CN113659558A - Control method of direct-current micro-grid hybrid energy storage system based on multi-step model prediction - Google Patents

Control method of direct-current micro-grid hybrid energy storage system based on multi-step model prediction Download PDF

Info

Publication number
CN113659558A
CN113659558A CN202110754949.0A CN202110754949A CN113659558A CN 113659558 A CN113659558 A CN 113659558A CN 202110754949 A CN202110754949 A CN 202110754949A CN 113659558 A CN113659558 A CN 113659558A
Authority
CN
China
Prior art keywords
power
energy storage
value
storage battery
voltage
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110754949.0A
Other languages
Chinese (zh)
Inventor
张晨
李正明
陈剑月
王啸尘
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangsu University
Original Assignee
Jiangsu University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jiangsu University filed Critical Jiangsu University
Priority to CN202110754949.0A priority Critical patent/CN113659558A/en
Publication of CN113659558A publication Critical patent/CN113659558A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J1/00—Circuit arrangements for dc mains or dc distribution networks
    • H02J1/14—Balancing the load in a network
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J1/00—Circuit arrangements for dc mains or dc distribution networks
    • H02J1/02—Arrangements for reducing harmonics or ripples
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J1/00—Circuit arrangements for dc mains or dc distribution networks
    • H02J1/10—Parallel operation of dc sources
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J7/00—Circuit arrangements for charging or depolarising batteries or for supplying loads from batteries
    • H02J7/0013—Circuit arrangements for charging or depolarising batteries or for supplying loads from batteries acting upon several batteries simultaneously or sequentially
    • H02J7/0014—Circuits for equalisation of charge between batteries
    • H02J7/0016—Circuits for equalisation of charge between batteries using shunting, discharge or bypass circuits
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J7/00—Circuit arrangements for charging or depolarising batteries or for supplying loads from batteries
    • H02J7/34—Parallel operation in networks using both storage and other dc sources, e.g. providing buffering
    • H02J7/345—Parallel operation in networks using both storage and other dc sources, e.g. providing buffering using capacitors as storage or buffering devices

Abstract

The invention discloses a control method of a direct-current micro-grid hybrid energy storage system based on multi-step model prediction, and belongs to the field of direct-current micro-grid hybrid energy storage control. On the basis of double closed-loop control of outer loop voltage control and inner loop power control, the method adopts voltage control based on droop characteristic at any time t, and samples the real-time voltage U of the DC busdc‑stoAnd a reference value U of the voltage of the direct current bus terminal* dc‑stoComparing and calculating the droop coefficient to obtain the reference power value P of the input power at the DC side*As an inner ringInput reference values for the multi-step model predictive control; inner loop according to reference power P*Low power density of P* batHigh power density P for battery module allocation* SCAnd distributing the energy to a super capacitor, respectively establishing a prediction model and a target function, obtaining the optimal combination of the switching states of the converter through the weighted comparison of error coefficients, combining the optimal combination with the SOC value of a storage battery pack in the hybrid energy storage system, reasonably distributing the energy, balancing the respective SOC values and stabilizing the voltage of the direct-current bus.

