Online EMD power distribution method for hybrid energy storage system

Abstract

The invention discloses an online EMD power distribution method for a hybrid energy storage system, and the method comprises the following steps: S1, importing historical power data, extracting average power and peak power, and carrying out the capacity configuration of the hybrid energy storage system through employing energy and power as constraint conditions; s2, setting a sliding window and adjusting a step length, realizing online EMD decomposition of power data, and reconstructing a signal component by using an entropy criterion as a basis; s3, based on the hybrid energy storage loss quantification model, calculating a capacity loss factor, and calculating energy storage capacity loss; and S4, introducing an equal cycle life constraint, dynamically adjusting a critical entropy value, and changing a reconstruction rule in real time so as to change high and low frequency components. According to the method, the capacity loss is adjusted in real time, so that the power of the hybrid energy storage system is reasonably distributed, the service life difference of the hybrid energy storage system is reduced, the performance reduction is small, the capacity loss is slow, the operation cost is low, the system stability is good, the hybrid energy storage performance is improved, and the service life is prolonged.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage management, and in particular to an online EMD power distribution method for a hybrid energy storage system. Background Art

[0002] Although traditional internal combustion engines are powerful and have relatively mature technology, their noise, pollution and energy consumption cannot be ignored, which leads to the limitations of traditional fuel vehicles and is destined to be replaced by new energy vehicles. With the continuous advancement of electric vehicle technology, the development of electric motors has gradually matured, and they have obvious advantages in environmental protection and energy efficiency.

[0003] For electric vehicles, the energy storage system is undoubtedly the most critical part, which determines the cost and also greatly affects the overall performance. Compared with a single energy storage device, the hybrid energy storage system can smooth power fluctuations and improve the quality of power. At the same time, the complementarity of lithium batteries and supercapacitors improves the energy and power density of energy storage and maximizes energy storage performance. However, the diversified energy storage system increases the complexity of the control strategy, and it is particularly important to give full play to the advantages of different energy storage media, optimize the overall operating performance of the hybrid energy storage system, and reasonably allocate the power of the hybrid energy storage system.

[0004] At present, most of the research on hybrid energy storage is mainly aimed at smoothing the output power fluctuations of electric vehicles and wind and solar power generation to ensure the stable operation of the hybrid energy storage system; at the same time, it optimizes power distribution, reduces the impact of high-frequency power on lithium batteries, and prolongs the service life of lithium batteries; or improves the reliability and economy of active distribution network systems, and does not consider the impact of the life difference during the use of hybrid energy storage systems. Due to the different discharge properties of supercapacitors and batteries, there are differences in the life loss of the two. If the difference is too large, the effective capacity of the lithium battery will decay rapidly relative to the supercapacitor, and the performance will be greatly reduced, which will limit the overall use effect of the energy storage system, and will accelerate the capacity loss of the supercapacitor and increase the operating cost of the hybrid energy storage system. Summary of the invention

[0005] The purpose of the present invention is to provide an online EMD power distribution method for a hybrid energy storage system to solve the above defects.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] An online EMD power distribution method for a hybrid energy storage system comprises the following steps:

[0008] S1. Import historical power data, extract average power and peak power, and use energy and power as constraints to configure the capacity of the hybrid energy storage system;

[0009] S2, setting the sliding window and adjusting the step size to realize the online EMD decomposition of the power data, and using the entropy criterion as a basis to reconstruct the signal components;

[0010] S3. Based on the hybrid energy storage loss quantification model, the capacity loss factors of lithium batteries and supercapacitors are calculated, and the capacity configuration and reconstructed signal components of the hybrid energy storage system are combined to obtain the energy storage capacity loss;

[0011] S4, introduce equal cycle life constraints, dynamically adjust the critical entropy value, and change the reconstruction rules in real time, thereby changing the high and low frequency components.

[0012] Preferably, in step S1, the capacity of the hybrid energy storage system is configured, and the specific steps are as follows:

[0013] S11. Extracting average power P from historical power data average , the data below the average power is used as the basis for estimating the capacity configuration of lithium batteries, and the capacity of lithium batteries is configured with power and energy as constraints;

[0014] S12. Extract peak power P from historical power data peak The difference between the peak power and the average power is used as the basis for estimating the supercapacitor capacity configuration, and the capacity of the supercapacitor is configured with power and energy as constraints.

[0015] Preferably, in step S11, the capacity of the lithium battery is configured, and the specific process is as follows:

[0016] The power constraint is that the rated power of the battery pack is not less than the average power, then:

[0017] P LB ≥P average (1)

[0018] Where P LB is the rated power of the battery, P LB There are the following relations:

[0019]

[0020] In the formula, E LB is the rated energy of the battery, Q LB is the rated capacity of the battery, U LB is the rated voltage of the battery; T is the entire working interval time, in hours;

[0021] Combining the above formula, we can get the lithium battery capacity configuration range:

[0022] E LB ≥P average T (3),

[0023] For lithium batteries, energy constraints control the battery's SOC fluctuation within a certain range to avoid overcharge and overdischarge, then:

[0024]

[0025] In the formula, SOC LBmax Indicates the upper limit of lithium battery SOC, SOC LBmin Indicates the lower limit of lithium battery SOC; E l is the energy consumed by the lithium battery, which is estimated from the average power in the original power data. The calculation formula is as follows:

[0026]

[0027] Where P l is the data with lower power than the average power in the original data, n is the number of data with higher power than the average power in the original data; t is the time of a single data, in seconds;

[0028] Combining the above formulas (4) and (5), the lithium battery capacity configuration range formula is as follows:

[0029]

[0030] Combining equations (3) and (6), we take the intersection of the two capacity configuration ranges to obtain the final lithium battery configuration capacity.

