Popular Science Article: How can an independent energy storage power station make money?

Imagine the power grid as a relay race that never stops. The power generators are the “runners,” responsible for generating electricity (running).

The loads on the consumption side are the “pedometers,” representing the total electricity consumption of all users (the number of steps consumed in the run).

The power grid’s frequency of 50Hz is the “pace” of this race. In principle, the amount of electricity generated must be immediately used up, just as the runner’s running speed must perfectly match the pedometer’s consumption rate for the pace (frequency) to remain stable at 50Hz.

The problem is that pedometers (user electricity consumption) are completely uncontrollable. Factories suddenly start operating, millions of people turn on their air conditioners at the same time, TV commercials are playing and everyone is using the restroom and turning on their lights… electricity consumption fluctuates randomly every second. This is like a pedometer suddenly accelerating wildly. If the athletes (power plants) can’t keep up, the pace (frequency) of the entire team will slow down (below 50Hz), which may lead to machine damage or even large-scale power outages.

Frequency modulation is designed to solve this speed mismatch problem.

So how does the frequency regulation market operate to meet demand? (The Mechanism of the Frequency Regulation Market)

The traditional approach is to designate a few key players (such as thermal power and hydropower) to be ready to accelerate or decelerate at any time. However, this is uneconomical and not responsive enough.

Now, with the “frequency regulation market,” its operating logic is as follows: The power grid dispatch center (the referee) predicts the required “rapid regulation capacity” for the next day and then issues a recruitment notice: “Tomorrow, x MW of frequency regulation capacity is needed. Who has this capacity? Who will participate in the bidding?” Your independent energy storage power station, some modified thermal power plants, hydropower stations, etc., can all apply: “I have the capacity, and my bid is xx yuan/MW.” The dispatch center selects the best and most economical “players” based on the bid and technical performance (like an athlete’s explosive power and obedience). Once selected, even if there’s no work all day, you still receive a “capacity fee.” This is to compensate for the opportunity cost of being constantly on standby and maintaining a state (keeping the battery level at an intermediate value, ready to charge and discharge at any time).

In the actual competition, when the pedometer (electrical load) fluctuates, the dispatch center (referee) immediately orders the selected “athletes”: “Energy storage power station A, immediately accelerate power generation (discharge)! Thermal power plant B, immediately slow down (reduce power generation)!” This adjustment command occurs every second.

The power grid will precisely measure “how much electricity you actually adjusted, how accurately you adjusted it, and how quickly.” Based on this actual workload, you will receive a “mileage fee” or performance bonus.

In this “competition,” independent energy storage power stations have obvious advantages over thermal and hydroelectric power stations. Extremely fast response: Traditional thermal power plants take several minutes to adjust to full power from receiving an instruction, while energy storage has a millisecond-level response. Precise control: If you are given a 5MW acceleration, you can accurately output 5MW, without the slightest error.

Two-way regulation: It can both “accelerate power generation” (discharge) and “decelerate absorption” (charge).

This is why AGC response speed and regulation accuracy are so important in the design of energy storage power stations—they directly determine your “performance coefficient” and earning potential. In the performance evaluation (K-value) of the frequency regulation market, energy storage power stations usually score very high, which is equivalent to a high “performance coefficient,” meaning they can earn more “performance bonuses” than traditional power plants for the same work.

You can think of the frequency regulation market as a bidding and performance-based system where you “hire high-paid sprint bodyguards.”

Frequency regulation mileage refers to the amount of electricity (MWh) you effectively regulate per second based on the instructions given by the grid. This is your actual “workload.”

The comprehensive frequency regulation performance index K-value is your “capability performance coefficient,” scored based on the speed, accuracy, and compliance of your energy storage station. This coefficient determines how many times higher your performance bonus you can earn for the same amount of work. The clearing price refers to the market determining a uniform hourly wage for employers through daily bidding.

Suppose today is “hiring day” in a local frequency regulation market. The grid company releases a demand: 100MW of frequency regulation capacity is needed today.

Your energy storage station, other energy storage systems, thermal power plants, etc., all submit bids: “I bid 90 yuan/MWH,” “I bid 80 yuan,”… The market computer sorts the bids from lowest to highest until the 100MW demand is met. The last bid selected (e.g., 80 yuan) becomes the uniform “hourly wage” for the entire market today—this is the clearing price.

Hourly wages alone aren’t enough; real earnings depend on performance. The power grid will score you in real time across three dimensions, calculating a K-value. Response speed: How long does it take you to reach the designated location after receiving an instruction? Energy storage is measured in seconds, while thermal power is measured in stages. Regulation accuracy: If you’re instructed to regulate 5MW, can you accurately regulate to 5.00MW? Response accuracy: Did you execute every instruction correctly? Did you adjust in reverse in a timely manner? These three scores combined determine your K-value. The competition (actual operation) begins. The grid frequency fluctuates minutely every second, and dispatch instructions are issued every second: “Accelerate 0.5MW,” “Decelerate 0.3MW,” etc. The computer will accumulate the power changes corresponding to all your correct and effective regulation actions, integrating the results to calculate your total frequency regulation mileage. For example, if you provide 100MWh of effective regulation power to the grid through countless charging and discharging operations throughout the day, your mileage is 100MWh.

