Control systems
& modernization
Siemens PCS 7, CEMAT, WinCC Classic, and legacy PLC systems. Requirements, architecture, migration planning, licensing, and plant-team coordination.
INDUSTRIAL AUTOMATION · CEMENT · SOFTWARE
I work at the meeting point of plant engineering and software—helping engineers solve real problems simply, reliably, and cost-effectively.
Navi Mumbai, India
01 / ABOUT ME
My work starts with the engineer who has to keep the plant running.
My background is in industrial automation, DCS and SCADA modernization, and hands-on collaboration with plant engineers. Cement manufacturing is a central part of that work—from crusher and raw material preparation through kiln, grinding, and packing, alongside power and waste heat recovery systems.
I’m equally interested in the everyday gaps: an awkward data export, a difficult OPC connection, a backup nobody has verified, or a report that takes too much manual effort. Those problems are the starting point for the software tools I’m building.
My direction is practical: make existing systems easier to understand and maintain, then use reliable plant data to support better decisions.
02 / WHERE I WORK
Engineering choices have to make sense in a running plant, with real constraints and real consequences.
Siemens PCS 7, CEMAT, WinCC Classic, and legacy PLC systems. Requirements, architecture, migration planning, licensing, and plant-team coordination.
OPC DA and UA, historians, SQL, CSV, and reporting. Turning available signals into data that engineers can inspect, trust, and use.
My industrial AI work focuses on cement process applications: preheater anomaly detection, kiln and mill optimization, soft sensors, and operator advisory tools.
03 / SELECTED PROJECTS
These are the kinds of projects I work on—from focused utilities to larger ideas still taking shape. I welcome conversations with engineers and developers who see a useful next step.
What is in this automation project? What changed? Which physical module does a signal belong to? These tools turn project files and exports into information an engineer can use.
A plant is preparing for an upgrade. The engineer needs to connect program references, I/O addresses and installed modules, then identify which changes or lifecycle concerns deserve attention.
Broader project-format coverage, clearer confidence and exception reporting, and worked examples that help engineers plan maintenance or migration. Plant engineers can contribute real questions; developers can help strengthen parsers and traceability.
A lean industrial data logger for collecting OPC UA and OPC DA values, preserving quality and timestamps, and keeping data available through connection and storage interruptions.
A useful trend needs more than a tag value. It needs the right timestamp, trustworthy quality, and a visible explanation when a connection or storage path fails.
Resource-efficient acquisition, easier diagnostics, array handling and controlled data exchange. Especially useful conversations involve real logging workloads and the limits of the installed plant PC.
Modern data applications still have to connect to older plant systems. This work explores practical bridges between OPC Classic and OPC UA, with attention to quality, security and diagnostics.
An older system exposes OPC Classic data, while a new application expects OPC UA. The bridge has to preserve the meaning of the data and make connection failures understandable.
Vendor interoperability, legacy-machine collectors, clearer certificate and DCOM diagnostics, and the proposed DA/UA mapping workflows. Engineers and developers with difficult connectivity cases would bring valuable test scenarios.
Testing an integration should not require a running plant. Simulation makes it possible to exercise trends, alarms, historian behavior and failure recovery before deployment.
A dashboard works with smoothly changing values. What happens when a sensor reports bad quality, a signal freezes, a counter resets or communication drops?
More process scenarios, reusable acceptance exercises and the planned OPC DA simulation path. Process specialists can contribute realistic behavior; integration developers can contribute failure cases their applications must handle.
Make archive data easier to extract, and make external lab values or predictions visible where an operator already works.
An engineer needs archive data for analysis, or an operator needs to compare a laboratory result with a prediction. Repeated manual exports and disconnected displays make both jobs harder.
Simpler export workflows, more target-system trials, and the planned native CSV viewer. A useful starting point is an anonymized data sample and the operator or reporting task it should support.
Turn a long hardening checklist into a repeatable configuration review for PCS 7 clients, servers and engineering stations.
Collected plant data becomes useful when people can see its context, follow a trend, review an alarm or produce a dependable daily log sheet.
A daily report has values, but can the reader tell whether the plant stopped or the recording stopped? The calculation and its data coverage need to remain understandable.
Section-specific report templates, clear KPI definitions, better report design and practical operator views. Plant and process engineers can help define the calculations, missing-data rules and layouts worth building.
Different plant situations call for different ways to reach people: SMS, phone notifications, MQTT, WhatsApp or spoken messages in the control room.
A local assistant for a curated cement reference library, connecting answers to the passages that support them. Part of my wider interest in AI that helps engineers make informed decisions.
Finding a promising repository is only the beginning. These companion tools help discover relevant resources, remember why they matter and find reusable material later.
XPert Viewer / XPertCSVViewer
A local Windows tool for reading CSV, text and Markdown. Delimited data can be edited, saved with its format preserved, or converted for sharing.
XPertLinkedinPost
An offline browser tool for turning rough drafts into polished posts, with suggestions and a realistic preview.
WRITING IN PROGRESS
A manuscript project: A Practical Engineer’s Guide to AI, Knowledge, Automation and Common Sense. The draft outline explores human judgment, operational knowledge, the economics of AI and the future control room.
This portfolio includes working tools, prototypes and ideas in development. It reflects the problems I work on and the directions I am exploring.
Working on a similar challenge? Let’s discuss what we could develop together.
04 / HOW I THINK
A good solution should leave the engineer more confident, with less effort needed to keep it working.
Start with what the plant needs. Keep the utility focused and remove features that add effort without adding value.
Verified backups, predictable restart behavior, rollback, and clear diagnostics matter as much as normal operation.
Resource use, older operating systems, existing networks, and the engineer’s time are practical design constraints.
Good timestamps, data quality, process context, and operator understanding are the foundations of useful industrial AI.
05 / LESSONS FROM THE FIELD
I share observations from troubleshooting and modernization on LinkedIn, with an emphasis on lessons engineers can put to work.
LEGACY SYSTEMS / RECOVERY
A cement-loading S5-115U went to STOP. The hardware was healthy; a missing data block caused a TRAF error, and the EPROM differed from the expected program by one block.
The lesson: age alone tells you little. A verified restoration path tells you much more.
TROUBLESHOOTING / FIRST PRINCIPLES
Multiple trips at almost the same time naturally drew suspicion toward the DCS. The eventual finding was a loose neutral connection.
The lesson: follow the evidence across electrical, instrumentation, and control systems.
06 / CONNECT & COLLABORATE
I’m interested in conversations with plant engineers, developers and process specialists about useful tools, practical industrial AI and the next step for these projects. Share the problem, the constraints and what a useful result would look like.