Description

Control method of direct-current micro-grid hybrid energy storage system based on multi-step model prediction
Technical Field
The invention relates to a direct current micro-grid hybrid energy storage control technology, in particular to a direct current micro-grid hybrid energy storage system control method based on multi-step model prediction.
Background
The hybrid energy storage device composed of the storage battery and the super capacitor has the advantages of high energy density and high power density, and can effectively maintain the stability of bus voltage and stabilize the power fluctuation of the bus when the voltage of the direct-current micro-grid bus fluctuates. Hybrid energy storage system control strategies typically employ control algorithms based on model predictive control. Compared with the traditional PI control, the model predictive control has the advantages of simple modeling, no need of PI parameter regulation, no pulse width regulation, relevant parameter regulation and the like, and the dynamic response speed is greatly superior to that of the traditional PI control method by the active predictive mode. The model predictive control has many advantages, and the difficulty of the control is mainly concentrated in that the model predictive control is based on the state of the controlled quantity at the time k, the state quantity at the time k +1 is obtained by predicting the model, then all effective output state combinations at the time k +1 are traversed, and the output state corresponding to the minimum value of the objective function is selected as the output of the system. However, because the model prediction only considers the optimal output state in one control cycle, the optimal solution of the prediction model in two or more control cycles is not considered, and the optimal information contained in other output states is also ignored, the algorithm has conservatism and is easy to fall into the local optimal solution, so that the model prediction algorithm can cause the oscillation of the control system to be aggravated and even cause the system to diverge under the condition that the system has larger disturbance or modeling error. In order to reduce the instability of a system caused by possible errors of a model prediction algorithm, the rolling optimization link of the model prediction control algorithm is improved, a multi-step prediction model prediction control algorithm is designed, the prediction values of the next two moments are calculated, namely the model is subjected to prediction of two control periods, and the system output state at the k +1 moment corresponding to the optimal output state of the last control period acts on the current k moment. The factors considered by the method are not comprehensive, although the stability of system voltage fluctuation can be improved to a certain extent, the optimal prediction path is determined only from the optimal output state of the last control period, which is considered to be insufficient, the optimal switch control combination acting on the hybrid energy storage system cannot be determined only from the result, and the instability of the system can be caused due to the lack of an effective judgment mechanism for the comprehensive error of the optimal solution and the suboptimal solution in each step; and the model predictive control aiming at the hybrid energy storage system lacks an effective control algorithm for balancing the charge states of the storage battery packs, so that the charge states of the storage battery packs are unbalanced easily, the service life of the storage battery is influenced, the charge states of part of the storage battery packs are too low and quit in advance, and the phenomena of system power distribution and circulation among the storage battery packs are influenced.
Disclosure of Invention
The invention aims to control a DC micro-grid hybrid energy storage system by a multi-step model prediction method added with an error weight coefficient, balance SOC among storage battery packs, prolong the service life of the storage battery packs and enable the voltage of a DC micro-grid to be more stable. The method comprises the steps of utilizing a storage battery pack and a super capacitor to form a hybrid energy storage device, having the characteristics of high response speed and long working time for the power fluctuation of a direct current micro-grid bus, taking a hybrid energy storage system of the direct current micro-grid as a controlled object, and building a direct current micro-grid hybrid energy storage system structure model formed by the storage battery pack and the super capacitor, wherein a photovoltaic power generation module is connected into the direct current micro-grid through a DC/DC converter, the DC/DC converter is defaulted to work in an MPPT mode, the storage battery module is formed by connecting two storage battery packs in parallel, and the super capacitor is controlled by a hybrid energy storage control system which is built in a multi-step model predictive control mode and is respectively connected into the direct current micro-grid through a bidirectional DC/DC converter.
The technical scheme adopted by the invention is as follows: the control method of the direct current micro-grid hybrid energy storage system based on multi-step model prediction comprises the following steps: the inductance resistance value is considered, so that the accuracy of the prediction model is higher, and the hybrid energy storage system consists of a storage battery module and a super capacitor, wherein UstoIs the voltage value of the energy storage system side, istoThe current output by the energy storage unit flows through the inductors L and RLIs an inductive resistance, and is characterized in that,due to RLThe influence on the output voltage is not negligible, and the inductance resistance R is used for establishing an equivalent circuitLAnd UdAlso taken into account, UdIs the voltage drop of the diode in forward conduction, Udc-stoThe voltage value of the direct current bus side is obtained; the control method specifically comprises the following steps:
step 1, on the basis of double closed-loop control of outer loop voltage control and inner loop power control, the outer loop adopts voltage control based on droop characteristics, the voltage value of a direct current bus is sampled in real time, and the reference power value P of the input power at the direct current side is finally obtained through calculation of a droop coefficient*；
