IoT on the Factory Floor: Connecting Legacy Manufacturing Equipment

What's inside

The full contents of IoT on the Factory Floor: Connecting Legacy Manufacturing Equipment, slide by slide. Read it here, or use the viewer above for the designed version.

1. 02

   The Brownfield Reality

   • Most active production equipment predates IP networking and was never designed to communicate externally
   • Replacing legacy assets outright is rarely justified when the mechanical equipment still performs reliably
   • Retrofitting means adding sensing and connectivity without altering certified control logic or voiding warranties
   • Asset age, documentation gaps, and inconsistent wiring standards make each machine a distinct integration problem
   • The technical challenge is compounded by organizational risk aversion around touching running production lines
2. 03

   Why This Is Hard, Not Just Slow

   • Many controllers use proprietary or serial protocols with no vendor-supported export path
   • Physical access to control cabinets may require OEM sign-off, especially for machines still under service contracts
   • Retrofits must avoid interfering with safety-rated circuits and existing PLC scan cycles
   • Downtime windows for installation are scarce and must be negotiated around production schedules
   • Plant floor conditions — heat, vibration, EMI — constrain what hardware can survive long term
3. 04

   Protocol Translation and Gateway Strategy

   • Legacy PLCs and SCADA systems typically speak Modbus, Profibus, DNP3, or proprietary serial protocols
   • Edge gateways translate these into IP-based protocols such as MQTT or OPC UA for onward transmission
   • Protocol gateways should sit in a passive, read-only configuration wherever possible to avoid write-access risk to control logic
   • OPC UA is emerging as a common target format because of its vendor-neutral data modeling
   • Gateway selection should prioritize protocol coverage for the specific PLC generations already installed on site
4. 05

   Non-Invasive Sensing for Machines Without Native Telemetry

   • Vibration sensors (accelerometers) can be clamped externally to detect bearing wear, imbalance, and misalignment
   • Current sensors (clamp-on CTs) infer motor load and duty cycle without opening electrical panels
   • Acoustic sensors can detect ultrasonic leaks, cavitation, and early-stage mechanical faults
   • These methods avoid rewiring the machine's internal control system, reducing warranty and safety risk
   • Non-invasive sensing is generally the fastest path to instrumenting equipment with no accessible data port
5. 06

   Data Historian Integration

   • A historian consolidates time-series data from PLCs, gateways, and sensors into a queryable long-term store
   • Integration typically routes through OPC UA or a middleware layer rather than direct historian-to-PLC connections
   • Tag naming conventions should be standardized early to avoid costly rework as instrumentation scales across lines
   • Historians enable trend analysis and root-cause investigation that point-in-time SCADA screens cannot support
   • Data retention and compression settings should be set deliberately — raw data volume grows quickly across many tags
6. 07

   Measuring OEE Through IoT

   • Overall Equipment Effectiveness combines availability, performance, and quality into a single line-level metric
   • IoT sensing automates data capture that was previously logged manually or not captured at all
   • Automated OEE reduces reporting lag from shift-end to near-real-time visibility
   • Consistent, sensor-derived data reduces disputes over downtime attribution between operations and maintenance
   • Baseline OEE should be established before instrumentation to make improvement claims measurable and credible
7. 08

   Downtime Reduction Use Cases

   • Condition monitoring on rotating equipment can flag developing bearing or motor faults before failure
   • Current-based sensing can detect abnormal load patterns indicating jams, blockages, or tooling wear
   • Illustrative scenario: a stamping line that logs micro-stops via IoT may surface a recurring changeover bottleneck invisible in shift reports — not a verified case study, offered to show the pattern
   • Alerting thresholds should be tuned conservatively at first to build operator trust and avoid alarm fatigue
   • Downtime reduction benefits typically materialize over multiple quarters as models are tuned to specific equipment behavior
8. 09

   Network Segmentation for Shop-Floor Connectivity

   • Operational technology (OT) networks should be logically and physically separated from corporate IT networks
   • A demilitarized zone (DMZ) architecture is a common industry pattern for controlled data flow between OT and IT
   • Sensor and gateway traffic should be isolated on dedicated VLANs with restricted routing to production controllers
   • Firewall rules should default to read-only, outbound-only data flow from the shop floor unless a specific control case requires otherwise
   • Segmentation also limits the blast radius of any single compromised device or gateway
9. 10

   Prioritizing Which Machines to Instrument First

   • Rank candidate machines by downtime cost, criticality to the line, and frequency of unplanned stops
   • Favor machines with existing accessible data ports or simple retrofit paths for early wins
   • Sequence instrumentation to build internal capability before tackling the most complex or highest-risk assets
   • Bottleneck (constraint) machines typically offer the highest return since their downtime caps overall line throughput
   • Avoid spreading initial budget thin across many machines — depth on a few assets validates the approach before scaling
10. 11

    Vendor and Integrator Selection for Brownfield Retrofits

    • Prioritize vendors with demonstrated experience retrofitting the specific PLC and control platforms in use on site
    • Require clarity on data ownership, export formats, and portability to avoid long-term lock-in to a single vendor's platform
    • Evaluate integrators on their protocol translation depth, not just dashboard or analytics presentation
    • Request references for brownfield (not greenfield) deployments specifically, as the skill sets differ materially
    • Contract terms should include defined support response times for production-floor connectivity issues
11. 12

    Building the Business Case

    • Frame the investment around specific, named downtime or quality costs rather than generic connectivity value
    • Industry-reported ranges suggest unplanned downtime is a material cost driver in discrete and process manufacturing, though the exact figure varies widely by industry and should be validated against internal maintenance records
    • Pilot-scale ROI should be measured against the established OEE baseline, not assumed improvement percentages
    • Include total cost of ownership — sensors, gateways, network upgrades, historian licensing, and integration labor
    • Present the case in phases so funding can be reassessed based on pilot results before scaling
12. 13

    Governance and Change Management

    • Establish clear ownership between OT/maintenance, IT/security, and operations for the instrumented data pipeline
    • Define who has write access, if any, versus who has read-only visibility into shop-floor data
    • Train maintenance and operations staff on interpreting sensor-derived alerts before rollout, not after
    • Document the retrofit approach per machine class to make future instrumentation repeatable rather than bespoke
    • Align incident response procedures to include OT network events, not just IT security events
13. 14

    Phased Line-by-Line Rollout Plan

    • Phase 1: Instrument one bottleneck or high-downtime-cost machine per line as a proof of concept
    • Phase 2: Validate data quality, historian integration, and OEE baseline before expanding sensor coverage
    • Phase 3: Extend to remaining critical machines on the pilot line, refining the gateway and network pattern
    • Phase 4: Replicate the validated pattern across additional lines, reusing tag standards and integrator relationships
    • Each phase should close with a documented lessons-learned review before capital is committed to the next
14. 15

    Next Steps and the Ask

    • Approve a bounded pilot scope: one line, a small number of prioritized machines, defined success metrics
    • Assign an OT/IT governance lead to own data ownership and network segmentation decisions before installation begins
    • Commission a vendor and integrator shortlist evaluation focused on brownfield retrofit experience
    • Set a review checkpoint at the end of the pilot phase to decide on scaling investment
    • Timeline and budget for the pilot phase to be finalized in a follow-up working session with plant engineering leads