Hybrid energy storage optimization control method based on fuzzy control dynamic adjustment

Abstract

The invention discloses a hybrid energy storage optimization control method based on fuzzy control dynamic adjustment. The method comprises the steps that S1, the SOC of a storage battery is calculated; s2, respectively acquiring and collecting power data of wind power generation, photovoltaic power generation and load, forming a power fluctuation signal according to the acquired power data, and then decomposing the power fluctuation signal into different frequency components; and S3, dynamically controlling the power response of the pumped storage system and the storage battery according to the frequency component of the power fluctuation signal and the SOC of the storage battery. According to the method, the power fluctuation signal is subjected to frequency decomposition, and the power distribution is dynamically adjusted by combining the frequency component of the power fluctuation signal and the SOC, so that the overall operation performance of hybrid energy storage is optimized, accurate matching of the power of different energy storage systems can be realized, the energy storage systems can be ensured to work in a safe range, and overcharge or overdischarge is avoided. The multi-state joint estimation is carried out on the storage battery model, so that the safe and stable operation of the system can be ensured.

Description

Technical Field

[0001] This invention relates to the field of energy storage control technology, and more specifically to a hybrid energy storage optimization control method based on fuzzy control dynamic adjustment. Background Technology

[0002] my country's energy transition is accelerating, and the proportion of renewable energy in primary energy consumption will continue to increase, rapidly replacing fossil fuels. However, the randomness and volatility of new energy power generation, such as wind and solar power, pose challenges to the safe operation of the power grid and threaten its security. The increasing number of new energy vehicles also increases the demand on the power grid, especially during peak hours. To ensure the safe and stable operation of the power grid, the introduction of energy storage devices is an effective technical means.

[0003] Hybrid energy storage systems (HESS) have become an effective means of mitigating power fluctuations from renewable energy sources due to their ability to simultaneously provide high power density and high energy density. The hybrid configuration of pumped hydro storage (PHS) and batteries (such as lithium-ion batteries) can fully leverage the former's advantages in low-frequency, high-capacity regulation and the latter's advantages in high-frequency, rapid response.

[0004] Hybrid energy storage systems can comprehensively leverage the advantages of different energy storage technologies, adapt to various scenario requirements, significantly improve the grid's regulation capabilities, and ensure the efficient consumption of renewable energy and grid security. However, current hybrid energy storage control strategies typically use fixed frequency band allocation (such as high-pass / low-pass filtering) to distribute power commands, without considering the dynamic changes in the battery's state of charge (SOC). This may lead to overcharging / over-discharging of the energy storage system, affecting its lifespan and safety.

[0005] Therefore, in the face of the existing problems, developing a hybrid energy storage optimization control method that can dynamically control the power response of pumped hydro storage system and battery based on the frequency components of power fluctuation signals and the SOC of the battery is an urgent technical problem to be solved. Summary of the Invention

[0006] The purpose of this invention is to provide a hybrid energy storage optimization control method based on fuzzy control dynamic adjustment, in order to solve the problem that existing hybrid energy storage control strategies do not consider the dynamic changes in the state of charge (SOC) of the battery, which may lead to overcharging / over-discharging of the energy storage system, affecting its lifespan and safety.

[0007] To address the aforementioned technical problems, this invention provides a hybrid energy storage optimization control method based on fuzzy control dynamic adjustment, comprising the following steps:

[0008] S1: Calculate the SOC of the battery;

[0009] S2: Collect power data from wind power generation, photovoltaic power generation, and load respectively, form a power fluctuation signal, and decompose the power fluctuation signal into different frequency components;

[0010] S3: Dynamically control the power response of the pumped storage system and the battery based on the frequency components of the power fluctuation signal and the SOC of the battery.

[0011] Furthermore, the SOC calculation model for the battery is as follows:

[0012]

[0013]

[0014] Among them, S x (t) represents the state of charge (SOC) of the battery at time t; Q rx Q represents the remaining capacity of the battery. x λ represents the rated capacity of the battery. x,ch and λ x,dis These are the charging and discharging efficiencies of the battery, respectively. and These represent the charging and discharging power of the battery, respectively; S x,min and S x,max These are the minimum and maximum SOC of the battery, respectively; P x,min and P x,max These are the upper and lower limits of the battery's charging and discharging power.

[0015] Further, step S3 includes:

[0016] When the frequency component of the power fluctuation signal is high-frequency power, the high-frequency power response of the battery is controlled separately.

[0017] When the frequency component of the power fluctuation signal is medium frequency power, the fuzzy control method is used to dynamically control the power response of the battery and the pumped storage system based on the frequency component of the power fluctuation signal and the SOC of the battery.

[0018] When the frequency component of the power fluctuation signal decomposition is low-frequency power, the low-frequency power response of the pumped storage system is controlled separately.

[0019] Furthermore, when controlling the high-frequency power response of the battery independently, the SOC of the battery is monitored in real time. If the SOC is less than the first threshold and the sum of the wind power generation and photovoltaic power generation is less than the load power, the pumped storage system is started to charge the battery until the SOC of the battery rises to the ratio between the first threshold and 50%, where the first threshold is less than 50%. If the SOC is greater than the second threshold and the sum of the wind power generation and photovoltaic power generation is greater than the load power, the pumped storage system is started to absorb the battery's electrical energy until the SOC of the battery drops to the ratio between 50% and the second threshold, where the second threshold is greater than 50%.

[0020] Furthermore, when the frequency component of the power fluctuation signal decomposition is intermediate frequency power, the power response of the battery and pumped storage system is distributed through a first-order low-pass filter. After the target power response of the hybrid energy storage system is filtered by the first-order low-pass filter, the power distribution response of the battery and pumped storage system is as follows:

[0021]

[0022]

[0023] Where T is the filtering time, s is the differential operator, and P HESS For the target power response of the hybrid energy storage system, P PS For the low-frequency power distribution response of the pumped storage system, P bat For high-frequency power distribution response of the battery.

[0024] Furthermore, when using a first-order low-pass filter to distribute power between the battery and the pumped storage system, the filtering time of the first-order low-pass filter is optimized based on the battery's state of charge (SOC). The filtering time optimization methods include:

[0025] When 0 ≦ SOC <SOC minL At that time, the optimized filtering time T1 = T max T max This is the maximum filtering time;

[0026] When SOC minL ≦SOC <SOC minH At that time, the optimized filtering time T1 is:

[0027]

[0028] When SOC minH ≦SOC <SOC maxL When this happens, the filtering time T of the first-order low-pass filter is maintained;

[0029] When SOC maxL ≦SOC <SOC maxHAt that time, the optimized filtering time T1 is:

[0030]

[0031] When SOC maxH When SOC is less than or equal to 100%, the filtering time T1 is set to 0.

[0032] Where 0≦SOC <SOC minL <SOC minH <SOC maxL <SOC maxH ≤100%.

[0033] Furthermore, when controlling the low-frequency power response of the pumped storage system alone, the SOC of the battery is monitored in real time. If the SOC is less than the first threshold and the sum of the wind power generation and photovoltaic power generation is less than the load power, the pumped storage system is started to charge the battery until the SOC of the battery rises to the ratio between the first threshold and 50%, where the first threshold is less than 50%. If the SOC is greater than the second threshold and the sum of the wind power generation and photovoltaic power generation is greater than the load power, the pumped storage system is started to absorb the battery's electrical energy until the SOC of the battery drops to the ratio between 50% and the second threshold, where the second threshold is greater than 50%.

[0034] Furthermore, the method also includes: performing joint state estimation of the battery from SOC to SOP to SOH, and adopting a three-level progressive mechanism of real-time-periodic-offline to perceive the state of the battery in all dimensions.

[0035] Furthermore, the combined state of charge (SOC), state of charge (SOP), and state of harmonics (SOH) estimation of the battery includes:

[0036] Based on the battery model, the extended Kalman filter algorithm is used to estimate the SOC.

[0037] Based on the battery model, the peak charge and discharge current of the battery model is calculated, and then the SOP is estimated by combining the voltage design limit, SOC limit and maximum charge and discharge current constraint.

[0038] By analyzing the time and cycle period required for the voltage to drop to a preset value during the discharge process of the battery model, an improved double-exponential model is constructed for SOH estimation.

[0039] Furthermore, the battery status is comprehensively perceived using a three-tiered progressive mechanism of real-time, periodic, and offline methods, including:

[0040] In the real-time processing layer, based on the dynamic equivalent circuit model of the battery model, the extended Kalman filter algorithm is used to drive the state observer through high-frequency voltage-current data stream to evaluate and update the SOC in real time.

[0041] In the cycle processing layer, historical running data is extracted using a sliding time window, and the time-varying characteristics of the parameter set of the dynamic equivalent circuit model are obtained by combining the least squares method, and the SOC and SOP are dynamically updated.

