⚡ electrical engineering

Valuing urban building flexibility portfolios under carbon-constrained electricity dispatch: A Shenzhen case study

This study evaluates a Shenzhen building portfolio under carbon constraints, demonstrating that while temporal load shifting primarily reduces operating costs, net energy reduction is essential for alleviating binding emissions budgets, thereby supporting the need for distinct market definitions for these two flexibility services.

Original authors: Kaixiang Zhang, Huanxin Liao, Yifan Zhu, Zhenshang Wang, Yuming Zhao, Jing Wang, Junyao Gao, Junhua Zhao

Published 2026-08-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Kaixiang Zhang, Huanxin Liao, Yifan Zhu, Zhenshang Wang, Yuming Zhao, Jing Wang, Junyao Gao, Junhua Zhao

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Cities are alive with electricity. Every office light, every air conditioner, and every elevator draws power from the grid, creating a massive, shifting demand that must be met the instant it is needed. For decades, the energy industry has treated buildings as static consumers, assuming their power usage is fixed and unchangeable. However, a growing body of research suggests that buildings are actually flexible resources. They can be coaxed into shifting their power use to different times of day or reducing their total consumption, acting like a giant, invisible battery that helps balance the grid. This flexibility becomes even more critical as the world tries to cut carbon emissions. When electricity generation is constrained by strict limits on greenhouse gases, the ability to move or reduce demand can become a powerful tool for keeping the lights on without burning more fuel. The challenge lies in understanding exactly how much value this flexibility holds and how to define it correctly, because moving power around is fundamentally different from using less power.

A team of researchers from the Chinese University of Hong Kong, Shenzhen, and Shenzhen Power Supply Co., Ltd., set out to measure this value in one of the world's most dynamic cities. They focused on a specific question: does it matter if we treat the ability to shift energy use as a different product than the ability to reduce energy use? To answer this, they gathered hourly power records from 143 real buildings in Shenzhen, ranging from offices to hotels. Instead of looking at each building individually, which would be computationally overwhelming, they grouped them into three representative types based on their usage patterns. These groups formed a "portfolio" that acted as a stand-in for the entire city's flexible demand. The researchers then fed this data into a sophisticated computer model that simulated how the power grid operates under strict carbon limits. This model acted as a digital twin of the region's electricity system, calculating the most efficient way to generate and import power while staying within a set budget for carbon emissions.

The study revealed two distinct ways these buildings can help the grid, and the difference between them is significant. The first type of help is shifting. Imagine a building that runs its air conditioning heavily during the hottest part of the day but can wait until the evening to do so without making anyone uncomfortable. This is an energy-neutral shift; the building uses the exact same amount of electricity, just at a different time. The researchers found that this kind of shifting is excellent for lowering the cost of running the power system. By moving demand away from expensive peak hours, the system saves money on fuel and operations. In their simulation, when the researchers allowed a large portion of the city's buildings to shift their usage, the system saved nearly 0.4 million RMB in a single week. Crucially, this saving came without changing the total amount of carbon emitted, because the electricity was simply being used at a different time.

The second type of help is reduction. This is when a building actually uses less electricity than it normally would, perhaps by turning off lights or raising the thermostat slightly. This action is different because it physically removes energy from the grid. The researchers discovered that reduction becomes incredibly valuable when the power system is under a tight carbon budget. In their model, they imposed a strict limit on carbon emissions, forcing the system to find ways to operate with less pollution. Without this limit, the buildings only reduced their usage when it was cheap to do so, saving about 1,480 megawatt-hours. But when the carbon budget was tightened, the system demanded more reduction to stay within the legal limits. The buildings responded by cutting an additional 803 megawatt-hours, bringing the total reduction to over 2,200 megawatt-hours. This extra reduction allowed the system to avoid burning expensive, high-carbon fuel, which in turn lowered the overall cost of the carbon constraint itself.

The study also tested how well their simplified model of three building groups compared to a model that tracked all 143 buildings individually. The results were remarkably close. The simplified model predicted the total system costs and savings with an error of less than one-hundredth of a percent. This suggests that for the purpose of valuing these services, grouping buildings by their general behavior is a reliable and efficient method. However, the researchers noted that the specific value of the flexibility depends heavily on how the products are defined. Shifting helps the grid manage its daily costs, while reduction helps it meet its long-term environmental goals. If a city tries to sell these services as a single, blended product, it might miss out on the specific benefits each one offers.

Ultimately, the work provides a clear roadmap for how cities can monetize their buildings' flexibility. It shows that shifting and reducing are not the same thing, and they should be treated as separate products in the energy market. Shifting is a tool for economic efficiency, smoothing out the daily peaks and valleys of power use. Reduction is a tool for environmental compliance, directly lowering the carbon footprint when the grid is under pressure. By distinguishing between the two, grid operators and building owners can better understand the true value of their actions. In a world where carbon limits are becoming stricter, knowing whether to move your power use or cut it could be the difference between a costly, carbon-heavy grid and a clean, efficient one. The researchers' simulations, based on real data from Shenzhen, prove that this flexibility is not just a theoretical idea but a tangible asset that can save money and protect the environment simultaneously.