[0031] Preferably, in step S12, the capacity of the supercapacitor is configured, and the specific process is as follows:

[0032] The power constraint is that the rated power of the supercapacitor bank is not less than the peak power, so:

[0033] P SC ≥P peak -P average (7)

[0034] Where P SC is the rated power of the supercapacitor bank;

[0035] From the capacitance energy relation Available supercapacitor capacity configuration range:

[0036]

[0037] In the formula, C SC Indicates the rated capacity of the supercapacitor, U SC is the rated voltage of the supercapacitor bank;

[0038] For supercapacitors, voltage constraints control the voltage fluctuation of supercapacitors within a certain range, and control the SOC of supercapacitors to change within a specified range, then:

[0039]

[0040] In the formula, E SC Indicates the rated energy of the supercapacitor, E h It is the power of the original data that exceeds the average power, and its formula is as follows:

[0041] E h =E power -E l (10)

[0042] In the formula, E power is the sum of the original power data;

[0043] Combining equations (9) and (10) and the voltage fluctuation limit r%, the capacitor energy relationship Get the supercapacitor capacity configuration range:

[0044]

[0045] Combining equations (8) and (11), we take the intersection of the two capacity configuration ranges to obtain the final supercapacitor configuration capacity.

[0046] Preferably, in step S2, online EMD decomposition is performed, and the signal components are reconstructed using entropy criteria as a basis, and the specific process is as follows:

[0047] S21. In order to ensure the continuity of EMD decomposition, the power data is treated as one in and one out, and several cycles are obtained. Considering the constraints of EMD decomposition, sufficient power data is required, so the window size is set to G, and when the cycle reaches the Gth cycle, that is, the window is filled for the first time, the EMD decomposition is started for it and the subsequent cycles, and it does not participate in the decomposition stage. The system is powered by the battery alone.

[0048] S22. In order to reasonably reconstruct the high and low frequency components and adjust the reconstruction rules conveniently, the information entropy is used to calculate the fluctuation amplitude of the power data, so as to facilitate the distinction between the high and low frequency components and reconstruct the power; the formula for entropy value calculation is as follows:

[0049]

[0050] In the formula, p(x i ) represents a random number x i The probability of occurrence, X represents a certain imf component, S represents the entropy value of this component, and x represents a random data point on this component;

[0051] Formula (12) is used to calculate the entropy of the i imf components obtained by EMD decomposition of each cycle, and the entropy value is recorded as Si , and compare it with the critical entropy value S e Make a judgment, if S i ≥S e , that is, if the entropy value exceeds the set value, the i-th component is high frequency, otherwise it is low frequency.

[0052] Preferably, the specific steps of step S3 are as follows:

[0053] S31. Introduce crack propagation theory, and derive the material fracture propagation rate formula based on the material fracture time formula; consider the influence of lithium battery charge state fluctuation, throughput cycle and temperature, and obtain the lithium battery life loss coefficient. Finally, after superposition, the lithium battery life loss factor is obtained as follows:

[0054] λ LB =∑L 1 ·L 2 ·L 3 (13)

[0055]

[0056] In the formula, λ LB Indicates the lithium battery capacity loss factor, L 1 Indicates the loss factor 1, L 2 Indicates the loss factor 2, L 3 Indicates loss factor 3; K co is the throughput coefficient; N LB is the equivalent throughput cycle, which indicates the ratio of the actual charge and discharge capacity to the nominal capacity in a time period; S dev_LB is the normalized deviation of the state of charge, which reflects the degree of deviation between the actual value and the average value; K ex is the discharge depth index; T LB 、T ref_LB are the real-time temperature of the lithium battery and the reference temperature of rated operation; T LBa 、T refa_LB are the absolute temperatures of the real-time temperature of the lithium battery and the reference temperature of the rated working temperature; T f represents the temperature coefficient in the thermal aging model; t cycle is the time of one cycle, t life Indicates the calendar life when the capacity decays to 80% under rated conditions, in seconds; K soc is the average charge state coefficient; S avg_LB is the average state of charge;

[0057] The low-frequency component obtained by decomposition and reconstruction in step S2 is given to the lithium battery, and combined with the lithium battery configuration capacity obtained in step S1, it is substituted into the lithium battery life loss factor formula (13)-(16) to obtain the loss factor of each cycle, and then superimposed one by one to obtain the real-time capacity loss of the lithium battery;

[0058] S32. The Arrhenius equation is introduced. According to its generalized formula, and considering the influence of voltage and temperature on the life of supercapacitor, the functional relationship between loss rate and voltage and temperature is obtained, thereby obtaining the life loss factor of supercapacitor:

[0059]

[0060] In the formula, λ SC Represents the supercapacitor capacity loss factor; l 0 is the rated calendar life of the supercapacitor, in seconds; E a_SC Indicates the real-time energy of the supercapacitor; C is the rated capacity of the supercapacitor; V 0 is the rated voltage of the supercapacitor; T ref_SC is the initial temperature of the supercapacitor; R th_SC P is the relationship coefficient between operating temperature and charge and discharge power; SC It represents the real-time power of the supercapacitor when it is working; τ is the cycle length in seconds; T 0 is the rated temperature of the supercapacitor;

[0061] The high-frequency component obtained by decomposition and reconstruction in step S2 is given to the lithium battery, combined with the supercapacitor configuration capacity obtained in step S1, and is substituted into the supercapacitor life loss factor formula (17) to obtain the loss factor of each cycle, and then superimposed one by one to obtain the real-time capacity loss of the supercapacitor;

[0062] Preferably, the specific steps of step S4 are as follows:

[0063] S41, the capacity loss of the lithium battery and the supercapacitor obtained in step S3 is used to determine how to adjust the critical entropy value through equal cycle life constraints, and acts on the power allocation of the next cycle, so as to achieve rapid adjustment of power allocation;

[0064] S42. Ratio the capacity loss of lithium battery and supercapacitor to that of supercapacitor, denoted as R L / S , if R L / S ≠1, indicating that there is a difference in the loss between the two, and the critical entropy value S needs to be adjusted ek+1 =S ek +ΔS, where S ek represents the critical entropy value of the kth period, S ek+1 represents the critical entropy value of the k+1th cycle, ΔS represents the critical entropy value adjustment value; otherwise, the critical entropy value remains unchanged;

[0065] S43, critical entropy value S e It is the basis for distinguishing high-frequency and low-frequency components. The entropy of several components decomposed in each cycle is calculated one by one to obtain the mean entropy of each component. Then the mean entropy of the cycle is calculated and used as the initial critical entropy value. The formula is as follows:

[0066]

[0067] Where n is the number of components in a single cycle, and N is the number of cycles;

[0068] If the loss ratio of lithium battery and supercapacitor exceeds the set threshold, that is, R L / S ≥1, the critical entropy value is reduced and the adjustment value ΔS<0; otherwise, the critical entropy value is increased and the adjustment value ΔS>0; the range of the critical entropy adjustment value is:

[0069] S min ≤|ΔS|≤S max (19)

[0070] ΔS+S e1 ≥0 (20),

[0071] In the formula, S max It represents the maximum value of all IMF component entropy values, S min Indicates the minimum value of all IMF component entropy calculations.

[0072] The beneficial effects of the present invention are:

[0073] The online EMD power distribution method for the hybrid energy storage system of the present invention dynamically adjusts the critical entropy value, changes the reconstruction rules, adjusts the capacity loss in real time, and reduces the life difference of the hybrid energy storage system through the online EMD and entropy calculation combined algorithm, thereby avoiding the problem that the overall performance of the hybrid energy storage system is affected by the large difference in cycle life, and even accelerates the aging speed of the energy storage equipment and increases the operation and maintenance cost.

[0074] The online EMD power distribution method for the hybrid energy storage system of the present invention realizes reasonable distribution of the power of the hybrid energy storage system by adjusting the capacity loss in real time, reduces the life span difference of the hybrid energy storage system, has little performance degradation, slow capacity loss, low operating cost, good system stability, and improves the use effect of the system.

[0075] The present invention is applicable to energy optimization management in the fields of electric vehicles and microgrids, and improves the performance and service life of hybrid energy storage. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 : A schematic structural diagram of the hybrid energy storage system of the present invention;

[0077] Figure 2 : Flow chart of online EMD power allocation for equal cycle life of hybrid energy storage system of the present invention;

[0078] Figure 3 : Online EMD schematic diagram of the present invention;

[0079] Figure 4 : Schematic diagram of entropy calculation of the present invention;

[0080] Figure 5 : Schematic diagram of power signal reconstruction adjustment of the present invention;

[0081] Figure 6 : Original power instance diagram;

[0082] Figure 7 : Comparison of loss factors of lithium batteries of the present invention (unadjusted vs ΔS=±0.9);

[0083] Figure 8 : Comparison of loss factors of lithium batteries of the present invention (ΔS=±0.9vsΔS=±1.8);

[0084] Fig. 9 : Comparison of loss factors of lithium batteries of the present invention (ΔS=±1.8vsΔS=±2.7);

[0085] Fig.10 : Comparison of the supercapacitor loss factor of the present invention (unadjusted vs ΔS = ± 0.9);

[0086] Fig.11 : Comparison of the supercapacitor loss factor of the present invention (ΔS=±0.9vsΔS=±1.8);

[0087] Fig.12 : Comparison of the supercapacitor loss factor of the present invention (ΔS=±1.8vsΔS=±2.7);

[0088] Fig.13 : Comparison diagram of lithium battery-supercapacitor capacity loss of the present invention.

[0089] Figure 14: Fig.13 Enlarged view of parts Ⅰ-Ⅷ in the middle.

[0090] Comparison diagram of capacity loss of lithium battery-supercapacitor of the present invention. DETAILED DESCRIPTION

[0091] The present invention is further described below in conjunction with the embodiments. It should be noted that these are merely examples and illustrations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the claims, they should be deemed to fall within the protection scope of the present invention.

[0092] Embodiment 1:

[0093] Figure 6 To collect the power data of an electric tractor in plowing state, the sampling time is 8633s and the sampling frequency is 1Hz.

[0094] Figure 1 is a system structure diagram of the hybrid energy storage system of the present invention, Figure 2 The flow chart of online EMD power allocation for the cycle life of the hybrid energy storage system of the present invention is shown as Figure 1 , Figure 2 As shown, the online EMD power distribution method of the hybrid energy storage system of the present invention comprises the following steps:

[0095] S1. Import historical power data, extract average power and peak power, and use energy and power as constraints to configure the capacity of the hybrid energy storage system.

[0096] S11. Extracting average power P from historical power data average =10.26kW, average power The following data is used as the basis for estimating the capacity configuration of lithium batteries. With power and energy as constraints, the capacity of lithium batteries is configured. The specific process is as follows:

[0097] The power constraint is that the rated power of the battery pack is not less than the average power, then:

[0098] P LB ≥P average (1)

[0099] Where P LB is the rated power of the battery, P LB There are the following relations:

[0100]

[0101] In the formula, E LB is the rated energy of the battery, Q LB is the rated capacity of the battery, U LB is the rated voltage of the battery; T is the entire working interval time, in hours;

[0102] Combining the above formula, we can get the lithium battery capacity configuration range:

[0103] E LB ≥P average T (3),

[0104] For lithium batteries, energy constraints control the battery's SOC fluctuation within a certain range to avoid overcharge and overdischarge, then:

[0105]

[0106] In the formula, SOC LBmax Indicates the upper limit of lithium battery SOC, SOC LBmin Indicates the lower limit of lithium battery SOC; E l is the energy consumed by the lithium battery, which is estimated from the average power in the original power data. The calculation formula is as follows:

[0107]

[0108] Where P l is the data with lower power than the average power in the original data, n is the number of data with higher power than the average power in the original data; t is the time of a single data, in seconds;

[0109] Combining the above formulas (4) and (5), the lithium battery capacity configuration range formula is as follows:

[0110]

[0111] Combining power constraints and energy constraints, we take the intersection of the two capacity configuration ranges, where the battery SOC is controlled to fluctuate within 0.2-0.8, that is, SOC LBmax =0.8, SOC LBmin =0.2; time T = 2.4h, t = 1s; from the historical power data, we can get n = 4689, Substitute the formula to get the final lithium battery configuration capacity range E LB ≥29.96kWh, therefore, 30kWh is selected as the capacity of the battery pack for lithium battery loss calculation.

[0112] S12. Extract peak power P from historical power data peak =24kW, the difference between peak power and average power is used as the basis for estimating the configuration of supercapacitor capacity, and the capacity of supercapacitor is configured with power and energy as constraints. The specific process is as follows:

[0113] The power constraint is that the rated power of the supercapacitor bank is not less than the peak power, so:

[0114] P SC ≥P peak -P average (7)

[0115] Where P SC is the rated power of the supercapacitor bank;

[0116] From the capacitance energy relation Available supercapacitor capacity configuration range:

[0117]

[0118] In the formula, C SC Indicates the rated capacity of the supercapacitor, U SC is the rated voltage of the supercapacitor bank;

[0119] For supercapacitors, voltage constraints control the voltage fluctuation of supercapacitors within a certain range, and control the SOC of supercapacitors to change within a specified range, then:

[0120]

[0121] In the formula, E SC Indicates the rated energy of the supercapacitor, E h It is the power of the original data that exceeds the average power, and its formula is as follows:

[0122] E h =E power -E l (10)

[0123] In the formula, E power is the sum of the original power data;

[0124] Combining equations (9) and (10) and the voltage fluctuation limit r%, the capacitor energy relationship Get the supercapacitor capacity configuration range:

[0125]

[0126] Combined with power constraints and voltage constraints, the intersection of the two capacity configuration ranges is taken. Among them, the supercapacitor voltage fluctuation is controlled within 20%, that is, r = 20%; at the same time, the SOC is controlled to change within 0.4-0.6, that is, SOC SCmax =0.6, SOC SCmin =0.4; take the rated voltage of the supercapacitor bank U SC =300V, put the data into the formula to get the final supercapacitor configuration capacity C SC ≥1.1497F. Therefore, 1.15F is selected as the capacity of the supercapacitor bank for supercapacitor loss calculation.

[0127] S2, set the sliding window and adjust the step size to realize the online EMD decomposition of power data, and use entropy calculation as a basis to reconstruct the power signal components. The specific process is as follows:

[0128] S21, Figure 3 It is a schematic diagram of online EMD decomposition, such as Figure 3 As shown. In order to ensure the continuity of EMD decomposition, the power data is treated as "one in and one out" to obtain several cycles; considering the constraints of EMD decomposition, sufficient power data is required, so the window size is set to G, and when the cycle reaches the Gth cycle, that is, the window is filled for the first time, the EMD decomposition of it and subsequent cycles is started, and it does not participate in the decomposition stage. The system is powered by the battery alone. In order to make the decomposition go smoothly, the window size is set to 100, that is, G = 100.

[0129] S22, Figure 4 is a schematic diagram of entropy calculation, such as Figure 3 In order to reasonably reconstruct the high and low frequency components and adjust the reconstruction rules conveniently, the information entropy is used to calculate the fluctuation amplitude of the power data, so as to distinguish the high and low frequency components and reconstruct the power. The formula for entropy value calculation is as follows:

[0130]

[0131] In the formula, p(x i ) represents a random number x i The probability of occurrence, X represents a certain imf component, S represents the entropy value of this component, and x represents a random data point on this component;

[0132] Formula (12) is used to calculate the entropy of several IMF components obtained by EMD decomposition of each cycle, and the entropy value is recorded as S i , and compare it with the critical entropy value S e Make a judgment, if S i ≥S e , that is, if the entropy value exceeds the set value, the i-th component is high frequency, otherwise it is low frequency.

[0133] S3. Based on the hybrid energy storage loss quantification model, calculate the capacity loss factors of lithium batteries and supercapacitors, combine the capacity configuration of the hybrid energy storage system and reconstruct the signal components, and thus obtain the energy storage capacity loss. The specific steps are as follows: give the low-frequency component obtained by decomposition and reconstruction in step S2 to the lithium battery, and the high-frequency component to the supercapacitor, and combine the hybrid energy storage system configuration capacity obtained in step S1, and bring it into the lithium battery and supercapacitor life loss factor formula to obtain the loss factor of each cycle, and then superimpose them one by one to obtain the real-time capacity loss of the hybrid energy storage.

[0134] S31. Introduce crack propagation theory, and derive the material fracture propagation rate formula based on the material fracture time formula; consider the influence of lithium battery charge state fluctuation, throughput cycle and temperature, and obtain the lithium battery life loss coefficient. Finally, after superposition, the lithium battery life loss factor is obtained as follows:

[0135] λ LB =∑L 1 ·L 2 ·L 3 (13)

[0136]

[0137] In the formula, λ LB Indicates the lithium battery capacity loss factor, L 1 Indicates the loss factor 1, L 2 Indicates the loss factor 2, L 3 Indicates loss factor 3; K co is the throughput coefficient; N LB is the equivalent throughput cycle, which indicates the ratio of the actual charge and discharge capacity to the nominal capacity in a time period; S dev_LB is the normalized deviation of the state of charge, which reflects the degree of deviation between the actual value and the average value; K ex is the discharge depth index; T LB 、T ref_LB are the real-time temperature of the lithium battery and the reference temperature of rated operation; T LBa 、T refa_LB are the absolute temperatures of the real-time temperature of the lithium battery and the reference temperature of the rated working temperature; T f represents the temperature coefficient in the thermal aging model; t cycle is the time of one cycle, t life Indicates the calendar life when the capacity decays to 80% under rated conditions, in seconds; K soc is the average charge state coefficient; S avg_LB is the average state of charge.

[0138] The low-frequency component obtained by decomposition and reconstruction in step S2 is given to the lithium battery. Combined with the lithium battery configuration capacity obtained in step S1, it is substituted into the lithium battery life loss factor formula (13)-(16) to obtain the loss factor of each cycle, and then they are superimposed one by one to obtain the real-time capacity loss of the lithium battery.