Daily frequency regulation revenue = (Capacity Fee) + (Mileage Performance Bonus) Frequency regulation revenue = Clearing Price × Your Winning Bid Capacity × Time + Clearing Price × K-Value × Frequency Regulation Mileage Substituting the assumed figures: On this day, the clearing price is 80 yuan/MW, the winning bid capacity is 100MW, the K-value is 4.0, the frequency regulation mileage is 100MWh, and the time is 24 hours. Capacity Fee = 80 yuan/MW × 100MW × 24 hours = 192,000 yuan; Mileage Performance Bonus = 80 yuan/MW × 4 × 100MWh = 32,000 yuan; Total daily revenue = 192,000 + 32,000 = 224,000 yuan. This is the ingenuity of the frequency regulation market mechanism. It uses a price discovery mechanism (clearing price) to ensure economic efficiency and uses performance indicators (K value) to incentivize technological progress, allowing the fastest athletes to receive the highest rewards, thereby ensuring the second-level security of the entire power grid in the most economical and efficient way.

When several independent energy storage power stations participate in frequency regulation simultaneously, the competition between them is not a “race,” but a “job application competition” centered around “cost-effectiveness” and “performance.” This competition occurs in two stages: “job application” and “performance ranking.” Continuing with the analogy of a “high-salary bodyguard company,” there are now multiple bodyguards (energy storage power stations) bidding for the service. Stage 1: Who gets the job? Every day, the power grid company conducts bidding for the frequency regulation service for the next day. Each energy storage power station must submit a “resume,” the core content of which is: “How much frequency regulation capacity can I provide?” For example, Station A bids 100MW, Station B bids 50MW. “What is my hourly wage requirement?” This is the most crucial part, namely the “frequency regulation capacity quotation,” in yuan/MW. For example, Station A bids 90 yuan/MW, Station B bids 85 yuan/MW, and Station C bids 95 yuan/MW. The power grid company sorts all bids from lowest to highest, much like arranging resumes by salary requirements. They then begin “hiring” from the cheapest bid until the total capacity hired meets the day’s demand. The last bid accepted becomes the unified “clearing price” for the entire market that day. For example, if the grid needs 200MW, and four stations—A (90 yuan), B (85 yuan), C (95 yuan), and D (100 yuan)—bid, the order would be: B (85 yuan) → A (90 yuan) → C (95 yuan) → D (100 yuan). After hiring B and A, the total capacity is 150MW, leaving a 50MW shortfall. Therefore, station C is hired. The total capacity is now 200MW, meeting the demand. Station C’s bid of 95 yuan/MW becomes the unified clearing price for the day. Stations B, A, and C all win the bid at 95 yuan/MW per hour. Station D, with its bid as high as 100 yuan, is eliminated in this round of competition and receives no income for the day.

The competitive strategy at this stage is to offer a competitive low price to ensure one’s own “operation,” while hoping that others’ bids won’t be too low, lest they lower the final clearing price. This is a subtle game. The second stage: Who “performs better”? All the winning power stations are operational, receiving a “standby wage” (capacity fee) of 95 yuan/MW. But the real bulk of income—the “performance bonus” (mileage fee)—is just beginning to be competed for. The performance bonus calculation formula is: Performance Bonus = Clearing Price × K-value × Actual Adjustment Mileage. K-value competition (“capacity coefficient” showdown): This is the ultimate showdown of technical strength between energy storage power stations. The K-value is determined by response speed, adjustment accuracy, and response correctness. The grid simultaneously issues instructions to stations A and B to “increase output by 10MW.” Station A accurately outputs 10.00MW within 0.1 seconds, with a K-value as high as 4.5; Station B outputs 9.8MW within 0.5 seconds, with a K-value of 2.0. For the exact same 1MWh regulation mileage, station A receives 2.25 times the performance bonus as station B (4.5 / 2.0 = 2.25 times). Grid regulation commands are real-time and random. The dispatch system is more inclined to assign regulation commands to stations with better performance, more stable status, and higher availability. If station A consistently responds the fastest and most accurately, the dispatch system will consider it “more reliable” and prioritize assigning commands to station A when fine-tuning is needed. This means that, within the same timeframe, a station with a higher K-value not only commands a higher “unit price” but may also “grab” more actual workload (mileage), thus achieving a dual lead in revenue and market share.