Step 2, inputting the reference power P of the DC side*The low-power-density part is used as the reference power P of the storage battery pack by using a low-pass filter and combining the charge and discharge characteristics of a storage battery and a super capacitor as the input reference value of the inner-loop multi-step model predictive control* batThe part which is distributed to the storage battery module and has high power density is used as reference power P of the super capacitor* SCIs distributed to the super capacitor;
step 3, respectively establishing a prediction model and a target function of the storage battery module and the super capacitor module, discretizing the models by adopting an Euler method, iterating an equation to obtain a multi-step prediction model, and then performing weighted comparison on error coefficients of the optimization path of each control period predicted by the multi-step model to obtain an optimal combination of the switching states of the converter and apply the optimal combination to the hybrid energy storage system so as to control the output of the energy storage device, thereby achieving the functions of realizing flexible energy management and restraining power fluctuation of a power generation unit or a load;
and 4, combining the storage battery pack in the hybrid energy storage system with the SOC value thereof, and setting a reasonable power deviation gain coefficient K0And differentially distributing the power predicted by the model to different storage battery packs to eliminate the unbalanced degree of the SOC of the storage battery packs, thereby balancing the SOC values of the storage battery packs and stabilizing the DC bus voltage.
Further, the calculation amount and the system steady-state problem are comprehensively considered, and the prediction step size of the multi-step prediction model is selected to be 2.
Further, the specific process of step 1 is as follows:
voltage U through real-time sampling direct current busdc-stoThe reference value U of the voltage of the direct current bus terminal is used* dc-stoComparing, and calculating droop coefficient to obtain reference current value I of DC input currentdcThen, the reference power value P of the input power at the DC side is obtained*And using the power value as an input reference value for prediction by the inner loop model.
Further, the characteristics of the storage battery pack and the super capacitor are fully combined, the power is reasonably distributed to achieve the optimal system scheduling, and in the step 2:
passing P through a low-pass filter*The method comprises the steps of dividing the hybrid energy storage device into a high-frequency part and a low-frequency part, using the low-frequency part as a power reference value of a storage battery pack and using the high-frequency part as a power reference value of a super capacitor according to self characteristics of a storage battery and the super capacitor, analyzing whether the hybrid energy storage device sends power or absorbs power according to bus power fluctuation of a direct-current micro-grid, and dividing the hybrid energy storage device into a Boost prediction model and a Buck prediction model.
Further, the prediction error of each step is taken into consideration, so as to avoid the occurrence of the phenomenon that the system oscillates or diverges due to the fact that the prediction model is determined only by the optimal solution of the last step, in the step 3:
acquiring bus voltage at the moment k of the direct-current micro-grid, and obtaining power values at the moment k +1 and the moment k +2 through calculation of a prediction model by taking bus current and voltage values of an energy storage device as input; comparing the power value at the moment k +1 with the power reference value, and obtaining the optimal solution x at the moment k +1 by a traversal algorithmimin1(k +1) and suboptimal solution ximin2(k +1), around the optimal solution x obtained in the first stepimin1(k +1) and suboptimal solution ximin2(k +1) respectively carrying out model prediction of the second step to obtain each predicted value x of the two solutions in the second stepij(k +2) and corresponding switch state combination Sij(k +1), j ═ 1, 2.., n; respectively calculating the optimal solution x of the first stepimin1(k +1) and suboptimal solution ximin2(k +1) the predicted value and the expected value obtained in the second step are incorrectTwo solutions x with minimum differenceijmin1(k +2) and xijmin2(k + 2); and then carrying out weighted comparison on error coefficients of the optimizing paths of each control period predicted by the multi-step model to obtain the optimal combination of the switching states of the converters, and applying the optimal combination to the hybrid energy storage system.
Further, a power gain deviation coefficient is set to eliminate the problem of SOC imbalance among the storage battery packs, and in step 4: calculating an average SOC value according to the SOC value of the initial state of the storage battery pack and the number of the storage battery packs, calculating the difference between the average SOC value and the average SOC value to obtain the unbalance degree delta SOC of the SOC value of the storage battery pack, and setting a reasonable power deviation gain coefficient K according to the type of the storage battery and the requirement of adjusting time0The power deviation is adjusted, and the SOC values of the battery packs are balanced by the unbalanced energy distribution.
The invention has the beneficial effects that:
1. considering that the randomness of input and output changes of the direct-current micro-grid system (influence factors comprise changes of loads and distributed energy output) can affect the stability of the direct-current bus voltage, and a multi-step model prediction algorithm with variable weight coefficients is constructed, so that the system breaks through the conservative property of single-step prediction, jumps out of a local optimal solution, avoids the phenomenon of oscillation and even divergence of the system, and has stronger stability and robustness.
2. The method is different from the traditional multi-step model prediction, aims at the problem of judging the optimal solution path of the multi-step prediction, considers more comprehensive prediction results on the premise of keeping the total prediction step length unchanged, and introduces the prediction results into a judgment criterion according to a variable weight coefficient theory when the prediction of the intermediate control period has large fluctuation, so that the judgment of the system prediction results is more comprehensive and comprehensive, the condition that the intermediate process possibly has large errors is ignored when only the final result is seen, and the requirement of the direct-current bus voltage control stability is met.
