Non-Invasive Cloud-Native Retrofitting of Production Monitoring via Authorized Web APIs: Field Evidence from 50 PLC-Connected Machines

Author (s)

Md. Baki Billah Ripon, Md. Abu Hasib, & Nadia Islam Noyshi

Abstract

The Brownfield manufacturers may require a dependable production data without altering the current programmable logic controllers (PLCs) or supervisory systems. It is a non-invasive field study on a single site that is conducted using a non-invasive monitoring overlay, which is based on an authorized manufacturing web API to investigate whether counter-derived production indicators can be trusted when there is a communication break. Machine records were acquired by Google Apps Script and state persistence, history and visualization were supported by Google Sheets. The effective telemetry window consisted of 36,600 observations of 50 machines connected to PLC during 732 refresh cycles (12.23 h). The successful cycles all yielded full 50-machine panels, temporal refresh coverage was 99.59, and the average API response time was 1,015.98 ms (95th percentile: 1,528.6 ms). The main result was a data-quality failure, not a timing constraint: there were 5 cycles of fleet-wide communication-loss followed by 245 cycles of restoration transitions that accounted for 99.975% of the unfiltered positive counter-difference total. A communication-sensitive classification was thus used to differentiate between OFFLINE and immediate RECOVERY observations and assessable RUN and STOP comparisons. With 500 offline/recovery observations excluded, 36,100 observations were assessable, and the corrected production-active fraction was 44.55% (machine-cluster bootstrap 95% confidence interval: 35.91%-53.20%). Another 166.58-h execution-log window had 9,943 failed attempts at acquisition, with 9,941 being HTTP 500 responses. These results determine four decision-relevant requirements: rule out recovery artifacts, condition production states based on continuity of communication, differentiate PLC connectivity and production activity, and show collector, source and machine communication health individually. The evidence confirms that minute-scale reporting is feasible in this specific deployment, but not that physical machine states, or that the results can be generalized to other locations, or that there are economic returns.

Key words: Brownfield manufacturing, Industrial data quality, Authorized web APIs, Communication-aware state classification, Production decision support, Production monitoring, Smart retrofitting, Operational visibility.

Download PDF

Cite This Article:

Ripon, M. B. B, Hasib, M. A., & Noyshi, N. I. (2026). Non-Invasive Cloud-Native Retrofitting of Production Monitoring via Authorized Web APIs: Field Evidence from 50 PLC-Connected Machines, Journal of Scientific Reports, 15(1), 236-257.  https://doi.org/10.58970/JSR.1249

About Author (s)

Md. Baki Billah Ripon (Corresponding author), Graduate, Industrial & Production Engineering, Rajshahi University of Engineering & Technology, Rajshahi, Bangladesh. ORCID: https://orcid.org/0009-0003-3544-4015

Md. Abu Hasib, Graduate, Industrial and Production Engineering, National Institute of Textile Engineering and Research, Dhaka, Bangladesh.

Nadia Islam Noyshi, Department of Agriculture, Bangladesh Agricultural University, Mymensingh, Bangladesh.

Download PDF

DOI: https://doi.org/10.58970/JSR.1249

Summarize this article with: