Emergent Energy
    Back to case studies
    Manufacturing — Injection Molding
    Meridian Plastics·February 27, 2026

    How Energy Data and Process Sensors Transformed Preventative Maintenance at Meridian Plastics

    62% reduction in unplanned downtime, 41% lower maintenance costs, 14.7% energy savings, and 46-day payback period

    Share:
    Introduction Manufacturing environments operate on razor-thin margins where every minute of unplanned downtime translates directly to lost revenue. For Meridian Plastics, a mid-size injection molding facility in Cleveland, Ohio running three production lines across 187,000 square feet, unplanned equipment failures were costing the company an estimated $2.3 million annually. Traditional time-based maintenance schedules were either too conservative—replacing components prematurely—or too aggressive, allowing critical failures to cascade through interconnected systems. The challenge was clear: Meridian needed a way to predict equipment degradation before it caused failures, without the capital expenditure of a full-scale IIoT platform. The solution came from an unexpected convergence—combining circuit-level energy monitoring with strategically placed process sensors to create a predictive maintenance framework that paid for itself within four months. The Problem: Reactive Maintenance in a Connected Process Meridian operates 14 injection molding machines ranging from 150 to 1,500 tons of clamping force, supported by ancillary systems including hydraulic power units, material dryers, chillers, and conveyor systems. Before the project, maintenance operated primarily in reactive mode. The facility employed a basic CMMS (Computerized Maintenance Management System) with time-based PM schedules, but these schedules were based on manufacturer recommendations rather than actual equipment condition. The consequences were significant: - Average of 4.2 unplanned shutdowns per month across all lines - Mean time to repair (MTTR) of 6.8 hours per incident - Spare parts inventory carrying costs of $340,000 annually due to uncertainty - Energy waste from degraded equipment estimated at 18-22% above baseline - Quality defects traceable to equipment condition averaging 3.1% of total output Perhaps most critically, when one machine in a production cell failed, it created a cascade effect. A failed chiller would cause mold temperature instability across multiple machines, a degraded hydraulic pump would produce inconsistent shot weights, and a worn barrel heater band would create material flow variations that showed up as surface defects hours later. The Solution: Convergent Data Architecture Rather than deploying a monolithic IIoT platform, Meridian implemented a layered monitoring approach using circuit-level energy meters and targeted process sensors, all feeding into a unified analytics platform. Layer 1: Circuit-Level Energy Monitoring Panoramic Power wireless energy sensors were installed on every circuit feeding production equipment—not just at the panel level, but at the individual machine and subsystem level. This granularity was critical. A panel-level meter would show total consumption for an entire production cell, but it couldn't distinguish between a hydraulic pump drawing 15% more power due to internal wear versus a barrel heater compensating for a failed thermocouple. The deployment included: - 47 circuit-level energy sensors across three production lines - Monitoring of individual motors, heaters, pumps, and auxiliary systems - 1-second sampling intervals for real-time anomaly detection - Historical trending with 15-minute aggregation for long-term analysis Layer 2: Process Sensors Strategic process sensors were added to capture the operational context that energy data alone couldn't provide: - Vibration sensors on critical rotating equipment (hydraulic pumps, material augers, cooling fans) - Temperature sensors on mold cooling water supply and return lines - Pressure transducers on hydraulic systems - Flow meters on cooling water circuits Layer 3: Analytics and Correlation Engine The breakthrough came from correlating energy signatures with process sensor data. A backend analytics engine normalized the data streams and applied pattern recognition algorithms to identify degradation signatures weeks before they would cause failures. Key Findings and Energy Signature Patterns The first three months of data collection revealed several critical patterns that became the foundation for the predictive maintenance program. Pattern 1: Hydraulic Pump Degradation Energy monitoring revealed that hydraulic pump motors showed a characteristic power consumption increase of 8-12% over a 3-6 week period before catastrophic failure. When correlated with vibration data, the pattern became even more predictive. A pump showing both elevated power draw AND increasing vibration amplitude in the 2-5 kHz frequency band had a 94% probability of failure within 14 days. Before this insight, hydraulic pumps were replaced on a fixed 18-month schedule at $4,200 per pump. The data showed that actual pump life varied from 11 to 28 months depending on load profile and duty cycle. Condition-based replacement saved $67,000 annually in avoided premature replacements while eliminating 89% of pump-related unplanned downtime. Pattern 2: Barrel Heater Band Degradation Injection molding barrel heaters operate at 400-700°F and are critical for material consistency. Energy data revealed that heater bands don't simply fail—they degrade progressively. A healthy 5 kW heater band drawing rated current at steady state would begin drawing 5-8% more current as resistance heating elements developed micro-fractures, then suddenly drop to near-zero as the element opened. The critical insight was the "pre-failure surge." In 87% of cases, a heater band showed a distinctive 12-18% current spike pattern 48-72 hours before complete failure. By correlating this with barrel zone temperature sensor data, the system could identify not just which heater was failing, but whether the adjacent zones could compensate, allowing maintenance to be scheduled during the next planned changeover rather than forcing an emergency shutdown. Pattern 3: Chiller Efficiency Degradation Cooling water chillers showed a gradual increase in compressor power consumption that correlated with decreasing cooling capacity. Energy data showed compressor power increasing by 2-3% per month as condenser coils fouled and refrigerant charge decreased. When combined with cooling water flow and temperature data, the system could calculate real-time coefficient of performance (COP) for each chiller. This enabled a shift from calendar-based chiller maintenance (quarterly coil cleaning, annual refrigerant check) to condition-based maintenance triggered by COP degradation thresholds. The result was a 23% reduction in chiller energy consumption and elimination of mold temperature-related quality defects. Pattern 4: Material Dryer Performance Resin dryers are critical for processing hygroscopic materials like nylon and polycarbonate. Energy monitoring revealed that dryer heater cycling patterns contained information about desiccant bed condition. A healthy dryer with fresh desiccant showed regular, predictable regeneration cycles. As desiccant degraded, regeneration cycles became longer and more frequent, with measurable increases in energy consumption. By tracking the ratio of drying energy to regeneration energy over time, the system could predict desiccant replacement needs 2-3 weeks in advance, preventing moisture-related defects that had previously caused $180,000 in annual scrap costs. Quantified Results After 12 Months