[0139] Due to inconsistent losses, the reconstruction rules are adjusted in real time, and the lithium battery life loss factor also changes accordingly, such as Figure 7 , Figure 8 , Fig. 9As shown in the figure, the red curve represents the unadjusted lithium battery loss factor, and the blue curve represents the adjusted lithium battery loss factor. It can be observed that the two curves overlap a lot. This is because the adjustment scale is limited. The adjusted critical entropy value does not affect the power distribution in some areas, resulting in the high and low frequency components remaining unchanged, thus making the loss factor curves overlap; however, from some detailed figures, it can be observed that the blue curve is overall low and relatively flat compared to the unadjusted red curve, which indicates that the lithium battery loss is reduced after the adjustment and the change is relatively stable; and when the adjustment range is increased, the lithium battery loss factor continues to decrease, the peaks of the loss factor curves in some areas are eliminated, and the adjustment effect is more obvious.

[0140] S32. The Arrhenius equation is introduced. According to its generalized formula, and considering the influence of voltage and temperature on the life of supercapacitor, the functional relationship between loss rate and voltage and temperature is obtained, thereby obtaining the life loss factor of supercapacitor:

[0141]

[0142] In the formula, λ SC Represents the supercapacitor capacity loss factor; l 0 is the rated calendar life of the supercapacitor, in seconds; E a_SC Indicates the real-time energy of the supercapacitor; C is the rated capacity of the supercapacitor; V 0 is the rated voltage of the supercapacitor; T ref_SC is the initial temperature of the supercapacitor; R th_SC P is the relationship coefficient between operating temperature and charge and discharge power; SC It represents the real-time power of the supercapacitor when it is working; τ is the cycle length in seconds; T 0 is the rated temperature of the supercapacitor;

[0143] The high-frequency component obtained by decomposition and reconstruction in step S2 is given to the lithium battery. Combined with the supercapacitor configuration capacity obtained in step S1, it is substituted into the supercapacitor life loss factor formula (17) to obtain the loss factor of each cycle, and then they are superimposed one by one to obtain the real-time capacity loss of the supercapacitor.

[0144] Due to the inconsistent loss, the reconstruction rules are adjusted in real time, and the supercapacitor loss factor also changes accordingly, such as Fig.10 , Fig.11 , Fig.12As shown in the figure, the red curve represents the unadjusted supercapacitor loss factor, and the blue curve represents the adjusted loss factor. It can be observed that the overall fluctuation and amplitude of the adjusted loss factor are higher than before the adjustment, and as the adjustment continues, the effect becomes more obvious, indicating that the supercapacitor loss increases significantly after the adjustment; and after increasing the adjustment scale, the supercapacitor loss factor increases further, and the loss factor in some areas shows peaks, and the adjustment effect becomes more obvious.

[0145] S4, introduce equal cycle life constraints, dynamically adjust the critical entropy value, and change the reconstruction rules in real time, thereby changing the high and low frequency components. The specific steps are as follows:

[0146] S41, Figure 5 Adjust the schematic diagram for reconstruction, such as Figure 5 The capacity loss of the lithium battery and the supercapacitor obtained in step S3 is used to determine how to adjust the critical entropy value through equal cycle life constraints, and act on the power allocation of the next cycle, thereby realizing rapid adjustment of power allocation.

[0147] S42. For the convenience of comparison, the capacity loss of lithium battery and supercapacitor is ratioed and recorded as R L / S , if R L / S ≠1, indicating that there is a difference in the loss between the two, and the critical entropy value S needs to be adjusted ek+1 =S ek +ΔS, where S ek represents the critical entropy value of the kth period, S ek+1 represents the critical entropy value of the k+1th cycle, ΔS represents the critical entropy value adjustment value; otherwise, the critical entropy value remains unchanged.

[0148] S43, critical entropy value S e It is the basis for distinguishing high-frequency and low-frequency components. The entropy of several components decomposed in each cycle is calculated one by one to obtain the mean entropy of each component. Then the mean entropy of the cycle is calculated and used as the initial critical entropy value. The formula is as follows:

[0149]

[0150] In the formula, n is the number of components in a single cycle, N is the number of cycles, and the initial critical entropy value S is calculated by substituting the data into e1 =2.9943.

[0151] The adjustment of the critical entropy value determines the trend of change after adjustment and also determines the magnitude of the adjustment. If the loss ratio of the lithium battery and supercapacitor exceeds the set threshold, that is, R L / S≥1, the critical entropy value is reduced and the adjustment value ΔS<0, which will make the supercapacitor bear more power components, thereby sacrificing part of the effective capacity of the supercapacitor, reducing the loss of lithium batteries, and making the cycle life of hybrid energy storage tend to be consistent; otherwise, the critical entropy value is increased and the adjustment value ΔS>0.

[0152] According to the calculated entropy difference of the components in each cycle and the adjusted critical entropy value should not be less than zero, the range of the critical entropy adjustment value can be determined:

[0153] S min ≤|ΔS|≤S max (19)

[0154] ΔS+S e1 ≥0 (20),

[0155] In the formula, S max It represents the maximum value of all IMF component entropy values, S min represents the minimum value of all IMF component entropy calculations, where S max =4.0257, S min =0.3977, and there is formula (20), from which the adjustment value range is obtained:

[0156] ΔS∈(-2.9943,-0.3977)∪(0.3977,4.0257).

[0157] In order to obtain experimental comparison of three different adjustment scales, test verification of three different adjustment scales of ΔS=±0.9, ΔS=±1.8, and ΔS=±2.7 were set.

