How a Water Utility Group Achieves Cost Reduction and Efficiency Gains through Industrial Park Energy Management
Industrial Park Energy Management Engineering Case: A Closed-Loop Practice from Metering and Optimization to Dual-Carbon Goals
With annual electricity consumption of approximately 132 million kWh and maximum demand of approximately 18.6 MW, a high-end equipment manufacturing industrial park carried out park-level energy management engineering centered on an energy management system (EMS), energy efficiency optimization, energy-saving retrofits, and dual-carbon goals (carbon peaking and carbon neutrality). The project was built around full-area metering, multi-energy-flow monitoring, load forecasting, demand control, and PV-storage coordination, covering key energy-consuming systems such as central HVAC, air compressors, production power, lighting, distributed PV, and energy storage. According to project acceptance data, the park achieved annual electricity savings of approximately 6.12 million kWh, a comprehensive energy-saving rate of approximately 4.6%, annual electricity cost savings of approximately RMB 3.88 million, a decrease in energy consumption per unit output value of approximately 5.2% compared with the baseline, and an annual CO2 emission reduction of approximately 3,680 tonnes, providing a reference engineering sample for similar industrial parks to advance digital energy management and low-carbon transition.
I. Project Background: High-Load Park Facing Dual Constraints of Carbon Goals and Costs
The project is located in an advanced manufacturing cluster in the Yangtze River Delta. The park covers approximately 25.3 hectares, with a total floor area of approximately 460,000 m², and hosts 126 enterprises mainly engaged in precision machinery, electronic components, automation equipment, and testing services. In 2023, the park's annual electricity consumption was approximately 132 million kWh, of which production load accounted for approximately 58%, HVAC and power equipment for approximately 27%, and lighting, office, and other auxiliary energy use for approximately 15%. The park's energy structure was dominated by electricity, with electricity consumption accounting for approximately 92%, natural gas for approximately 7%, and steam for approximately 1%.
With enterprise capacity expansion and increasing summer high-temperature loads, the park's maximum demand increased from 15.8 MW in 2021 to 18.6 MW in 2023, with an annual peak load growth rate of approximately 8.5%. At the same time, the park had built approximately 2.1 MW of rooftop distributed PV and 1.5 MW/3 MWh of electrochemical energy storage, but lacked unified dispatch among PV, storage, chiller plants, air compressors, and production equipment; energy operation was still mainly based on manual experience and local control. Under the constraints of dual-carbon goals, the park management proposed phased targets: reduce carbon emission intensity per unit output value by 18% by 2025 compared with 2020, achieve carbon peaking before 2030, and gradually establish a measurable, traceable, and optimizable energy management system.
Before the project was launched, the park had three prominent problems:
- Inadequate metering infrastructure: The park originally had approximately 280 metering points, mainly covering main incoming lines and some buildings, with a sub-metering coverage rate of approximately 61%; key systems such as chiller plants, air compressors, boiler rooms, and workshop power lacked independent metering.
- Lagging data utilization: Data on water, electricity, gas, cooling, and heating were scattered across manual meter reading, equipment built-in monitoring, and third-party platforms; monthly energy reports were usually delayed by 5 to 7 days, making it difficult to support energy efficiency optimization and anomaly diagnosis.
- Energy-saving potential was not systematically identified: There was a lack of coordination among production scheduling, HVAC operation, air compressor pressure settings, and peak-valley electricity prices, causing some equipment to maintain high load during non-production periods; demand charges, basic charges, and peak charges all had room for optimization.
II. Technical Challenges: Multi-Energy Flows, Heterogeneous Equipment, and Data Governance
Park energy management is not simply installing electricity meters and deploying software, but an integration project involving multiple energy flow systems including electricity, gas, cooling, heating, and water. During preliminary diagnosis, the project team found that the park's equipment came from diverse sources and communication protocols were inconsistent. Smart meters mainly used DL/T 645 and Modbus RTU protocols, chiller plant control systems used BACnet, air compressor systems used Modbus TCP, some PLC devices supported OPC UA, and PV inverters and storage PCS provided vendor-specific proprietary protocols. Without a unified data access standard, the EMS would find it difficult to form a complete energy view.