The calculation of the K-value is not a simple linear correspondence, but a multi-dimensional, standardized, real-time data-based assessment system. It’s like a comprehensive “college entrance examination report card,” assessing speed, accuracy, and stability. Typically, the comprehensive FM performance index K = K1 × K2 × K3, where:

K1: Response speed index (assessing “explosive power”);

K2: Adjustment accuracy index (assessing “control”);

K3: Response accuracy/stability index (assessing “endurance”).

Below, we’ll substitute data from an example to see how it’s calculated. Detailed step-by-step calculation explanation.

Suppose the power grid issues a command: at time T0, the target output will increase by 10MW.

Step 1: Calculate the response speed K1 (compared to “burst force”)—the time required from the command issuance (T0) until the actual output first reaches 90% of the target change.

The shorter the time, the higher the score. For example, the requirement is to achieve this within 1 minute (60 seconds). A baseline time (e.g., 15 seconds) is usually set; a response time less than or equal to the baseline time receives full marks, and scores decrease linearly beyond that. Station A calculates that it reaches 9MW (90% of 10MW) in 0.1 seconds.

This speed is much faster than the baseline time, K1 ≈ 1.0 (full marks).

Station B calculates that it reaches this in 0.5 seconds. Also much faster than the baseline, K1 ≈ 1.0 (full marks). For the response speed alone, as long as it’s fast enough (milliseconds), both A and B can get full marks. The real difference lies in the last two items.

Step 2: Calculate the adjustment accuracy K2 (compared to “control force”) over a period after the adjustment process stabilizes (e.g., 30 to 60 seconds after the command is issued), the average deviation between the actual output and the target output. The smaller the deviation, the higher the score. K2 = 1 – (mean absolute deviation / command change).

If the deviation is 0, then K2 = 1.0 (full marks). Station A stabilizes at 10.00MW with an average deviation of 0. K2 = 1.0. Station B stabilizes at an average output of around 9.8MW with an average deviation of 0.2MW. K2 = 1 – (0.2/10) = 1 – 0.02 = 0.98. Station A achieves full marks for accuracy, while Station B loses 0.02 points due to insufficient accuracy.

Step 3: Calculate the response accuracy K3 (compared to “endurance and sense of direction”). This is the most complex and easiest to lose points on. It assesses two aspects: the correctness of the adjustment direction. Since the command requires an increase in output, your output curve must always increase positively without any reverse fluctuations. If the command is “+10MW”, but you discharge to +10MW first and then charge back down to +8MW, this is called “reverse regulation” and will result in a severe deduction of points.

The accuracy of power contribution is the ratio of the actual regulated power provided (area under the curve) to the theoretically required regulated power provided during the command’s duration. Higher accuracy results in a higher score.

K3 is a comprehensive score, requiring your output curve to be not only fast and accurate but also stable, perfectly matching the command like a smooth straight line.

Station A stabilizes at 10.00MW from 0.1 seconds onwards, with a perfect straight line and no reverse fluctuations, indicating 100% accurate power contribution. K3 ≈ 1.0.

Station B may experience two situations leading to point deductions: a slight overshoot or oscillation during the rise from 0MW to 9.8MW (e.g., initially reaching 10.2MW and then falling back to 9.8MW); or failure to maintain stability after stabilization, with slight fluctuations around 9.8MW. These will be considered poor regulation quality. Assume K3 = 0.95.

The final K-value calculation and result comparison are as follows: Station A’s comprehensive K-value K_A = K1 × K2 × K3 × proportional coefficient = 1.0 × 1.0 × 1.0 × 5 = 5.0. In the market, the product result is usually magnified (e.g., multiplied by 5) to make the K-value range between 0 and 5. Station A achieved a theoretically perfect score or near-perfect score.

Station B’s comprehensive K-value K_B = 1.0 × 0.98 × 0.95 × 5 = 4.655. Even though Station B’s response is 0.5 seconds faster, its K-value is significantly lower than the perfect Station A simply because of a 0.2MW difference in accuracy and minor fluctuations in the adjustment process. The K-value is not simply about “fast” or “accurate,” but a 360-degree quantitative assessment of adjustment quality without any blind spots.

For energy storage power station projects, optimizing the BMS balancing strategy, PCS control algorithm, and EMS command parsing aims to make each response infinitely close to the perfect curve of “instantaneous, accurate, and smooth”, so as to obtain the highest “performance bonus” with the highest K value in the fierce market competition.

In fact, the profitability of independent energy storage power stations extends beyond frequency regulation and capacity compensation; it has evolved into a diversified “revenue mix.”

The core profit models primarily include: Participating in the electricity spot market to profit from peak-valley price differences through “buying low and selling high”—this is the most promising fundamental source of income;

Providing various power ancillary services such as ramping and reactive power support;

Leasing energy storage capacity to renewable energy power stations to meet their allocation and storage requirements, thereby obtaining stable rental income;

Responding to grid demand-side management by discharging during peak electricity demand and receiving compensation.