[0158] Under the adjustment of three scales, the real-time capacity loss data of lithium batteries and supercapacitors at different nodes are obtained as shown in Table 1:

[0159] Table 1. Real-time capacity loss data of lithium batteries and supercapacitors at different nodes under three different adjustment scales

[0160] Adjustment scale / serial number 1 2 3 4 5 6 7 8 <![CDATA[LB - Unadjusted (×10 -7 )]]> 718.1024 1414.1788 2131.2638 2853.9031 3585.4973 4321.2584 5076.9588 5826.6787 <![CDATA[LB-±0.9(×10 -7 )]]> 716.9309 1413.0220 2129.8454 2852.2984 3583.8092 4319.3686 5074.9200 5824.4748 <![CDATA[LB-±1.8(×10 -7 )]]> 714.4306 1410.3122 2125.3531 2847.6208 3579.0973 4312.7314 5068.2470 5817.7489 <![CDATA[LB-±2.7(×10 -7 )]]> 713.2246 1409.1014 2122.3013 2843.7907 3574.8848 4306.1459 5061.6615 5811.1633 <![CDATA[SC - Unadjusted (×10 -7 )]]> 647.5351 1295.6249 1943.7782 2591.3131 3238.7002 3887.3837 4535.0937 5184.5240 <![CDATA[SC-±0.9(×10 -7 )]]> 650.5004 1297.8410 1945.9209 2591.9085 3238.7581 3887.9807 4537.8740 5188.3150 <![CDATA[SC-±1.8(×10 -7 )]]> 661.0796 1307.8131 1971.0650 2614.0578 3257.9869 3914.4319 4564.1278 5214.7862 <![CDATA[SC-±2.7(×10 -7 )]]> 687.7306 1334.4969 2009.3861 2656.9439 3303.4091 3992.9097 4642.6057 5293.2646 Loss Ratio - Unadjusted 1.1090 1.0915 1.0965 1.1013 1.1071 1.1116 1.1195 1.1239 Loss ratio - ±0.9 1.1021 1.0887 1.0945 1.1005 1.1065 1.1110 1.1183 1.1226 Loss ratio - ±1.8 1.0807 1.0784 1.0783 1.0893 1.0986 1.1018 1.1105 1.1156 Loss ratio - ±2.7 1.0371 1.0559 1.0562 1.0703 1.0822 1.0784 1.0903 1.0978

[0161] By observing Table 1, it can be seen that at all sampling points, the loss of lithium batteries decreased after the adjustment was added, the loss of supercapacitors increased after the adjustment was added, and the loss ratio was also approaching 1, which met expectations, that is, sacrificing the capacity of supercapacitors and reducing the loss of lithium batteries. The difference in life span between the two is reduced, which verifies the effectiveness of the proposed method.

[0162] Compare the real-time capacity loss of lithium batteries and supercapacitors, such as Fig.13As shown in the figure, due to the different discharge properties of lithium batteries and supercapacitors, and their different roles in the energy storage system, there is a difference in the capacity loss of the two in the same working range, and the difference will gradually increase as the tractor works. At this time, it is necessary to add an adjustment reconstruction algorithm to reduce this difference.

[0163] Since the difference before and after adjustment is small, the area with obvious features is enlarged for further analysis, and 14(ag) is obtained, where the red curve represents the cumulative change of the loss factor before adjustment, the blue represents the loss change curve when ΔS = ±0.9, the green corresponds to ΔP = ±1.8, and the black corresponds to ΔS = ±2.7.

[0164] observe Figure 14-a For lithium batteries, the red curve is at the highest point, indicating that the lithium battery loss is high when no adjustment is added. After adding the reconstruction algorithm, the lithium battery loss gradually decreases, and as the adjustment scale increases, the reduction becomes more obvious; Observation Figure 14-b ,For the supercapacitor, on the contrary, the red curve is at the lowest point, indicating that the ,supercapacitor loss is low when no adjustment is added. After ,the reconstruction algorithm is added, the loss of the supercapacitor gradually increases, ,and with the increase of the adjustment scale, the increase is greater, which also meets ,the expectations of the reconstruction adjustment power allocation method. Figure 14-c to Figure 14-g Taken from different adjustment cycles, it can be observed that the blue and green curves are closer on the ordinate than the red curve, indicating that after adding the reconstruction adjustment power allocation, the capacity loss difference between lithium batteries and supercapacitors is reduced.

[0165] When the adjustment scale increases to ΔS = ±2.7, corresponding to the black curve, Figure 14-c It can be observed that before adjustment cycle 31, the loss of lithium batteries is always higher than that of supercapacitors. However, due to the reconstruction allocation of the adjustment algorithm, after the 31st adjustment cycle, the capacity loss curves of lithium batteries and supercapacitors intersect, that is, at this moment, the capacity loss of the two is consistent, and then the loss of supercapacitors exceeds that of lithium batteries and continues to rise, but the rising trend relative to lithium batteries gradually disappears after adjustment cycle 32. At this time, the adjustment reconstruction algorithm allocates more components to lithium batteries, hoping that the two will be consistent again; from Figure 14-d It can be seen that when the cycle period reaches 46, the black loss curve of the supercapacitor overlaps with the lithium battery loss curve again during the decline, indicating that under the effect of the adjustment and reconstruction algorithm, the real-time capacity loss of the two has reached consistency again; after that, the adjustment and reconstruction algorithm continues to work, and the adjusted lithium battery and supercapacitor loss curves (blue, green, black) are getting closer to each other than the unadjusted ones (red). Figure 14-e , Figure 14-f as well as Figure 14-g This also verifies the effectiveness of the adjustment algorithm.

[0166] Combined with the above loss curves, it can be seen that in the initial stage of operation, when the capacity loss of lithium batteries and supercapacitors is small, with the intervention of the adjustment algorithm, the capacity loss of the two is consistent more than once, but as the energy storage continues to work, the capacity loss of lithium batteries and supercapacitors continues to increase, and the loss difference also gradually increases. However, the degree of adjustment of the reconstruction and adjustment of power distribution is limited. Therefore, in the second half of the energy storage work, the capacity loss difference between the two can only be reduced; after completing a working cycle, the capacity loss and loss difference of lithium batteries and supercapacitors are as shown in Table 2:

[0167] Table 2. Capacity loss and loss difference of lithium battery and supercapacitor

[0168] Parameters / scaling Unadjusted ΔS=±0.9 ΔS=±1.8 ΔS=±2.7 <![CDATA[Lithium battery capacity loss (×10 -7 )]]> 6228.9195 6226.6193 6219.8347 6213.2036 <![CDATA[Supercapacitor capacitance loss (×10 -7 )]]> 5534.7888 5538.9728 5566.4113 5645.0660 Loss ratio 1.1254 1.1241 1.1174 1.1006 <![CDATA[Loss difference (×10 -7 )]]> 694.1307 687.6465 653.4234 568.1376 Adjustment effect % \ 0.9342 5.8645 18.1512

[0169] It can be seen from Table 2 that when the reconstruction adjustment algorithm is not added, after a complete working cycle, the capacity loss of the supercapacitor and the lithium battery will produce a difference of 694.1307 units. When the adjustment scale ΔS = ±2.7, the loss difference between the two is reduced to 568.1376 units, a reduction of 18.1512%, and the adjustment effect is more obvious.

[0170] The online EMD power distribution method for the hybrid energy storage system of the present invention dynamically adjusts the critical entropy value, changes the reconstruction rules, adjusts the capacity loss in real time, and reduces the life difference of the hybrid energy storage system through the online EMD and entropy calculation combined algorithm, thereby avoiding the problem that the overall performance of the hybrid energy storage system is affected by the large difference in cycle life, and even accelerates the aging speed of the energy storage equipment and increases the operation and maintenance cost.

[0171] The online EMD power distribution method for the hybrid energy storage system of the present invention realizes reasonable distribution of the power of the hybrid energy storage system by adjusting the capacity loss in real time, reduces the life span difference of the hybrid energy storage system, has little performance degradation, slow capacity loss, low operating cost, good system stability, and improves the use effect of the system.

[0172] The present invention is applicable to energy optimization management in the fields of electric vehicles and microgrids, and improves the performance and service life of hybrid energy storage.

[0173] The above is an exemplary description of the invention. Obviously, the specific implementation of the present invention is not limited to the above-mentioned method. As long as such non-substantial improvements are made by adopting the method concept and technical solution of the present invention, or the concept and technical solution of the invention are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.

Claims

1. A hybrid energy storage system online EMD power allocation method, characterized in that: The following steps are involved: S1. Import historical power data, extract average power and peak power, and use energy and power as constraints to configure the capacity of the hybrid energy storage system; S2, setting the sliding window and adjusting the step size to realize the online EMD decomposition of the power data, and using the entropy criterion as a basis to reconstruct the signal components; S3. Based on the hybrid energy storage loss quantification model, the capacity loss factors of lithium batteries and supercapacitors are calculated, and the capacity configuration and reconstructed signal components of the hybrid energy storage system are combined to obtain the energy storage capacity loss; S4, introduce equal cycle life constraints, dynamically adjust the critical entropy value, and change the reconstruction rules in real time, thereby changing the high and low frequency components.

2. The online EMD power allocation method for a hybrid energy storage system according to claim 1, characterized in that: In step S1, the capacity of the hybrid energy storage system is configured, and the specific steps are as follows: S11. Extracting average power P from historical power data average , the data below the average power is used as the basis for estimating the capacity configuration of lithium batteries, and the capacity of lithium batteries is configured with power and energy as constraints; S12. Extract peak power P from historical power data peak The difference between the peak power and the average power is used as the basis for estimating the supercapacitor capacity configuration, and the capacity of the supercapacitor is configured with power and energy as constraints.

3. The online EMD power allocation method for a hybrid energy storage system according to claim 2, characterized in that: In step S11, the capacity of the lithium battery is configured, and the specific process is as follows: The power constraint is that the rated power of the battery pack is not less than the average power, then: P LB ≥P average (1), Where P LB is the rated power of the battery, P LB There are the following relations: In the formula, E LB is the rated energy of the battery, Q LB is the rated capacity of the battery, U LB is the rated voltage of the battery; T is the entire working interval time, in hours; Combining the above formula, we can get the lithium battery capacity configuration range: E LB ≥P average ·T (3), For lithium batteries, energy constraints control the battery's SOC fluctuation within a certain range to avoid overcharge and overdischarge, then: In the formula, SOC LBmax Indicates the upper limit of lithium battery SOC, SOC LBmin Indicates the lower limit of lithium battery SOC; E l is the energy consumed by the lithium battery, which is estimated from the average power in the original power data. The calculation formula is as follows: Where P l is the data with lower power than the average power in the original data, n is the number of data with higher power than the average power in the original data; t is the data sampling time in seconds; Combining the above formulas (4) and (5), the lithium battery capacity configuration range formula is as follows: Combining equations (3) and (6), we take the intersection of the two capacity configuration ranges to obtain the final lithium battery configuration capacity.

4. The online EMD power allocation method for a hybrid energy storage system according to claim 3, characterized in that: In step S12, the capacity of the supercapacitor is configured, and the specific process is as follows: The power constraint is that the rated power of the supercapacitor bank is not less than the peak power, so: P SC ≥P peak -P average (7), Where P SC is the rated power of the supercapacitor bank; From the capacitance energy relation Available supercapacitor capacity configuration range: In the formula, C SC Indicates the rated capacity of the supercapacitor, U SC is the rated voltage of the supercapacitor bank; For supercapacitors, voltage constraints control the voltage fluctuation of supercapacitors within a certain range, and control the SOC of supercapacitors to change within a specified range, then: In the formula, E SC Indicates the rated energy of the supercapacitor, E h It is the power of the original data that exceeds the average power, and its formula is as follows: AND h =And power -AND l (10), In the formula, E power is the sum of the original power data; Combining equations (9), (10) and the voltage fluctuation limit r%, the capacitor energy relationship is Get the supercapacitor capacity configuration range: Combining equations (8) and (11), we take the intersection of the two capacity configuration ranges to obtain the final supercapacitor configuration capacity.