3. In order to avoid the phenomenon of unbalance of the SOC among the storage battery packs when the optimal switch combination of multi-step model prediction only acts on a system, the SOC evaluation is carried out by utilizing a charge accumulation method, and a power deviation gain coefficient is set, so that the storage battery pack with more energy shares more power, the SOC of the storage battery packs is balanced, the circulation phenomenon among the storage battery packs is effectively avoided, the service life of the storage battery is prolonged, and the working stability of the storage battery packs is improved.
Drawings
Fig. 1 is a topological structure of a dc microgrid hybrid energy storage control system adopted by the present invention;
FIG. 2 is a schematic diagram of the hybrid energy storage system of the present invention operating in a BOOST mode;
FIG. 3 shows T when the hybrid energy storage system of the present invention is operating in the BOOST mode2A schematic diagram of the inductor L during discharge when the inductor L is turned off;
FIG. 4 shows T when the hybrid energy storage system of the present invention is operating in the BOOST mode2Conducting and discharging the inductor L by the energy storage system;
FIG. 5 shows T when the hybrid energy storage system of the present invention is operating in the BOOST mode2An equivalent model diagram when the device is turned off;
FIG. 6 is an algorithmic schematic of multi-step model prediction in accordance with the present invention;
fig. 7 is a general control block diagram of the dc microgrid hybrid energy storage system established by the present invention;
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
Step 1, on the basis of double closed-loop control of outer loop voltage control and inner loop power control, the outer loop adopts voltage control based on droop characteristics, the voltage value of a direct current bus is sampled in real time, and the reference power value P of the input power at the direct current side is finally obtained through calculation of a droop coefficient*；
Step 2, inputting the reference power P of the DC side*The low-power-density part is used as the reference power P of the storage battery pack by using a low-pass filter and combining the charge and discharge characteristics of a storage battery and a super capacitor as the input reference value of the inner-loop multi-step model predictive control* batThe part which is distributed to the storage battery module and has high power density is used as reference power P of the super capacitor* SCIs distributed to the super capacitor;
according to the method, a mixed energy storage system of a direct-current micro-grid is taken as a controlled object, as shown in figure 1, a direct-current micro-grid mixed energy storage system structure model consisting of a storage battery and a super capacitor is built, wherein a photovoltaic power generation module is connected into the direct-current micro-grid through a DC/DC converter and works in an MPPT mode by default, the storage battery module is formed by connecting two groups of storage battery packs in parallel, and the super capacitor is controlled by a mixed energy storage control system which is built in a multi-step model predictive control and is respectively connected into the direct-current micro-grid through a bidirectional DC/DC converter. As the prediction times of the multi-step model prediction increase in a geometric progression along with the increase of the prediction period, a large operation burden is caused on a processor, and a system response delay phenomenon can be caused. Considering the problems of calculation amount and system steady state, the invention selects the prediction step length to be 2.
The invention focuses on a hybrid energy storage control strategy. The strategy is based on double closed loop control of outer loop voltage control and inner loop power control. At any time t, the outer ring employs voltage control based on droop characteristics. Voltage U through real-time sampling direct current busdc-stoAnd the reference value U of the voltage of the direct current bus terminal* dc-stoBy comparison, by calculation of the droop coefficient, i.e. from the formula
Figure BDA0003144066240000051
Determining the droop coefficient, wherein UmaxIs the maximum voltage value, ImaxAt the maximum current value, ULIs ImaxThe corresponding voltage value. Finally formed by the formula
Figure BDA0003144066240000052
Calculating the reference power value P of the DC side input power*Where k is the droop coefficient, U* dc-stoIs a reference value of the voltage at the DC bus terminal, Udc-stoIs the dc bus terminal voltage. And taking the power value as an input reference value of the inner loop multi-step model predictive control. Inner loop according to reference power P*By the action of a low-pass filter according to the battery overloadCharge-discharge characteristics of the stage capacitor, low power density P* batHigh power density P for distribution to battery modules* SCAnd distributing the energy to the super capacitor, respectively establishing a prediction model and a target function, performing error coefficient weighted comparison on the optimizing paths predicted by the multi-step model to obtain the optimal combination of the switching states of the converter, applying the optimal combination to the hybrid energy storage system, and combining the SOC value of the storage battery pack in the hybrid energy storage system, reasonably distributing the energy, balancing the respective SOC value and stabilizing the direct-current bus voltage. 1
Taking the hybrid energy storage unit working in the Boost mode as an example, the schematic diagram is shown in fig. 2, and the current direction flows from left to right. When T is2In the off state, as shown in FIG. 3, due to the MOSFET transistor T2Disconnection, T2The branch at (D) corresponds to an open circuit, so that current flows through the freewheeling diode D1Terminal voltage U flowing to load2The capacitor C is a filter capacitor, and the energy stored in the inductor L is transferred to the power supply through a freewheeling diode D1Is released to U2. When T is2In the conducting state, the load end is short-circuited, as shown in FIG. 4, and the current passes through T2Reflux hybrid energy storage terminal voltage U1The whole circuit is composed of a hybrid energy storage end U1And charging the inductor.
