with idle operation and thereby reducing CO₂ emissions. Figure 14 describes CO₂ emissions before and after using the interlock system.

Figure 14. CO₂ emissions before using interlock

Figure 14 presents a time series comparison of daily CO₂ emissions over 15 days, under two conditions: before (blue line) and after (orange line) the implementation of the interlock system. The blue curve shows higher, more fluctuating emission levels, indicating inefficient operation with significant energy consumption during idle periods. In contrast, the orange curve is consistently lower and more stable, reflecting reduced emissions after the interlock prevents unnecessary operation.

The dashed horizontal lines represent the average emissions for each condition. The mean value before implementation is approximately 19.36 kg CO₂/day, while after implementation it decreases to around 9.36 kg CO₂/day. This clear downward shift demonstrates that the interlock system effectively reduces both variability and overall emissions, confirming its impact on improving energy efficiency and environmental performance.

4.4 Statistical review

A statistical comparison was performed to quantify and validate the operational improvement following implementation of the interlock system. The mean value in Eq. (9) was used to represent the average daily condition of each evaluated parameter during the observation period. Meanwhile, the standard deviation in Eq. (10) was calculated to assess the variation and stability of daily operational performance.

The results indicate that the interlock system improved both the average performance and operational stability. Availability increased from 92.10 ± 3.57% to 95.42 ± 1.45%, indicating not only higher equipment availability but also lower daily fluctuation. Actual Performance increased from 89.12 ± 1.60% to 93.16 ± 1.68%, while modified OEE increased from 82.10 ± 4.10% to 88.89 ± 2.31%. These improvements suggest that the interlock system enhanced process synchronization and reduced unnecessary idle time during operation.

The absolute change, calculated using Eq. (11), further confirms the magnitude of the improvement. The modified OEE increased by 6.79 percentage points, representing the combined effect of improved availability and performance. In addition, idle energy consumption decreased from 23.33 ± 10.46 kWh to 11.28 ± 3.35 kWh, while CO₂ emissions decreased from 19.36 ± 8.68 kg to 9.36 ± 2.78 kg. The reduction in both energy consumption and emissions indicates that the interlock system contributed not only to productivity improvement but also to more energy-efficient and environmentally sustainable operation.

Welch’s independent t-test, expressed in Eqs. (12) and (13), was applied because the before- and after-implementation data represent two different operational periods and may have unequal variances. This approach is more appropriate than the conventional Student’s t-test when operational data show different levels of variation between the two periods. The adjusted degrees of freedom were calculated using the Welch–Satterthwaite equation to ensure a more reliable significance test.

At the significance level α = 0.05, the decision criterion in Eq. (14) was used to assess statistical significance. The results show that Availability improved significantly with p = 0.0036. Actual Performance and modified OEE also showed highly significant improvements (p < 0.001). Furthermore, idle energy consumption and CO₂ emissions were significantly reduced (p = 0.0006). These findings demonstrate that the observed improvements were statistically significant and were not merely descriptive differences between the two operational periods.

Overall, the implementation of the interlock system significantly improved operational effectiveness, reduced idle energy consumption, and lowered CO₂ emissions. Therefore, the proposed interlock system can be considered an effective approach to improving manufacturing process efficiency and supporting energy-saving and emission-reduction initiatives, as shown in Table 3.

Table 3. Statistical comparison of operational performance before and after interlock system implementation

This study was supported by Research and Technology Transfer Office - Bina Nusantara University.

The authors gratefully acknowledge the Directorate General of Electricity, Ministry of Energy and Mineral Resources (ESDM), Indonesia, for providing public access to the data used in this study through the official Gatrik ESDM website. Although newer emission factors were unavailable from an official source during the study, the 2019 value remains a reasonable approximation because changes in the national electricity generation mix generally occur gradually. Nevertheless, this limitation is acknowledged, and the emission results should be interpreted with caution. Future studies are encouraged to use updated official emission factors when newer data become available. The data source was last updated on 02 February 2026 at 9:45 PM (WIB) https://gatrik.esdm.go.id/frontend/download_index/?kode_category=emisi_pl. Data Tables 1 and 2 are available from the Zenodo Repository.