Technical Summary: Valuing Urban Building Flexibility Portfolios under Carbon-Constrained Electricity Dispatch

Problem Statement
Urban building portfolios are frequently modeled as aggregate demand, a practice that obscures the distinct operational and economic effects of temporal energy shifting versus net energy reduction. Current demand-response models often fail to explicitly distinguish between energy-neutral shifts (changing when electricity is consumed) and verified net reductions (changing how much is consumed). This distinction is critical because shifting alters consumption timing relative to price and emission signals, while reduction alters the total volume of consumption, potentially relieving binding emissions constraints. Furthermore, existing valuation frameworks often lack a direct connection between measured building aggregation, explicit product definitions, and specific outcomes in operating costs and emissions budgets.

Methodology
The authors propose a subsystem valuation framework that converts hourly building records into dispatch-oriented portfolios, evaluated within a two-stage optimization structure:

Data Processing and Portfolio Construction:
- Using hourly load data from 148 buildings in Shenzhen, the study screens and clusters 143 buildings into three substantive archetypes based on area-normalized load profiles (lighting, plug, AC, motive power, special loads).
- The clustering (k-means with $K=5$ , reduced to 3 groups after removing outliers) preserves the physical scale of total load and controllable AC envelopes.
- Portfolios are scaled to represent specific shares ( $\rho$ ) of the Shenzhen subsystem's weekly electricity demand, up to a stress-test limit of 60% ( $\rho=0.60$ ).
Two-Stage Dispatch Framework:
- Upper Layer (Provincial DC-OPF): A Guangdong provincial Direct Current Optimal Power Flow model generates fixed boundary signals for the city model. These include the import price and an offer-derived provincial average operational emissions proxy. This layer ensures city decisions do not feedback into provincial dispatch.
- Lower Layer (Shenzhen Carbon-Constrained Dispatch): A city-level model co-optimizes local generation, imports, and building flexibility subject to an operational emissions budget.
- Flexibility Definitions: The model defines two distinct product types:
  - A-type (Shifting): Energy-neutral shifting ( $s^+, s^-$ ) bounded by AC envelopes and ramp limits, with daily energy conservation.
  - B-type (Reduction): Adds a bounded net reduction ( $r$ ) to the shifting capability, subject to daily reduction limits and separate cost structures.
Evaluation Metrics:
- The study reports changes in subsystem operating costs, accounted emissions, the emissions-budget dual (marginal cost of tightening the budget), and the dual value of reduction availability.
- Scenarios include varying emissions budget tightness (0.05% to 0.60% reduction from baseline) and flexibility costs.

Key Contributions

Explicit Product Differentiation: The framework explicitly separates energy-neutral shifting and net reduction into distinct variables with separate costs and constraints, allowing for a granular analysis of their respective system values.
Physical-Scale Aggregation: The methodology constructs portfolios that retain the absolute scale and controllable envelopes of the underlying building data, moving beyond statistical similarity to physical deliverability.
Marginal Value Analysis: The study quantifies the "emissions-budget dual" to determine the marginal operating cost of tightening the carbon constraint, distinguishing between cost-driven reduction and carbon-induced response.
Aggregation Validation: The paper benchmarks the 3-archetype model against a 143-building detailed model and a single-cluster ( $K=1$ ) model to quantify aggregation errors in both total objectives and matched savings.

Results

Rigid Demand Response: Under a 0.60% emissions budget tightening, the rigid demand model increases operating costs by 0.258% and raises the emissions-budget dual to 242.51 RMB/tCO2. The system responds by substituting local dispatch with imports, despite the proxy indicating higher emissions for imports, highlighting an accounting-driven response.
A-type (Shifting) Performance: At a 60% portfolio share and 0.60% budget tightening, A-type shifting reduces operating costs by 0.389 million RMB (0.268%) but leaves the emissions-budget dual nearly unchanged. This confirms that shifting primarily provides temporal operating-cost value without significantly altering the binding emissions constraint.
B-type (Reduction) Performance:
- Economic vs. Carbon-Induced: Without an emissions constraint, B-type flexibility dispatches 1,480 MWh of reduction (driven by a 200 RMB/MWh cost). With the 0.60% budget tightened, reduction increases to 2,283 MWh. The additional 803 MWh represents the carbon-constraint-induced response.
- System Impact: Under the tight budget, B-type flexibility saves 1.102 million RMB (0.760%) relative to rigid demand and significantly lowers the emissions-budget dual from 242.51 to 88.41 RMB/tCO2.
Aggregation Accuracy: The 3-cluster ( $K=3$ ) model shows negligible difference (0.011%) in the total objective compared to the 143-building model. However, matched savings differ by 3.28% for A-type and 1.41% for B-type, indicating that while clustering is descriptive for homogeneous parameters, specific group parameters are needed for precise valuation.

Significance and Claims
The paper claims that the distinction between temporal shifting and net reduction is fundamental to building flexibility valuation. The results support the definition of distinct market products for these services:

Shifting (A-type) provides value primarily through temporal optimization of operating costs and does not significantly relieve binding emissions budgets.
Net Reduction (B-type) provides dual value: it reduces operating costs and, crucially, can relieve binding accounted-emissions constraints, thereby lowering the marginal cost of the carbon budget.

The authors conclude that building-flexibility valuation and product design must preserve this distinction to accurately capture the operational consequences of different demand-response strategies. The study does not claim to establish annual or citywide realizable potential but demonstrates the operational consequences of treating these services separately within a controlled, single-week subsystem model.

Technical Summary: Valuing Urban Building Flexibility Portfolios under Carbon-Constrained Electricity Dispatch

More like this