The energy storage system consists of a storage battery module and a super capacitor, wherein UstoIs the voltage value of the energy storage system side, istoThe current output by the energy storage unit flows through the inductor L. RLThe inductance resistance has a large influence on the output voltage, and is not negligible for the accuracy of model building. Therefore, the invention sets the inductance resistance R when establishing the equivalent circuitLAnd UdAre also contemplated. U shapedIs the voltage drop of the diode in forward conduction, Udc-stoIs the voltage value on the side of the direct current bus.
And 3, respectively establishing a prediction model and a target function of the storage battery pack module and the super capacitor module, discretizing the models by adopting an Euler method, iterating an equation to obtain a multi-step prediction model, and then performing weighted comparison on error coefficients of the optimization path of each control period predicted by the multi-step model to obtain an optimal combination of the switching states of the converter and apply the optimal combination to the hybrid energy storage system so as to control the output of the energy storage device, thereby achieving the functions of realizing flexible energy management and restraining power fluctuation of the power generation unit or the load.
In MOSFET tube T2During the turn-off period, the equivalent circuit diagram is shown in FIG. 5, and the electric energy stored in the inductor L passes through the freewheeling diode D1Releasing the voltage to the direct current bus side, and obtaining a relation between the inductive current and the voltage of the direct current bus side according to kirchhoff's law:
Figure BDA0003144066240000061
Figure BDA0003144066240000062
discretizing and sorting the formulas (1) and (2) by an Euler method to obtain:
Figure BDA0003144066240000063
Figure BDA0003144066240000064
wherein: i is the current flowing out of the energy storage unit, TSFor a sampling period, RLIs an inductor resistor, C is a filter capacitor, Usto(k) The terminal voltages at two ends of the energy storage unit at the moment k are obtained; i.e. isto(k) Outputting current for the energy storage unit at the moment k; u shapedc-sto(k) The voltage of the direct current bus at the moment k; i.e. isto(k +1) is the output current of the energy storage unit at the moment of k +1, Udc-stoAnd (k +1) is the terminal voltage of two ends of the direct current bus side at the moment of k + 1.
Similarly, when the MOSFET T is turned off, the same applies2When conducting, the energy storage system passes through T2Charging inductor L, reading inductor current from k moment, direct current bus side voltage value and energy storageAnd discretizing the side voltage value by using an Euler method to obtain an inductance current value and a direct current bus side voltage value at the k +1 moment.
Figure BDA0003144066240000065
Figure BDA0003144066240000066
And further finishing the bidirectional DC/DC in a Boost mode to obtain:
Figure BDA0003144066240000067
Figure BDA0003144066240000068
wherein S represents T2S e 0,1, where 1 represents T2On, 0 represents T2And (6) turning off.
And (3) obtaining the inductance current value and the direct current bus side voltage value at the k +1 moment by establishing a prediction model according to the inductance current value and the direct current bus side voltage value read from the k moment. And multiplying the inductive current and the voltage at the side of the direct current bus, and calculating the power predicted value at the k +1 moment by the k moment as follows:
P(k+1|k)＝isto(k+1)*Udc-sto(k+1) (9)
through a feedback correction link, according to a formula:
P(k+1)＝P(k+1|k)+[P(k)-P(k|k-1)] (10)
and obtaining a predicted value P (k +1) of the power at the k +1 moment after feedback correction, and then performing iterative operation.
The bidirectional DC/DC converter is used for controlling the charging and discharging of the energy storage element, so that the energy is charged to the energy storage system from the direct-current microgrid and the energy is discharged to the direct-current microgrid in a bidirectional flow manner.
Therefore, the invention selects the difference value of the predicted power and the expected power value in each prediction period to be multiplied by the variable weight coefficient to form the objective function. Considering that single-step model prediction is easy to fall into a local optimal solution, the calculation amount of the system increases exponentially every more than one step of prediction, so that the hardware requirement of the system is too high, the operation time is too long, and the response delay phenomenon occurs. Therefore, the model prediction control of the DC micro-grid hybrid energy storage system adopts two-step model prediction, namely the predicted value is calculated to the time of k +2 from the switching state at the time of k, so that the local optimal solution is skipped, the exponential increase of the operand is avoided, the response speed of the hybrid energy storage system to the power fluctuation of the DC micro-grid is accelerated, and the hardware investment cost is reduced.
As shown in fig. 6. Knowing x (k), x is calculated from the prediction modeli(k +1) and Si(k) N, n is the combined number of all output states, and two values with the minimum error between the predicted value at the moment of k +1 and the expected value are selected as the optimal solution ximin1(k +1) and suboptimal solution ximin2(k +1), around the optimal solution x obtained in the first stepimin1(k +1) and suboptimal solution ximin2(k +1) respectively carrying out model prediction of the second step to obtain each predicted value x of the two solutions in the second stepij(k +2) and corresponding switch state combination Sij(k +1), j ═ 1, 2.., n; respectively calculating the optimal solution x of the first stepimin1(k +1) and suboptimal solution ximin2(k +1) two solutions x with the smallest error between the predicted value and the expected value obtained in the second stepijmin1(k +2) and xijmin2(k+2)。