[0042] In the offline processing layer, the capacity decay trajectory is analyzed by long-term data mining and performance degradation characteristics, and key feature parameters are extracted to quantify SOH.

[0043] The beneficial effects of this invention are as follows: By performing frequency decomposition on the power fluctuation signal and combining the frequency components of the power fluctuation signal with the state of charge (SOC) to dynamically adjust the power allocation, the overall operating performance of hybrid energy storage is optimized. This not only enables precise power matching between different energy storage systems but also ensures that the energy storage system operates within a safe range, avoiding overcharging or over-discharging. Furthermore, by performing multi-state joint estimation on the battery model, the safe and stable operation of the system can be ensured. Attached Figure Description

[0044] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, use the same reference numerals to denote the same or similar parts. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0045] Figure 1 This is a flowchart of one embodiment of the present invention;

[0046] Figure 2 Block diagram of the control strategy when the frequency component of the power fluctuation signal is high-frequency power;

[0047] Figure 3 Strategy diagram for classifying the state of charge of batteries;

[0048] Figure 4 This is a graph showing the SOC membership function of a battery.

[0049] Figure 5 The graph shows the membership function of power imbalance.

[0050] Figure 6 Membership function graph for filter time adjustment;

[0051] Figure 7 A block diagram of the control strategy when the frequency component of the power fluctuation signal is low-frequency power.

[0052] Figure 8 The block diagram of the second-order RC equivalent circuit model of the battery;

[0053] Figure 9 A diagram showing the estimated SOC of a battery.

[0054] Figure 10 Estimated SOP (Start of Production) diagram for a storage battery;

[0055] Figure 11 A diagram showing the estimated state of harm (SOH) of a storage battery.

[0056] Figure 12 Framework diagram for joint state estimation of battery (SOC-SOP-SOH);

[0057] Figure 13 Estimated SOC-SOP-SOH of the battery. Detailed Implementation

[0058] like Figure 1 The hybrid energy storage optimization control method based on fuzzy control dynamic adjustment shown includes the following steps:

[0059] S1: Calculate the SOC of the battery;

[0060] S2: Collect wind power generation data, photovoltaic power generation data, and load power data respectively, form a power fluctuation signal, and decompose the power fluctuation signal into different frequency components;

[0061] S3: Dynamically control the power response of the pumped storage system and the battery based on the frequency components of the power fluctuation signal and the SOC of the battery.

[0062] According to one embodiment of this application, the SOC calculation model for a storage battery is as follows:

[0063]

[0064]

[0065] Among them, S x (t) represents the state of charge (SOC) of the battery at time t; Q rx Q represents the remaining capacity of the battery. x λ represents the rated capacity of the battery. x,ch and λ x,dis These are the charging and discharging efficiencies of the battery, respectively. and These represent the charging and discharging power of the battery, respectively; S x,min and S x,max These are the minimum and maximum SOC of the battery, respectively; P x,min and P x,max These are the upper and lower limits of the battery's charging and discharging power.

[0066] Considering both safety and service life, and also having the ability to smooth out wind power output at the next moment, this embodiment sets SOC constraints and charge / discharge constraints to ensure that the energy storage device operates within its SOC and charge / discharge power allowable range.

[0067] According to one embodiment of this application, step S3 includes:

[0068] When the frequency component of the power fluctuation signal is high-frequency power, the high-frequency power response of the battery is controlled separately. By adopting the independent response of the battery to high-frequency power, its fast charging and discharging advantages are fully utilized, and the dynamic adjustment capability of the system is improved.

[0069] When the frequency component of the power fluctuation signal is medium-frequency power, the fuzzy control method is used to dynamically control the power response of the battery and the pumped storage system based on the frequency component of the power fluctuation signal and the SOC of the battery. By using fuzzy control to dynamically allocate medium-frequency power, the coordinated operation of pumped storage and battery is optimized, so that the advantages of the battery can be fully utilized according to the dynamic SOC of the battery. This not only helps to realize the coordinated scheduling between energy storage systems, but also ensures its fast and high-precision response capability.

[0070] When the frequency component of the power fluctuation signal is low-frequency power, the low-frequency power response of the pumped storage system is controlled separately. By using the pumped storage system to independently respond to low-frequency power, pumped storage typically has a considerable capacity, is much cheaper than lithium-ion batteries, and has a long service life, and can usually eliminate significant power fluctuations.

[0071] According to one embodiment of this application, when controlling the high-frequency power response of the battery independently, the SOC of the battery is monitored in real time. If the SOC is less than a first threshold and the sum of wind power generation and photovoltaic power generation is less than the load power, the pumped storage system is started to charge the battery until the SOC of the battery rises to the ratio between the first threshold and 50%, and then stops. During this process, even if the bus power changes, for example, the sum of wind power generation and photovoltaic power generation exceeds the load power, the pumped storage should not stop. If the SOC is greater than a second threshold and the sum of wind power generation and photovoltaic power generation exceeds the load power, the pumped storage system is started to absorb the battery's electrical energy until the SOC of the battery drops to the ratio between 50% and the second threshold, so that the battery releases additional energy to ensure that it can absorb wind and solar energy. The first threshold is less than 50%, and the second threshold is greater than 50%.

[0072] According to one embodiment of this application, when the frequency component of the power fluctuation signal decomposition is intermediate frequency power, the power response of the battery and the pumped storage system is distributed through a first-order low-pass filter. The transfer function of the first-order low-pass filter is:

[0073]

[0074] After the target power response of the hybrid energy storage system is filtered by a first-order low-pass filter, the power distribution response of the battery and the pumped hydro storage system is as follows:

[0075]

[0076]

[0077] Where T is the filtering time, s is the differential operator, and P HESS For the target power response of the hybrid energy storage system, P PS For the low-frequency power distribution response of the pumped storage system, P bat For high-frequency power distribution response of the battery.

[0078] According to one embodiment of this application, when using a first-order low-pass filter to distribute power between a battery and a pumped-storage system, the filtering time of the first-order low-pass filter is optimized based on the battery's state of charge (SOC). The filtering time optimization method includes:

[0079] When 0 ≦ SOC <SOC minL When the signal is high, it indicates that the battery's state of charge (SOC) is low, and the battery bank is absorbing additional electrical energy from the wind-solar system. In this case, a longer filtering time can be considered to allow the battery to absorb more energy, thereby raising its SOC to a reasonable range. Assume the maximum filtering time is T. max Then, the optimized filtering time T1 = T max .

[0080] When SOC minL ≦SOC <SOC minH At that time, the optimized filtering time T1 is:

[0081]

[0082] During charging, as the battery's state of charge (SOC) increases, the filtering time T1 decreases from its maximum value T. max The value changes to T, which will correspondingly transfer some of the fluctuating power initially absorbed by the pumped storage system to the battery for absorption.

[0083] When SOC minH ≦SOC <SOC maxL During this period, the battery's state of charge (SOC) is at its optimal level, requiring no additional operation. Therefore, it is only necessary to keep the filtering time constant, i.e., T1 = T.

[0084] When SOC maxL ≦SOC <SOCmaxH At that time, the optimized filtering time T1 is:

[0085]

[0086] In this scenario, as the battery's state of charge (SOC) increases, a portion of the fluctuating power that was originally handled by the battery is continuously transferred to the pumped-storage hydroelectric system. The pumped-storage system then needs to absorb the energy that should have been absorbed by the battery, so the filtering time decreases as the battery's SOC increases. When the battery's SOC reaches its upper limit, the filtering time drops to 0, and the pumped-storage system takes over the fluctuating power.

[0087] When SOC maxH When SOC is less than or equal to 100%, the battery's state of charge (SOC) is high and it cannot continue to absorb additional power from wind-solar power; all energy should be absorbed by the pumped storage system, at which point T1 = 0.

[0088] Where 0≦SOC <SOC minL <SOC minH <SOC maxL <SOC maxH ≤100%, the classification of the state of charge of a storage battery is as follows: Figure 3 As shown.

[0089] According to one embodiment of this application, when controlling the low-frequency power response of the pumped storage system alone, the SOC of the battery is monitored in real time. If the SOC is less than a first threshold and the sum of wind power generation and photovoltaic power generation is less than the load power, the pumped storage system is started to charge the battery until the SOC of the battery rises to the ratio between the first threshold and 50%, where the first threshold is less than 50%. If the SOC is greater than a second threshold and the sum of wind power generation and photovoltaic power generation is greater than the load power, the pumped storage system is started to absorb the battery's electrical energy until the SOC of the battery drops to the ratio between 50% and the second threshold, where the second threshold is greater than 50%.