5. The online EMD power allocation method for a hybrid energy storage system according to claim 4, characterized in that: In step S2, the online EMD decomposition is performed, and the signal components are reconstructed using entropy calculation as a basis. The specific process is as follows: S21. In order to ensure the continuity of EMD decomposition, the power data is treated as "one in and one out" to obtain several cycles. Considering the constraints of EMD decomposition, sufficient power data is required, so the window size is set to G, and when the cycle reaches the Gth cycle, that is, the window is filled for the first time, the EMD decomposition is started for it and the subsequent cycles, and it does not participate in the decomposition stage. The system is powered by the battery alone. S22. In order to reasonably reconstruct the high and low frequency components and adjust the reconstruction rules conveniently, the information entropy is used to calculate the fluctuation amplitude of the power data, so as to distinguish the high and low frequency components and reconstruct the power. The formula for entropy value calculation is as follows: In the formula, p(x i ) represents a random number x i The probability of occurrence, X represents a certain imf component, S represents the entropy value of this component, and x represents a random data point on this component; Formula (12) is used to calculate the entropy of the i imf components obtained by EMD decomposition of each cycle, and the entropy value is recorded as S i , and compare it with the critical entropy value S e Make a judgment, if S i ≥S e , that is, if the entropy value exceeds the set value, the i-th component is high frequency, otherwise it is low frequency.

6. The online EMD power allocation method for a hybrid energy storage system according to claim 5, characterized in that: The specific steps of the S3 step are as follows: S31. Introduce crack propagation theory, and derive the material fracture propagation rate formula based on the material fracture time formula; At the same time, consider the influence of lithium battery charge state fluctuation, throughput cycle and temperature, and obtain the lithium battery life loss coefficient. Finally, after superposition, the lithium battery life loss factor is obtained as follows: λ LB =∑L1·L2·L3 (13), In the formula, λ LB Indicates the capacity loss factor of lithium battery, L1 indicates loss factor 1, L2 indicates loss factor 2, and L3 indicates loss factor 3; K co is the throughput coefficient; N LB is the equivalent throughput cycle, which indicates the ratio of the actual charge and discharge capacity to the nominal capacity in a time period; S dev_LB is the normalized deviation of the state of charge, which reflects the degree of deviation between the actual value and the average value; K ex is the discharge depth index; T LB , T ref_LB are the real-time temperature of the lithium battery and the reference temperature of rated operation; T LBa 、T refa_LB are the absolute temperatures of the real-time temperature of the lithium battery and the reference temperature of the rated working temperature; T f represents the temperature coefficient in the thermal aging model; t cycle is the time of one cycle, t life Indicates the calendar life when the capacity decays to 80% under rated conditions, in seconds; K soc is the average state of charge coefficient; S avg_LB is the average state of charge. The low-frequency component obtained by decomposition and reconstruction in step S2 is given to the lithium battery, and combined with the lithium battery configuration capacity obtained in step S1, it is substituted into the lithium battery life loss factor formula (13)-(16) to obtain the loss factor of each cycle, and then superimposed one by one to obtain the real-time capacity loss of the lithium battery; S32. The Arrhenius equation is introduced. According to its generalized formula, and considering the influence of voltage and temperature on the life of supercapacitor, the functional relationship between loss rate and voltage and temperature is obtained, thereby obtaining the life loss factor of supercapacitor: In the formula, λ SC represents the supercapacitor capacity loss factor; l0 is the rated calendar life of the supercapacitor in seconds; E a_SC Indicates the real-time energy of the supercapacitor; C is the rated capacity of the supercapacitor; V0 is the rated voltage of the supercapacitor; T ref_SC is the initial temperature of the supercapacitor; R th_SC P is the relationship coefficient between operating temperature and charge and discharge power; SC Indicates the real-time power of the supercapacitor when it is working; τ is the cycle length in seconds; T0 is the rated temperature of the supercapacitor; The high-frequency component obtained by decomposition and reconstruction in step S2 is given to the lithium battery. Combined with the supercapacitor configuration capacity obtained in step S1, it is substituted into the supercapacitor life loss factor formula (17) to obtain the loss factor of each cycle, and then they are superimposed one by one to obtain the real-time capacity loss of the supercapacitor.

7. The online EMD power allocation method for a hybrid energy storage system according to claim 6, characterized in that: The specific steps of the S4 step are as follows: S41, the capacity loss of the lithium battery and the supercapacitor obtained in step S3 is used to determine how to adjust the critical entropy value through equal cycle life constraints, and acts on the power allocation of the next cycle, so as to achieve rapid adjustment of power allocation; S42. Ratio the capacity loss of lithium battery and supercapacitor to that of supercapacitor, denoted as R L / S , if R L / S ≠1, indicating that there is a difference in the loss between the two, and the critical entropy value S needs to be adjusted ek+1 =S ek +ΔS, where S ek represents the critical entropy value of the kth period, S ek+1 represents the critical entropy value of the k+1th cycle, ΔS represents the critical entropy value adjustment value; otherwise, the critical entropy value remains unchanged; S43, critical entropy value S e It is the basis for distinguishing high-frequency and low-frequency components. The entropy of several components decomposed in each cycle is calculated one by one to obtain the mean entropy of each component. Then the mean entropy of the cycle is calculated and used as the initial critical entropy value. The formula is as follows: Where n is the number of components in a single cycle, and N is the number of cycles; If the loss ratio of lithium battery and supercapacitor exceeds the set threshold, that is, R L / S ≥1, the critical entropy value is reduced, and the adjustment value ΔS<0, part of the power borne by the lithium battery is given to the supercapacitor to reduce the capacity loss of the lithium battery; otherwise, the critical entropy value is increased, and the adjustment value ΔS>0; the range of the critical entropy adjustment value is: S min ≤ΔS≤S max (19), ΔS+S e1 ≥0 (20), In the formula, S max It represents the maximum value of all IMF component entropy values, S min Indicates the minimum value of all IMF component entropy calculations.

Images