The direct current micro-grid bus voltage pursues certain stability, the output voltage value of the direct current micro-grid bus voltage is allowed to fluctuate within a certain range, only the optimal solution of the final result is seen, a large prediction error possibly occurs in the middle prediction process is ignored, and the optimal solution of the last control period can cause the oscillation and even the divergence of the prediction result when the optimal solution acts on the system.
In order to avoid the situation that the error of the suboptimal solution of the first step is larger in the model prediction process, the prediction performance of the second step is better, the control state of the system is determined blindly according to the optimal solution of the second step and the error amount is increased, so the invention introduces the concept of variable weight coefficients, the optimal solution and the suboptimal solution of the first control period and two optimal solutions corresponding to the second control period derived from the optimal solution and the suboptimal solution are respectively solved according to the error values of the optimal solution and the expected value, the corresponding weight coefficients are respectively multiplied by the error values, the switch state combination at the k moment corresponding to the minimum value of the objective function is obtained, the comprehensive performance of the system is optimal, and the state value at the k moment is acted on the system. The specific formula is as follows:
let n kinds of models participate in the variable weight coefficient combination, wherein the weight coefficient of the ith prediction model at t time is Wi(t), the predicted value is recorded as
Figure BDA0003144066240000081
Corresponding prediction error is noted
Figure BDA0003144066240000082
Where i is 1, 2. The weight-varying coefficient of each model at time t can be obtained by the following formula:
Figure BDA0003144066240000083
setting an objective function J ═ min { W }1(t)e1(t)+W2(t)e11(t),W3(t)e2(t)+W4(t)e22(t)} (12)
Wherein e is1(t) is the difference between the optimal solution and the expected value of the first model predictive control cycle, W1(t) is its error weight coefficient, e11(t) predicting based on the optimal solution of the first model predictive control cycle, the difference, W, between the optimal solution and the desired value obtained in the second model predictive control cycle2(t) is the error weight coefficient corresponding thereto; in the same way, e2(t) is the difference between the second best solution and the expected value of the first model predictive control cycle, W3(t) is its error weight coefficient, e22(t) predicting based on the first model predictive control cycle suboptimal solutionMeasuring the difference between the optimal solution and the expected value, W, obtained in the second model predictive control cycle4And (t) is the corresponding error weight coefficient.
In addition, the invention is suitable for absorbing or releasing high-frequency power according to the characteristics that the charging and discharging speed of the super capacitor is high, the energy density of the storage battery is high, the stored energy is more, and the super capacitor is suitable for absorbing or releasing low-frequency power. The low-pass filter is used for realizing power distribution, the low-frequency part of power fluctuation is distributed to the storage battery, and the high-frequency part of the power fluctuation is distributed to the super capacitor. The super capacitor is mainly used for optimizing the charge-discharge operation of the storage battery module, and the voltage stability of the direct-current bus is maintained by quickly compensating the power shortage when the output power and the pulsating load power of the photovoltaic power generation module suddenly change; the storage battery module needs to meet the requirement of long-time large energy transmission, and how to reasonably distribute the energy transmitted by each storage battery pack port is very important, which is related to the service life of each storage battery pack, the SOC balance and the energy transmission efficiency of the whole system.
And 4, combining the storage battery pack in the hybrid energy storage system with the SOC value thereof, and setting a reasonable power deviation gain coefficient K0And differentially distributing the power predicted by the model to different storage battery packs to eliminate the unbalanced degree of the SOC of the storage battery packs, thereby balancing the SOC values of the storage battery packs and stabilizing the DC bus voltage.
The invention is based on the formula
Figure BDA0003144066240000084
Obtaining SOC value of each storage battery pack (13)
In the formula: ceIs the energy storage unit capacity; SOC is the initial residual capacity of the storage battery pack; i.e. ibattIs the output current of the battery pack.
Then, the average value of the SOC of each battery pack port is obtained, and since the storage battery module used in the invention is formed by connecting two storage battery packs in parallel, here:
Figure BDA0003144066240000091
defining Δ SOC as the unbalance degree of SOC of k ports of the storage battery pack:
Figure BDA0003144066240000092
the purpose of the energy flow balance control is to make the SOC of each storage battery pack tend to be consistent, and essentially, the target value P of energy transmitted to each storage battery pack endbat(k)_refAnd a current reference value Ibat(k)_refIs reasonably given. The invention firstly equally divides the low-frequency power value obtained by the low-pass filter at each storage battery pack port, and carries out power deviation adjustment according to the SOC unbalance degree of the storage battery pack on the basis of the average power value born by each storage battery pack port to obtain the following equation set:
Figure BDA0003144066240000093
in the formula: k0The power deviation gain coefficient is related to the SOC regulation speed and can be set according to the model of the storage battery and the regulation time requirement.
The present invention can be realized in light of the above. Other variations and modifications which may occur to those skilled in the art without departing from the spirit and scope of the invention are intended to be included within the scope of the invention.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an illustrative embodiment," "an example," "a specific example," or "some examples" or the like mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the invention have been shown and described, it will be understood by those of ordinary skill in the art that: various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims (6)