    Unplanned Downtime Events Per Month — Before vs. After

    • Before (Baseline)
    • After Implementation

    Annual Maintenance Cost Breakdown ($K)

    • Before ($K)
    • After ($K)

    Energy Savings Breakdown by System

    • Chiller Optimization
    • Hydraulic Efficiency
    • Heater Band Mgmt
    • Other Systems

    OEE Component Trends Over 12 Months (%)

    • Availability %
    • Performance %
    • Quality %
    The combined energy and process sensor monitoring system delivered measurable results across every key performance indicator: Downtime Reduction: Unplanned downtime decreased from 4.2 events per month to 1.6 events—a 62% reduction. Remaining unplanned events were primarily related to systems not yet instrumented (material handling, robotics) and external factors (power quality events, raw material issues). Maintenance Cost Reduction: Total maintenance spending decreased 41%, from $1.84 million to $1.09 million annually. This included a 34% reduction in spare parts inventory (from $340,000 to $224,000) enabled by better failure prediction and lead time management. Energy Savings: Identifying and correcting degraded equipment reduced total facility energy consumption by 14.7%, saving approximately $186,000 annually at prevailing utility rates. The largest contributors were chiller optimization (38% of savings), hydraulic system efficiency (28%), and heater band management (19%). Quality Improvement: Product defect rate decreased from 3.1% to 0.9%, directly attributable to more consistent process conditions maintained by properly functioning equipment. This translated to $412,000 in reduced scrap and rework costs. OEE Improvement: Overall Equipment Effectiveness improved from 71.3% to 84.6%, driven by increased availability (less downtime), improved performance (consistent cycle times with properly maintained equipment), and better quality (fewer defects). Implementation Economics Total project cost including hardware, installation, and analytics platform: $187,000. Annual savings across all categories: $1.47 million. Simple payback period: 46 days. Three-year ROI: 2,257%. The surprisingly rapid payback was driven by the multiplicative effect of improvements. Reducing downtime didn't just save direct production losses—it eliminated overtime costs for catch-up production, reduced expedited shipping fees, improved customer satisfaction scores, and freed maintenance staff to focus on improvement projects rather than emergency repairs. Continuous Improvement Framework Beyond predictive maintenance, the converged data platform enabled a continuous improvement methodology that Meridian calls "Energy-Informed Kaizen." By analyzing energy consumption patterns during production, the team identified optimization opportunities that traditional lean manufacturing approaches had missed. For example, energy data revealed that one production cell consumed 23% more energy during startup than an identical cell, despite producing the same parts. Investigation showed that the operator was following a different startup sequence that resulted in extended hydraulic warm-up cycles. Standardizing the startup procedure across all cells saved $34,000 annually. Similarly, energy monitoring identified that cycle time variations of just 0.3-0.5 seconds—too small to trigger traditional SPC alerts—were associated with 8-12% energy consumption variations. These micro-variations traced back to check valve wear in hydraulic systems, providing yet another early warning indicator for the predictive maintenance program. Lessons Learned Several key lessons emerged from the Meridian deployment: 1. Granularity matters. Panel-level monitoring would have captured only 30-40% of the insights that circuit-level monitoring revealed. The incremental cost of per-circuit monitoring was less than 15% of the total project cost but delivered more than 60% of the value. 2. Context is everything. Energy data alone identified anomalies, but process sensor data provided the context needed to determine urgency and appropriate response. The combination was far more valuable than either data stream independently. 3. Start with the data, not the algorithms. Meridian initially planned to deploy commercial predictive maintenance AI, but found that simple threshold-based rules derived from three months of baseline data captured 80% of the predictive value at a fraction of the complexity. 4. Cultural change is the hardest part. Maintenance technicians initially resisted data-driven recommendations, preferring to rely on experience and intuition. The breakthrough came when the system correctly predicted a hydraulic pump failure that an experienced technician had dismissed as normal. After that, adoption accelerated rapidly. Conclusion The Meridian Plastics case demonstrates that effective predictive maintenance doesn't require massive capital investment in IIoT platforms. By combining granular energy monitoring with targeted process sensors and straightforward analytics, manufacturers can achieve dramatic reductions in unplanned downtime, maintenance costs, energy waste, and quality defects. The key is recognizing that energy consumption is not just a cost to be minimized—it is a rich, real-time signal that reveals the health and performance of every piece of equipment in the facility. When combined with process context, that signal becomes a powerful tool for both predictive maintenance and continuous improvement.

    See similar results at your facility

    Let Emergent Metering show you what circuit-level monitoring and energy intelligence can do for your operation.

    About Emergent Metering Solutions

    Emergent Metering Solutions provides commercial and industrial metering hardware, installation support, and energy analytics services. We specialize in electric meters, water meters, BTU meters, compressed air meters, gas meters, and steam meters with Modbus RTU, BACnet IP, pulse output, and wireless communication options. Our Managed Intelligence services deliver automated reporting, anomaly detection, tenant billing, and AI-powered consumption forecasting. We support compliance with IECC 2021, ASHRAE 90.1-2022, NYC Local Law 97, Boston BERDO 2.0, DC BEPS, California LCFS, and EU CSRD requirements.

    Contact our engineering team for meter selection guidance, system design, and project quotes.

    Explore More Resources

    We use cookies to analyze site traffic and improve your experience. Privacy Policy