October 30, 2026
Who Pays for the Grid? Prosumers, Reverse Flow, and the Challenge of Charging for Wire Use
Series "Cost allocation in the era of prosumers" – Article 1 of 3
Abstract
This is the first article in a three-part series that presents, in an accessible manner, the key findings from TR Soluções' technical report on cost allocation in the era of prosumers. Here we outline the problem and the concepts needed to understand it, without requiring prior knowledge of the power sector: the portion of the electricity bill that pays for the distribution network (the "wire") is divided based on logic that assumes electricity flows in one direction only, from the substation to the consumer. Distributed generation has broken that premise. In the following articles, we show how the study proposes to resolve the problem and what it found in Cemig's data.
How this series is organized
- Article 1 (this one) – the problem: basic concepts, how the wire is currently paid for, and what changed with prosumers.
- Article 2 – the method: how to separate consumption and injection and calculate the cost of each.
- Article 3 – the results: when the grid becomes more expensive, what price signals emerge, and what the study's limitations are.
At the end of this article is a glossary (Box 1) with terms used throughout the series.
1. Introduction: the portion of the bill that pays for the wire
The electricity bill pays for several things simultaneously: the energy generated, the transport of that energy to your region, transmission losses along the way, sector charges, and the use of the distribution company's local network. This series addresses one specific component of that bill: the "wire", that is, the infrastructure of poles, cables, transformers, and substations belonging to the utility.
In tariff terminology, the cost of this infrastructure is part of the so-called Parcela B (Portion B) and is charged in the network-use component through the Wire B component of the TUSD (Distribution System Use Tariff). The question that organizes this entire series is simple to ask and difficult to answer: how should this cost be fairly divided among consumers and generators?
2. Three concepts to begin with
2.1 Energy and power are not the same thing. Imagine a faucet and a water tank. Power (measured in kW) is the flow rate of the faucet: how much electricity is being drawn at one instant. Energy (measured in kWh) is the volume that has accumulated in the tank: everything that was consumed over a period. The energy bill depends on volume; the size of the network depends on the moment of maximum flow.
Figure 1 – Energy is the area under the curve; peak power is the highest point
Source: own elaboration. Illustrative diagram, no actual data.
2.2 Electricity travels across "stages" of voltage. Electricity exits power plants, travels along transmission lines, and arrives at the utility at high voltage. From there, it descends through successive levels: high voltage (A2 and A3 groups, used by large industries), medium voltage (A4), and low voltage (BT), where residences, shops, and small businesses are located. Each level requires its own equipment, and lower levels depend on all those above.
Figure 2 – The journey of electricity and voltage levels in the distribution network
Source: own elaboration. Illustrative diagram.
2.3 The network is built for the peak. A highway is designed for the hour of heaviest traffic, not for the daily average. The same applies to the electrical network: it must have the capacity for the hour of maximum loading, even if for much of the day there is spare capacity. This is why investment in network, and therefore the cost to be divided, is determined by the peak, not just by the total volume of energy consumed.
Figure 3 – Network capacity is defined by the peak; outside it, there is spare capacity
Source: own elaboration. Illustrative diagram.
3. How the wire is currently divided
Dividing the wire bill means answering: how much does each type of user contribute to the need to expand the network? The Brazilian regulatory response uses the Marginal Cost of Capacity (CMC), which estimates, in BRL/kW, how much it costs to increase network capacity to serve each type of user at each hour of the day. Those who have greater impact during peak hours and use more network layers have higher CMC.
There is also a cascade effect. To reach a residence in low voltage, electricity passes through nearly the entire chain of assets, from high voltage to the street transformer. A large high-voltage consumer (A2) uses only the top portion of that chain. This is why capacity cost increases as you move down the voltage level.
Figure 4 – Which network layer is used by whom and the average CMC of each level
Source: TR Soluções report, ESTRUTURA platform, Aneel data from Cemig 2023. Diagram of layers elaborated by the authors of this article.
In practice, Aneel calculates these costs at each periodic tariff review, based on typical curves: 24-hour profiles obtained from measurement campaigns that represent groups of similar consumers and network sections. The method follows Submodule 7.2 of PRORET and operates statically, once per cycle.
4. What changed: the street became two-way
The prosumer is the unit that both consumes and generates energy, generally with solar panels. During the day, solar generation can exceed the home's consumption, and the surplus is injected into the grid. At night, without sun, the home relies on the grid again.
Figure 5 – Profile of a house with solar generation: energy surplus at midday, grid dependence at night
Source: own elaboration. Illustrative diagram.
When many homes do this in the same neighborhood, the surplus can exceed local consumption and "flow up" through the transformer toward medium voltage. This is reverse flow: it is as if a street designed for one-way traffic began to have traffic in both directions.
Figure 6 – One-way flow in late afternoon and two-way flow at midday
Source: own elaboration. Illustrative diagram.
5. Why the current model has difficulty
Data from the utility analyzed by TR Soluções shows the scale of the phenomenon: of the 228 typical load curves, 112 represent prosumers; of the 141 network curves, 15 exhibit flow reversal, 11 of them in A4/BT transformers. The current model was designed for unidirectional flow, and this creates three effects:
- Mathematics that fails: when generation and consumption are equal and net flow approaches zero, the formulas produce division by zero and negative coefficients with no physical interpretation.
- Data discarded: curves with flow reversal were discarded. According to verbal information from Aneel technicians, the model also cannot accommodate all remaining curves.
- Results affected: discarding curves is never neutral, as each curve contributes to the marginal cost of its class and voltage level.
Figure 7 – Typical curves: what exists in the data and what the current model utilized
Source: TR Soluções report, ESTRUTURA platform, Aneel data from Cemig 2023.
6. Why this matters for the tariff
If the analysis only "sees" consumption, injection has no price signal of its own. Today, the reference tariff for generators (TUSDg Wire B) derives from the consumption tariff, with fixed weights of 10% for the Peak period and 90% for Off-Peak. The study describes it as an echo of the consumption tariff, not as a measure of the actual cost of reverse flow.
When the tariff does not reflect who stresses the grid, cross-subsidies emerge: one group of users pays part of the cost imposed by another group. The example in Figure 8, with hypothetical numbers, illustrates the logic: if two users pay the same amount but impose different costs on the grid, whoever exerts less pressure subsidizes whoever exerts more.
Figure 8 – Cross-subsidy: what each user pays versus the cost they impose on the grid (hypothetical example)
Source: own elaboration. Illustrative diagram.
7. The central idea and the next step
Rather than abandoning official methodology, the study proposes separating the analysis into two environments: Direct Flow (the network serving consumption) and Reverse Flow (the network evacuating surplus generation). Each environment receives its own Marginal Cost of Capacity, so that consumption and injection no longer mix and each can receive an appropriate price signal.
It is important to delimit the scope: the study quantifies a reference tariff that reflects costs, but does not evaluate whether it should be applied. That depends on regulatory policy and the legal framework, such as the Electrical Energy Compensation System and Law No. 14,300/2022.
In the next article, we explain step by step how this separation works.
| Term | What it means |
|---|---|
| Power (kW) | How much electricity is used at one instant. Defines the required size of the network. |
| Energy (kWh) | Total consumed over a period. Defines the energy bill. |
| Voltage levels | A2 (88 to 138 kV) and A3 (69 kV): high voltage; A4 (2.3 to 25 kV): medium; BT (below 2.3 kV): low. |
| TUSD, Wire B, and Parcela B | TUSD is the distribution network use tariff; Wire B remunerades the utility's infrastructure (Portion B of revenue). |
| CMC | Marginal Cost of Capacity: cost of expanding the network to serve each user at each hour. |
| Typical curve | Representative 24-hour profile of consumption, generation, or network loading, obtained by measurement. |
| Prosumer and MMGD | Unit that consumes and generates energy (typically solar); MMGD is micro and mini-scale distributed generation. |
| Reverse flow | Energy injected by prosumers that travels the network in the opposite direction of consumption. |
| Cross-subsidy | Situation in which one group pays part of the cost imposed by another. |
| SCEE | Electrical Energy Compensation System, which regulates how surplus injection is compensated. |
* About this series. The three articles summarize the technical report "Cost allocation in the era of prosumers: separation of price signals for load and generation" (TR Soluções, version 3.8, 2026), available at https://doi.org/10.5281/zenodo.23439480. The results use data from Cemig (2023 tariff revision).