1. A control method of a direct current micro-grid hybrid energy storage system based on multi-step model prediction is characterized in that inductance resistance is considered, so that the accuracy of a prediction model is higher, the hybrid energy storage system is composed of a storage battery module and a super capacitor, and UstoIs the voltage value of the energy storage system side, istoThe current output by the energy storage unit flows through the inductors L and RLIs an inductive resistance, due to RLThe influence on the output voltage is not negligible, and the inductance resistance R is used for establishing an equivalent circuitLAnd UdAlso taken into account, UdIs the voltage drop of the diode in forward conduction, Udc-stoThe voltage value of the direct current bus side is obtained; the control method specifically comprises the following steps:
step 1, on the basis of double closed-loop control of outer loop voltage control and inner loop power control, the outer loop adopts voltage control based on droop characteristics, the voltage value of a direct current bus is sampled in real time, and the reference power value P of the input power at the direct current side is finally obtained through calculation of a droop coefficient*；
Step 2, inputting the reference power P of the DC side*The low-power-density part is used as the reference power P of the storage battery pack by using a low-pass filter and combining the charge and discharge characteristics of a storage battery and a super capacitor as the input reference value of the inner-loop multi-step model predictive control* batThe part which is distributed to the storage battery module and has high power density is used as reference power P of the super capacitor* SCIs distributed to the super capacitor;
step 3, respectively establishing a prediction model and a target function of the storage battery module and the super capacitor module, discretizing the models by adopting an Euler method, iterating an equation to obtain a multi-step prediction model, and then performing weighted comparison on error coefficients of the optimization path of each control period predicted by the multi-step model to obtain an optimal combination of the switching states of the converter and apply the optimal combination to the hybrid energy storage system so as to control the output of the energy storage device, thereby achieving the functions of realizing flexible energy management and restraining power fluctuation of a power generation unit or a load;
and 4, combining the storage battery pack in the hybrid energy storage system with the SOC value thereof, and setting a reasonable power deviation gain coefficient K0And differentially distributing the power predicted by the model to different storage battery packs to eliminate the unbalanced degree of the SOC of the storage battery packs, thereby balancing the SOC values of the storage battery packs and stabilizing the DC bus voltage.
2. The method for controlling the direct current microgrid hybrid energy storage system based on the multi-step model prediction is characterized in that the prediction step size of the multi-step prediction model is selected to be 2 in consideration of the problems of calculation amount and system steady state.
3. The method for controlling the direct current microgrid hybrid energy storage system based on the multi-step model prediction as claimed in claim 1, characterized in that the specific process of the step 1 is as follows:
voltage U through real-time sampling direct current busdc-stoThe reference value U of the voltage of the direct current bus terminal is used* dc-stoComparing, and calculating droop coefficient to obtain reference current value I of DC input currentdcThen, the reference power value P of the input power at the DC side is obtained*And using the power value as an input reference value for prediction by the inner loop model.
4. The method for controlling the direct current microgrid hybrid energy storage system based on the multi-step model prediction as claimed in claim 1, wherein in the step 2:
passing P through a low-pass filter*The method is divided into a high-frequency part and a low-frequency part, the low-frequency part is used as a power reference value of a storage battery pack according to the self characteristics of a storage battery and a super capacitor, the high-frequency part is used as a power reference value of the super capacitor, and then the direct-current micro-grid bus power is usedAnd (4) analyzing whether the hybrid energy storage device emits power or absorbs power by means of rate fluctuation, namely dividing the hybrid energy storage device into a Boost prediction model and a Buck prediction model.
5. The method for controlling the direct current microgrid hybrid energy storage system based on the multi-step model prediction as claimed in claim 2, characterized in that in the step 3:
acquiring bus voltage at the moment k of the direct-current micro-grid, and obtaining power values at the moment k +1 and the moment k +2 through calculation of a prediction model by taking bus current and voltage values of an energy storage device as input; comparing the power value at the moment k +1 with the power reference value, and obtaining the optimal solution x at the moment k +1 by a traversal algorithmimin1(k +1) and suboptimal solution ximin2(k +1), around the optimal solution x obtained in the first stepimin1(k +1) and suboptimal solution ximin2(k +1) respectively carrying out model prediction of the second step to obtain each predicted value x of the two solutions in the second stepij(k +2) and corresponding switch state combination Sij(k +1), j ═ 1, 2.., n; respectively calculating the optimal solution x of the first stepimin1(k +1) and suboptimal solution ximin2(k +1) two solutions x with the smallest error between the predicted value and the expected value obtained in the second stepijmin1(k +2) and xijmin2(k + 2); and then carrying out weighted comparison on error coefficients of the optimizing paths of each control period predicted by the multi-step model to obtain the optimal combination of the switching states of the converters, and applying the optimal combination to the hybrid energy storage system.
6. The method for controlling the direct current microgrid hybrid energy storage system based on the multi-step model prediction as claimed in claim 1, characterized in that in the step 4:
calculating an average SOC value according to the SOC value of the initial state of the storage battery pack and the number of the storage battery packs, calculating the difference between the average SOC value and the average SOC value to obtain the unbalance degree delta SOC of the SOC value of the storage battery pack, and setting a reasonable power deviation gain coefficient K according to the type of the storage battery and the requirement of adjusting time0The power deviation is adjusted, and the SOC values of the battery packs are balanced by the unbalanced energy distribution.
CN202110754949.0A 2021-07-01 2021-07-01 Control method of direct-current micro-grid hybrid energy storage system based on multi-step model prediction Pending CN113659558A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110754949.0A CN113659558A (en) 2021-07-01 2021-07-01 Control method of direct-current micro-grid hybrid energy storage system based on multi-step model prediction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110754949.0A CN113659558A (en) 2021-07-01 2021-07-01 Control method of direct-current micro-grid hybrid energy storage system based on multi-step model prediction