[0090] The specific power control strategy based on fuzzy control is as follows:

[0091] (1) Input variables

[0092] Battery SOC (State of Charge): Represents the remaining percentage of battery charge, ranging from 0% to 100%. In fuzzy control, its fuzzy set is divided into {low (L), medium-low (ML), medium (M), medium-high (MH), high (H)}.

[0093] Power Imbalance: It represents the difference between the wind-solar power generation and the load power, and its value range can be set according to the actual situation, such as -1000kW to 1000kW. In fuzzy control, its fuzzy set is divided into {Negative Big (NB), Negative Small (NS), Zero (ZO), Positive Small (PS), Positive Big (PB)}.

[0094] (2) Output variable

[0095] Filter Time Adjustment (FTA): It represents the adjustment amount of the filtering time T of the first-order low-pass filter, and the unit is second. Its fuzzy set is divided into {Negative Big (NB), Negative Small (NS), Zero (ZO), Positive Small (PS), Positive Big (PB)}.

[0096] Membership function

[0097] (1) Membership function of input variable

[0098] The membership function of the SOC of the battery is as Figure 5 shown.

[0099] Low (L): When SOC ≤ 20%, the membership degree is 1; when 20% < SOC ≤ 30%, the membership degree linearly decreases to 0; when SOC > 30%, the membership degree is 0.

[0100] Medium Low (ML): When 20% < SOC ≤ 30%, the membership degree linearly increases to 1; when 30% < SOC ≤ 40%, the membership degree linearly decreases to 0; when SOC ≤ 20% or SOC > 40%, the membership degree is 0.

[0101] Medium (M): When 30% < SOC ≤ 40%, the membership degree linearly increases to 1; when 40% < SOC ≤ 60%, the membership degree is 1; when 60% < SOC ≤ 70%, the membership degree linearly decreases to 0; when SOC ≤ 30% or SOC > 70%, the membership degree is 0.

[0102] Medium High (MH): When 60% < SOC ≤ 70%, the membership degree linearly increases to 1; when 70% < SOC ≤ 80%, the membership degree linearly decreases to 0; when SOC ≤ 60% or SOC > 80%, the membership degree is 0.

[0103] High (H): When 70% < SOC ≤ 80%, the membership degree linearly increases to 1; when SOC > 80%, the membership degree is 1; when SOC ≤ 70%, the membership degree is 0.

[0104] (2) The membership function of power imbalance is as Figure 5 shown.

[0105] Negative Large (NB): When the power imbalance ≤ -500 kW, the membership degree is 1; when -500 kW < power imbalance ≤ -250 kW, the membership degree linearly decreases to 0; when the power imbalance > -250 kW, the membership degree is 0.

[0106] Negative Small (NS): When -500 kW < power imbalance ≤ -250 kW, the membership degree linearly increases to 1; when -250 kW < power imbalance ≤ 0 kW, the membership degree linearly decreases to 0; when the power imbalance ≤ -500 kW or power imbalance > 0 kW, the membership degree is 0.

[0107] Zero (ZO): When -250 kW < power imbalance ≤ 0 kW, the membership degree linearly increases to 1; when 0 kW < power imbalance ≤ 250 kW, the membership degree linearly decreases to 0; when the power imbalance ≤ -250 kW or power imbalance > 250 kW, the membership degree is 0.

[0108] Positive Small (PS): When 0 kW < power imbalance ≤ 250 kW, the membership degree linearly increases to 1; when 250 kW < power imbalance ≤ 500 kW, the membership degree linearly decreases to 0; when the power imbalance ≤ 0 kW or power imbalance > 500 kW, the membership degree is 0.

[0109] Positive Large (PB): When 250 kW < power imbalance ≤ 500 kW, the membership degree linearly increases to 1; when the power imbalance > 500 kW, the membership degree is 1; when the power imbalance ≤ 250 kW, the membership degree is 0.

[0110] (3) The membership function of the filter time adjustment amount is as Figure 6 shown.

[0111] Negative Large (NB): When FTA ≤ -20 s, the membership degree is 1; when -20 s < FTA ≤ -10 s, the membership degree linearly decreases to 0; when FTA > -10 s, the membership degree is 0.

[0112] Negative Small (NS): When -20 s < FTA ≤ -10 s, the membership degree linearly increases to 1; when -10 s < FTA ≤ 0 s, the membership degree linearly decreases to 0; when FTA ≤ -20 s or FTA > 0 s, the membership degree is 0.

[0113] Zero (ZO): When -10 s < FTA ≤ 0 s, the membership degree linearly increases to 1; when 0 s < FTA ≤ 10 s, the membership degree linearly decreases to 0; when FTA ≤ -10 s or FTA > 10 s, the membership degree is 0.

[0114] Positive Small (PS): When 0s < FTA ≤ 10s, the membership degree linearly increases to 1; when 10s < FTA ≤ 20s, the membership degree linearly decreases to 0; when FTA ≤ 0s or FTA > 20s, the membership degree is 0.

[0115] Positive Big (PB): When 10s < FTA ≤ 20s, the membership degree linearly increases to 1; when FTA > 20s, the membership degree is 1; when FTA ≤ 10s, the membership degree is 0.

[0116] Rule Explanation:

[0117] When the measured battery SOC is Low (L) and the power imbalance is Negative Big (NB), it indicates that the measured battery has insufficient power and the system has excessive power. It is necessary to significantly increase the filtering time (output is PB) so that the measured battery absorbs more energy to quickly increase the SOC.

[0118] When the measured battery SOC is Medium (M) and the power imbalance is Zero (ZO), it indicates that the measured battery power is within a reasonable range and the system power is balanced. There is no need to adjust the filtering time (output is ZO), and the current filtering time can be maintained.

[0119] When the measured battery SOC is High (H) and the power imbalance is Positive Big (PB), it indicates that the measured battery has sufficient power and the system has insufficient power. It is necessary to significantly reduce the filtering time (output is NB) so that the measured battery absorbs less energy to prevent overcharging, and at the same time let the pumped-storage energy storage承担 more power regulation tasks.

[0120] According to an embodiment of the present application, the method further includes: performing a joint state estimation of SOC - SOP - SOH on the battery, and adopting a three - level progressive mechanism of real - time - periodic - offline to comprehensively perceive the state of the battery in all dimensions.

[0121] According to an embodiment of the present application, the joint state estimation of SOC - SOP - SOH on the battery includes:

[0122] State of Charge (SOC) estimation: Based on the battery model, the Extended Kalman Filter algorithm is used to perform SOC estimation. The evaluation effect of the SOC estimation is as Figure 9 shown;

[0123] The Kalman Filter algorithm actually combines the open - circuit voltage method and the ampere - hour method, and continuously corrects the SOC reference value obtained by the ampere - hour method with the relationship between the open - circuit voltage and SOC depending on the accuracy of the model. The Extended Kalman Filter algorithm, based on the Kalman Filter algorithm, uses linearization and the Jacobian matrix to enable the algorithm to handle nonlinear system problems. Considering the real - time requirement of the energy storage SOC estimation, therefore, the second - order RC model of the chemical energy storage system (such as Figure 8(As shown) is simplified to a first-order RC to improve the solution rate of state estimation.

[0124] Based on the above-mentioned chemical energy storage system model, the dynamic equations and state-space equations of the equivalent circuit are established according to Kirchhoff's laws as follows:

[0125]

[0126] U t =U OCV (SOC)-U1-I·R0(2)

[0127] x=[SOC U1](3)

[0128]

[0129]

[0130] The state equations and measurement equations for the nonlinear discrete-time system are as follows:

[0131] x k =f(x) k-1 ,u k-1 )+w k 6)

[0132] y k =g(x k )+v k (7)

[0133] Among them, w k It is process error; v k These are measurement errors. Both errors follow a normal distribution and are governed by covariance matrices Q and R, respectively.

[0134] To linearize the above equation, we first define the Jacobian matrix to obtain:

[0135]

[0136] Discretizing the state equations by combining equations (4) and (5) yields:

[0137]

[0138]

[0139] Peak Power (SOP) Estimation: Based on the battery model, the peak charge / discharge current of the battery model is calculated. Then, SOP is estimated by combining the voltage design limit, SOC limit, and maximum charge / discharge current constraint. The SOC estimation evaluation effect is as follows: Figure 10 As shown;

[0140] The method for estimating the peak power (SOP) of a battery is as follows:

[0141] Based on the above battery model, the current of the power battery at any given time can be calculated as follows:

[0142]

[0143] Considering the voltage design limit U t,min ≤U t ≤U t,max That is, the battery must always operate at the charging cutoff voltage U. t,max and discharge cutoff voltage U t,min Between. Therefore, the estimated peak charging and discharging current is:

[0144]

[0145] In the formula and These are the estimated values ​​of the peak discharge current and peak charging current based on the power battery model at time k, respectively.

[0146] Based on the maximum SOC limit of the power battery given in energy storage control max and minimum limit SOC min The method for calculating the maximum current that a battery can withstand under current conditions is called the peak current estimation method based on battery SOC. The peak discharge current at time k can be obtained from the following formula. and peak charging current

[0147]

[0148] The peak current calculation method based on the SOC design limit of the power battery takes into account the peak current during safe operation within the time interval Δt, ensuring the safety of the actual charging and discharging process of the power battery.

[0149] Based on the two peak current calculation methods mentioned above, and the limits for the use of power batteries (maximum discharge current - I), max and minimum charging current - I min The multi-constraint dynamic peak current obtained from comprehensive calculation is:

[0150]

[0151] In the formula, and These are the peak discharge current and peak charging current at time k, respectively.

[0152] State of Health (SOH) estimation: By analyzing the time and cycle number of the battery model during discharge, an improved bi-exponential model is constructed for SOH estimation. The SOH estimation evaluation results are as follows: Figure 11 As shown;

[0153] This method proposes a SOH estimation model based on an improved double-exponential model. This model simultaneously considers two health indicators, Cycle and TDV, to fit the relationship between SOH and cycle life and TDV of the battery. The improved double-exponential model is as follows:

[0154] Q Discharge =a·e b·Cycle +c·e d·Cycle(TDV) (15)

[0155] Compared to the traditional bi-exponential model that only considers the cycle number, this model has stronger robustness and adaptability, and can better handle model lag caused by changes in the charging and discharging process or sudden cessation.

[0156] According to one embodiment of this application, the battery status is comprehensively perceived using a three-level progressive mechanism of real-time-periodic-offline, including:

[0157] In the real-time processing layer, the dynamic equivalent circuit model based on the battery model (parameter set: ohmic internal resistance R0, polarization resistance R1, polarization capacitance C1) adopts the extended Kalman filter algorithm, drives the state observer through high-frequency voltage-current data stream, evaluates and updates the SOC in real time, and can also record the time-series changes of key state parameters synchronously.

[0158] In the periodic processing layer, historical running data is extracted using a sliding time window, and the time-varying characteristics of the parameter set of the dynamic equivalent circuit model are obtained by combining the least squares method, thereby dynamically updating the SOC and SOP.

[0159] In the offline processing layer, the capacity decay trajectory is analyzed by long-cycle data mining and performance degradation characteristics, and key feature parameters are extracted to quantify SOH. Through SOH state updates, the capacity information is further updated into the SOC and SOP estimation algorithms to improve the estimation accuracy.

[0160] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A hybrid energy storage optimization control method based on fuzzy control dynamic adjustment, characterized in that, Including the following steps: S1: Calculate the SOC of the battery; S2: Collect power data from wind power generation, photovoltaic power generation and load respectively, and form a power fluctuation signal based on the acquired power data. Then decompose the power fluctuation signal into different frequency components. S3: Based on the frequency components of the power fluctuation signal and the SOC of the battery, dynamically control the power response of the pumped storage system and the battery.

2. The hybrid energy storage optimization control method based on fuzzy control dynamic adjustment according to claim 1, characterized in that, The SOC calculation model for the battery is as follows: Among them, S x (t) represents the state of charge (SOC) of the battery at time t; Q rx Q represents the remaining capacity of the battery. x λ represents the rated capacity of the battery. x,ch and λ x,dis These are the charging and discharging efficiencies of the battery, respectively. and These represent the charging and discharging power of the battery, respectively; S x,min and S x,max These are the minimum and maximum SOC of the battery, respectively; P x,min and P x,max These are the upper and lower limits of the battery's charging and discharging power.

3. The hybrid energy storage optimization control method based on fuzzy control dynamic adjustment according to claim 1 or 2, characterized in that, Step S3 includes: When the frequency component of the power fluctuation signal is high-frequency power, the high-frequency power response of the battery is controlled separately. When the frequency component of the power fluctuation signal is medium frequency power, the fuzzy control method is used to dynamically control the power response of the battery and the pumped storage system based on the frequency component of the power fluctuation signal and the SOC of the battery. When the frequency component of the power fluctuation signal decomposition is low-frequency power, the low-frequency power response of the pumped storage system is controlled separately.

4. The hybrid energy storage optimization control method based on fuzzy control dynamic adjustment according to claim 2, characterized in that, When controlling the high-frequency power response of the battery independently, the SOC of the battery is monitored in real time. If the SOC is less than a first threshold and the sum of wind power and photovoltaic power is less than the load power, the pumped storage system is started to charge the battery until the SOC of the battery rises to the ratio between the first threshold and 50%, where the first threshold is less than 50%. If the SOC is greater than a second threshold and the sum of wind power and photovoltaic power is greater than the load power, the pumped storage system is started to absorb the battery's electrical energy until the SOC of the battery drops to the ratio between 50% and the second threshold, where the second threshold is greater than 50%.

5. The hybrid energy storage optimization control method based on fuzzy control dynamic adjustment according to claim 3, characterized in that, When the frequency component of the power fluctuation signal decomposes to intermediate frequency power, the power response of the battery and pumped storage system is distributed through a first-order low-pass filter. After the target power response of the hybrid energy storage system is filtered by the first-order low-pass filter, the power distribution response of the battery and pumped storage system is as follows: Where T is the filtering time, s is the differential operator, and P HESS For the target power response of the hybrid energy storage system, P PS For the low-frequency power distribution response of the pumped storage system, P bat For high-frequency power distribution response of the battery.

6. The hybrid energy storage optimization control method based on fuzzy control dynamic adjustment according to claim 5, characterized in that, When using a first-order low-pass filter to distribute power between a battery and a pumped-storage system, the filtering time of the first-order low-pass filter is optimized based on the battery's state of charge (SOC). Filtering time optimization methods include: When 0 ≦ SOC <SOC minL At that time, the optimized filtering time T1 = T max T max This is the maximum filtering time; When SOC minL ≦SOC <SOC minH At that time, the optimized filtering time T1 is: When SOC minH ≦SOC <SOC maxL When this happens, the filtering time T of the first-order low-pass filter is maintained; When SOC maxL ≦SOC <SOC maxH At that time, the optimized filtering time T1 is: When SOC maxH When SOC is less than or equal to 100%, the filtering time T1 is set to 0. Where 0≦SOC <SOC minL <SOC minH <SOC maxL <SOC maxH ≤100%.

7. The hybrid energy storage optimization control method based on fuzzy control dynamic adjustment according to claim 3, characterized in that, When controlling the low-frequency power response of the pumped storage system independently, the state of charge (SOC) of the battery is monitored in real time. If the SOC is less than a first threshold and the sum of wind power and photovoltaic power is less than the load power, the pumped storage system is started to charge the battery until the SOC of the battery rises to the ratio between the first threshold and 50%, where the first threshold is less than 50%. If the SOC is greater than a second threshold and the sum of wind power and photovoltaic power is greater than the load power, the pumped storage system is started to absorb the battery's electrical energy until the SOC of the battery drops to the ratio between 50% and the second threshold, where the second threshold is greater than 50%.

8. The hybrid energy storage optimization control method based on fuzzy control dynamic adjustment according to claim 1, characterized in that, The method also includes: performing joint state estimation of the battery (SOC-SOP-SOH) and using a three-level progressive mechanism (real-time-periodic-offline) to perceive the state of the battery in all dimensions.

9. The hybrid energy storage optimization control method based on fuzzy control dynamic adjustment according to claim 8, characterized in that, The combined state estimation of the battery (SOC-SOP-SOH) includes: Based on the battery model, the extended Kalman filter algorithm is used to estimate the SOC. Based on the battery model, the peak charge and discharge current of the battery model is calculated, and then the SOP is estimated by combining the voltage design limit, SOC limit and maximum charge and discharge current constraint. By analyzing the time and cycle period required for the voltage to drop to a preset value during the discharge process of the battery model, an improved double-exponential model is constructed for SOH estimation.

10. The hybrid energy storage optimization control method based on fuzzy control dynamic adjustment according to claim 9, characterized in that, The battery status is comprehensively monitored using a three-tiered progressive mechanism: real-time, periodic, and offline. In the real-time processing layer, based on the dynamic equivalent circuit model of the battery model, the extended Kalman filter algorithm is used to drive the state observer through high-frequency voltage-current data stream to evaluate and update the SOC in real time. In the cycle processing layer, historical running data is extracted using a sliding time window, and the time-varying characteristics of the parameter set of the dynamic equivalent circuit model are obtained by combining the least squares method, and the SOC and SOP are dynamically updated. In the offline processing layer, the capacity decay trajectory is analyzed by long-term data mining and performance degradation characteristics, and key feature parameters are extracted to quantify SOH.

Images