Data governance was another challenge. Some meters in the park had been installed for more than 8 years, with issues such as incorrect multiplier settings, missing samples, and clock desynchronization. Initial verification found abnormal multipliers at 17 metering points, time deviations exceeding 15 minutes at 9 acquisition points, and 23 water and gas meters not included in unified data acquisition. If energy consumption were calculated directly from raw data, errors could reach 3% to 5%, affecting energy savings accounting and carbon emission accounting. The project team therefore established an implementation path of "verify first, then acquire, then optimize," conducting accuracy tests, multiplier reviews, and clock synchronization for key metering points.
The uncertainty of production loads also increased system control difficulty. Enterprises in the park did not operate continuously, and order fluctuations caused multiple electricity peaks within a day in workshops. An electronic component enterprise concentrated test equipment operation during the daytime, while a precision machinery enterprise arranged heat treatment processes at night; load curves showed clear industry-specific differences. If the EMS optimized only based on historical average loads, it easily overlooked changes in production plans. The project team therefore combined load forecasting with production scheduling, weather data, equipment status, and historical operating curves to build a multi-factor forecasting model.
Control safety was also a key constraint in engineering implementation. The park energy system was connected to production, and any misoperation could affect product quality or cause downtime. Therefore, the project adopted a tiered control strategy: implementing automatic closed-loop control for non-production-critical equipment, implementing advisory control for production-related equipment, and setting manual confirmation and permission isolation for high-risk operations.
III. Solution: EMS Platform, Edge Acquisition, and Source-Grid-Load-Storage Coordination
The overall solution was centered on an energy management system (EMS), building a four-layer architecture: "perception layer—network layer—platform layer—application layer." The perception layer was responsible for data acquisition from electricity, gas, cooling, heating, water, PV, storage, HVAC, and air compressor equipment; the network layer transmitted data back via industrial Ethernet, RS485, LoRa, and 5G private networks; the platform layer completed data cleaning, storage, modeling, and analysis; and the application layer provided functions such as energy consumption monitoring, energy efficiency optimization, energy-saving diagnosis, demand management, carbon accounting, and operation and maintenance work orders.
In terms of the metering system, the project expanded the park's metering points from approximately 280 to approximately 1,460, achieving sub-metering for main incoming lines, buildings, workshops, key equipment, and utility systems. Electricity metering coverage reached 100%; independent metering was installed for chiller plants, air compressors, boiler rooms, and energy storage systems. Acquisition frequency for key equipment was increased to 1 minute; main incoming lines and important branches were archived every 15 minutes; and operating data from some HVAC chillers and storage PCS supported second-level sampling for anomaly diagnosis and control optimization.
EMS Core Modules
- Energy Overview
- Sub-metering Statistics
- Load Forecasting (MAPE ~4.8%)
- Efficiency Optimization
- Demand Control
- Carbon Management
Source-Grid-Load-Storage Strategy
Unified dispatch of 2.1 MW PV and 1.5 MW/3 MWh storage. Prioritize production load when PV output is high; discharge storage before evening peaks; charge strategically during off-peak hours. SOC controlled between 20%-90% with 15% backup reserve.
IV. Implementation Process: Phased Retrofit and Production-Interruption-Free Switchover
The project was implemented in six stages over approximately 38 weeks: diagnosis (6 weeks), design (4 weeks), field implementation (12 weeks), commissioning (6 weeks), trial operation (12 weeks), and acceptance evaluation. To reduce impact on production, the project adopted zone-based construction and gray-release rollout strategies. The first stage deployed only acquisition devices without changing existing control logic. Subsequent stages progressively optimized public areas, then conducted closed-loop tests on chiller plants and air compressors with automatic rollback mechanisms if alarms occurred. Finally, demand control and PV-storage coordination were incorporated into automatic dispatch with manual confirmation thresholds. During the entire implementation, no production interruption occurred due to energy system retrofitting.