Publications (1)

Publication Number Publication Date
CN113659558A true CN113659558A (en) 2021-11-16

Family

ID=78477904

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110754949.0A Pending CN113659558A (en) 2021-07-01 2021-07-01 Control method of direct-current micro-grid hybrid energy storage system based on multi-step model prediction

Country Status (1)

Country Link
CN (1) CN113659558A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114336548A (en) * 2021-12-31 2022-04-12 苏州汇川控制技术有限公司 Short-circuit parameter determination method, short-circuit parameter determination equipment, storage medium and short-circuit protection method
CN114709866A (en) * 2022-03-15 2022-07-05 深圳市京泉华科技股份有限公司 Fractional order model prediction control method for power of electricity-hydrogen hybrid energy storage system
CN115102246A (en) * 2022-06-10 2022-09-23 荣信汇科电气股份有限公司 SOC balance control method for multiple battery clusters in novel energy storage unit
CN115912323A (en) * 2023-03-09 2023-04-04 四川大学 Hybrid energy storage control method and device for improving transient performance of direct-current micro-grid
CN116674425A (en) * 2023-06-07 2023-09-01 湖南文理学院 Coordinated control method and system for power battery pack based on total amount consistency
CN117239711A (en) * 2023-11-13 2023-12-15 四川大学 Energy storage control method and device for improving power supply quality of well group of oil pumping unit

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
戢理: "基于模型预测控制的直流微电网HESS控制策略研究", 中国优秀硕士学位论文全文数据库, pages 8 - 38 *
田明杰: "基于蓄电池与超级电容器的直流微网混合储能研究", 中国优秀硕士学位论文全文数据库, pages 41 - 42 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114336548A (en) * 2021-12-31 2022-04-12 苏州汇川控制技术有限公司 Short-circuit parameter determination method, short-circuit parameter determination equipment, storage medium and short-circuit protection method
CN114709866A (en) * 2022-03-15 2022-07-05 深圳市京泉华科技股份有限公司 Fractional order model prediction control method for power of electricity-hydrogen hybrid energy storage system
CN115102246A (en) * 2022-06-10 2022-09-23 荣信汇科电气股份有限公司 SOC balance control method for multiple battery clusters in novel energy storage unit
CN115912323A (en) * 2023-03-09 2023-04-04 四川大学 Hybrid energy storage control method and device for improving transient performance of direct-current micro-grid
CN116674425A (en) * 2023-06-07 2023-09-01 湖南文理学院 Coordinated control method and system for power battery pack based on total amount consistency
CN116674425B (en) * 2023-06-07 2023-12-01 湖南文理学院 Coordinated control method and system for power battery pack based on total amount consistency
CN117239711A (en) * 2023-11-13 2023-12-15 四川大学 Energy storage control method and device for improving power supply quality of well group of oil pumping unit
CN117239711B (en) * 2023-11-13 2024-02-02 四川大学 Energy storage control method and device for improving power supply quality of well group of oil pumping unit

Similar Documents

Publication Publication Date Title
CN113659558A (en) Control method of direct-current micro-grid hybrid energy storage system based on multi-step model prediction
US8004232B2 (en) Method of battery charging and power control in conjunction with maximum power point tracking
CN108879730B (en) Hybrid energy storage system and wind power fluctuation stabilizing method based on same
CN104242329A (en) Micro-grid hybrid energy storage system power distribution method based on fuzzy control rules
CN110303945B (en) Self-adaptive optimization balance control method for electric quantity of storage battery pack
CN109378881B (en) Bidirectional self-adaptive equalization control method for power battery pack
CN113541268A (en) Energy storage system and control method thereof
CN105375518A (en) Fuzzy control method and system for photovoltaic maximum power point tracking (MPPT)
CN103546034B (en) A kind of compounding feedforward control type Hysteresis control system
CN113488983A (en) Virtual direct current motor based on power distribution and method for jointly stabilizing direct current bus voltage through model prediction
CN114256913A (en) Redistribution battery pack charging and discharging management strategy based on optimal operation efficiency
Zhong et al. A bus-based battery equalization via modified isolated cuk converter governed by adaptive control
CN115133520B (en) Storage battery energy coordination control method suitable for light storage integrated system
Lu et al. A novel comprehensive SOC-voltage control scheme for lithium-ion battery equalization
CN112491099B (en) Battery active equalization control method and device
CN114884162A (en) Battery state-of-charge prediction optimization control method suitable for distributed energy storage type MMC
CN114362529A (en) Control method for direct-current micro-grid energy storage bidirectional DC/DC converter
CN110556868B (en) Wind power access system energy storage charging and discharging control method considering voltage stability
CN114552660A (en) Dual-multiplexing full-working-condition topology self-switching island-supported energy router
CN114362226A (en) Distributed energy storage type MMC battery charge state three-level balance control strategy under complex working condition
CN111864777A (en) Hybrid energy storage system and control method and system thereof
Tang et al. Research on Bidirectional DC/DC Converter Based on Passive Control
CN115133518A (en) Method for judging stability of large signal of parallel system of energy storage converter
Koseoglou et al. An improved active Li-ion battery equalization scheme for enhancing the performance of a nearly zero energy building and electric vehicle microgrid
Song et al. Study on output impedance characteristics of S 3 R and S 4 R topological based on SAR satellite platform power controller

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination