We implemented a real-time KPI dashboard solution for our production floor using FactoryTalk MES 12.0. Our goal was to give operators immediate visibility into production scheduling performance without them needing to access the main MES interface.
The challenge was pulling live data from the production scheduling module and displaying it on large monitors throughout the plant. We needed metrics like schedule adherence percentage, current vs planned production rates, and upcoming changeover times.
We leveraged the MES API to create a custom integration that queries scheduling data every 30 seconds and pushes updates to our plant floor displays. The solution uses REST API calls to extract KPI calculations and formats them for our visualization layer.
The implementation has significantly improved our response time to scheduling deviations. Operators now see real-time alerts when production falls behind schedule, allowing immediate corrective action. What approaches have others taken for similar plant floor visualization needs?
Good questions. We use a middleware service that authenticates once with the MES API using service account credentials, then maintains the session. The displays pull from this middleware via websockets for push notifications. We tested 15, 30, and 60 second intervals. Thirty seconds gave us the best balance between real-time feel and server load. At 15 seconds we saw unnecessary database hits since scheduling data doesn’t change that frequently. The middleware also caches certain reference data like work center names and product codes to reduce API calls.
This is an excellent use case for real-time MES API integration with plant floor visualization. Let me provide a comprehensive perspective on implementing this type of solution effectively.
For the real-time dashboard component, the 30-second refresh interval is well-chosen for scheduling KPIs. The key is matching refresh rates to data volatility - production counts need faster updates (10-15 seconds) while schedule adherence can refresh slower (30-60 seconds). Consider implementing adaptive refresh rates based on production state. During active runs, increase frequency; during idle periods, reduce it to minimize system load.
Regarding MES API integration, the middleware pattern you’ve implemented is the industry best practice. This architecture provides several advantages: centralized authentication management, request throttling, data transformation, and the ability to aggregate multiple API calls into single responses for the displays. Consider adding API versioning support in your middleware so you can handle MES upgrades without breaking existing displays. Also implement request queuing during high-load periods to prevent overwhelming the MES server.
For plant floor visualization, focus on information hierarchy and actionability. Your three-tier approach (operators see actionable metrics, supervisors see analytical metrics) is correct. Add visual thresholds using color coding - green for on-target, yellow for approaching variance, red for exceeding thresholds. Include trend indicators (arrows showing direction) not just current values. For changeover countdown timers, add notifications at 30, 15, and 5 minutes to give operators adequate preparation time.
Additional recommendations: Implement a configuration UI so production managers can adjust KPI thresholds and display layouts without developer involvement. Add shift comparison views showing current performance versus previous shifts. Consider mobile responsiveness so supervisors can view dashboards on tablets during floor walks. Log all displayed KPI values to a time-series database for historical analysis and to identify patterns in response times to schedule deviations.
The impact you’re seeing on response times validates the approach. Real-time visibility transforms reactive troubleshooting into proactive management. Great implementation of FactoryTalk MES 12.0 capabilities.
We started with three core KPIs that operators could actually influence: current hour production count vs target, schedule adherence percentage for the current shift, and minutes until next scheduled changeover. For supervisors we added a separate view with deeper metrics like overall equipment effectiveness and labor utilization. We use the built-in KPI calculations from the production scheduling module where possible, but had to create custom calculations for our specific adherence formula which factors in approved schedule changes and maintenance windows. The API provides the raw data points and our middleware handles the custom math.
Absolutely, error handling was critical. We implemented a similar pattern with a 5-minute data cache and status indicators on each display showing connection health and last update timestamp. If the API doesn’t respond within 3 seconds, we fall back to cached data and show a yellow warning banner. After 10 minutes of failed updates, it turns red. We also log all API failures to our monitoring system so IT gets alerted if there’s a sustained outage. The exponential backoff was key - started at 5 seconds, maxing out at 2 minutes between retries.
This sounds like a great implementation. We’ve been looking at similar visualization options for our assembly lines. How did you handle the authentication for the API calls? Are the displays pulling data directly or going through a middleware layer? Also curious about your refresh rate choice - 30 seconds seems reasonable but wondering if you tested different intervals.
The middleware approach is smart for reducing authentication overhead. Did you implement any error handling for when the MES API is temporarily unavailable? In our environment we’ve had occasional network hiccups that would cause display failures. We ended up adding a local cache with the last known good values and a visual indicator showing data staleness. Also implemented exponential backoff on retry attempts to avoid hammering the API during outages.
What KPIs did you prioritize for the initial rollout? We’re planning something similar and trying to determine which metrics provide the most value to floor operators versus supervisors. Also interested in whether you built custom calculations or used the built-in KPI engine from the production scheduling module.