Chiller Plant Retrofit: By adding sensors and VFDs, the system optimized chiller combinations. When cooling load was below 45%, screw chillers were prioritized, increasing COP from ~4.1 to ~4.6 and improving plant efficiency by ~12%.
V. Energy Efficiency Optimization and Energy-Saving Strategies
Central HVAC System
The EMS dynamically adjusted temperatures and frequencies based on outdoor conditions and electricity prices. Pre-cooling 30 minutes before peak hours utilized building thermal inertia. This system saved approximately 2.1 million kWh annually (34.3% of total savings).
Air Compressor System
Pressure band optimization reduced average operating pressure by ~0.04 MPa, resulting in ~6% energy reduction. Nighttime shutdown of small compressors further saved energy. Total savings: ~0.98 million kWh annually.
Lighting & Auxiliary Systems
Upgraded to light-sensing and time-control LED fixtures with motion sensing in garages. Tiered office lighting controls. Savings: ~0.36 million kWh annually. Low investment with short implementation cycles.
Demand Management
Automated warnings triggered adjustments in HVAC temps, compressor pressures, and storage discharge when predicted demand approached thresholds. Monthly max demand dropped from 18.6 MW to 17.4 MW (~6.5% reduction), significantly cutting demand charges.
PV-Storage Coordination
Self-consumption rate rose from 78% to 89%. Daily storage cycles increased from 1.2 to 1.8. Annual comprehensive revenue from this subsystem was ~RMB 1.32 million, driven by arbitrage, demand reduction, and improved self-consumption.
VI. Key Parameters and System Integration
The project defined strict parameters to ensure the EMS was operable, not just visual. Data completeness required ≥99.5%, with alarm response <15 mins. Time sync deviation ≤5 seconds. Raw data retained for 2 years, aggregated for 5 years. Security features included role-based permissions, two-factor authentication, and four-level command hierarchy (read-only, advisory, auto, emergency stop).
Integration with business systems was critical. The EMS interfaced with electricity billing, MES (for production shifts/orders), and asset management systems. This allowed answering complex questions like "which enterprise/equipment/time period caused abnormal energy consumption?" Algorithms combined layered forecasting with rule engines to translate analysis into executable strategies.
VII. Project Results: Verifiable Savings and Carbon Reductions
Third-party verification confirmed a comprehensive energy-saving rate of ~4.6%. Approximately 55% of savings came from system optimization rather than equipment replacement. The total investment of ~RMB 21 million yielded a static payback period of ~3.2 years. Including PV/storage benefits, annual comprehensive energy costs decreased by ~RMB 5.2 million.
Operational efficiency improved drastically. Anomaly discovery time dropped from ~3 days to <15 minutes. Unplanned shutdowns halved. Manual statistical workload decreased by ~60%. Enterprise feedback highlighted stable energy use and transparent costs, with one electronics firm seeing an 11% drop in demand charges without impacting production plans.
VIII. Outlook: Expanding to Carbon Assets and Virtual Power Plants
This case demonstrates that industrial parks must move beyond single-equipment updates to establish digital energy management systems. The "diagnose-meter-optimize-operate" path is scalable. Future steps include extending to carbon asset management, linking energy data with product footprints, and participating in green electricity trading. If the park's green electricity share rises from 8% to 25%, Scope 2 emissions could drop another 15-20%.
Virtual Power Plants (VPP) represent a major opportunity. With adjustable loads estimated at 2.5-3.5 MW, the park can participate in regional demand response and ancillary services, transforming from a cost center into a revenue-generating energy node. Technological evolution will involve digital twins for simulation and AI dispatch for continuous optimization under dynamic market and weather conditions.
Ultimately, park energy management is a long-term operational process. It provides the data foundation, direct financial returns, and strategic pathway for manufacturing parks to achieve dual-carbon goals through systematic engineering implementation.