Multi-time scale model prediction energy management method for wind-wave-storage combined system

Abstract

A multi-time-scale model prediction energy management method for a wind-wave-storage combined system comprises the steps that according to the current frequency, the energy storage SOC and the wind wave power margin, the output power is rapidly adjusted by actively adjusting the motor opening degree of a wind wave combined power generation unit and the reference power response of an energy storage unit; and frequency optimization and energy storage state adjustment are carried out, so that system energy management optimization is realized. According to the invention, ms-level power management can be realized, system frequency optimization can be realized, hour-level SOC management can improve the long-term operation stability of the system, the potential of wind wave renewable energy in energy management can be fully excavated, and the stability and survivability of the system can be improved.

Description

Technical Field

[0001] The present invention relates to a technology in the field of island energy management, specifically a multi-time scale model prediction energy management method for a wind-wave-storage combined system. Background Art

[0002] The wind-wave-storage microgrid system is an important energy supply mode for future island energy systems. However, the intermittent and volatile characteristics of wind and wave energy lead to frequent power fluctuations in the system, which is not conducive to the stable control of the system frequency. In addition, small-capacity energy storage systems are difficult to ensure long-term stable output under extreme working conditions, which is not conducive to the stable operation of the system on multiple time scales of ms and h. Summary of the invention

[0003] In view of the shortcomings of existing PI-based energy storage collaborative control methods that fail to effectively consider the complex state and operation constraints of the system or are only effective for control variables but cannot take into account the numerous parameters and state variables of the system, the present invention proposes a multi-time scale model prediction energy management method for a wind-wave-storage combined system, which can simultaneously achieve ms-level power management to achieve system frequency optimization, and hour-level SOC management to improve the long-term operation stability of the system, fully tap the potential of wind and wave renewable energy in energy management, and improve system stability and survivability.

[0004] The present invention is achieved through the following technical solutions:

[0005] The present invention relates to a multi-time scale model prediction energy management method for a wind-wave-storage combined system. According to the current frequency, energy storage SOC and wind and wave power margin, the output power is quickly adjusted by actively adjusting the motor opening of the wind and wave combined power generation unit and the reference power response of the energy storage unit, while frequency optimization and energy storage state adjustment are performed, thereby realizing system energy management optimization.

[0006] The present invention relates to a wind-wave-storage combined power supply system for realizing the above-mentioned method, comprising: a wind-wave combined power generation unit and an energy storage unit, wherein: the wind-wave combined power generation unit is a hydraulic transmission power generation device, which is composed of two hydraulic motors coaxially connected to realize the superposition of wind energy and wave energy and jointly drive the synchronous power generation stage to directly output AC power. The unit realizes output power regulation by controlling the opening of the two hydraulic motors, and the energy storage unit adopts droop control to be incorporated into the system bus, and realizes output power regulation by controlling the reference power. Technical Effects

[0007] The present invention conducts refined modeling of the state space of frequency-motor opening-energy storage reference power-energy storage SOC of the wind-wave-island storage microgrid system, clarifies the system variables and their interaction mechanisms that affect the comprehensive operating characteristics of the system, and establishes a prediction model for system frequency and energy storage charge state; according to the control requirements of the microgrid system at different time scales, a system dual-mode energy management control strategy is proposed. At the ms time scale, the system takes frequency optimization as the main goal, and at the h time scale, the system takes energy storage SOC regulation as the main goal; a system multi-time scale operation mode switching method is designed, and the weight coefficient of the cost function is adaptively adjusted according to the system frequency deviation, thereby realizing flexible switching of the system's multiple operation modes. Compared with the prior art, the system state constraints and operation constraints of the present invention are more precise, and can realize the coordinated allocation of wind-wave-storage output power in the multi-time scale energy management process, and use the surplus wind and wave energy to adjust the energy storage SOC, fully tap the potential of wind and waves in system energy management, and further improve the stability and economy of the island microgrid system. At the same time, it can adjust the control strategy in real time according to the system state, realize flexible switching between ms-level power management and h-level state management, and take into account ms-level system frequency control and h-level system energy optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 It is the principle diagram of the present invention;

[0009] Figure 2 This is a structural diagram of a wind-wave combined power generation system;

[0010] Figure 3 It is a schematic diagram of the variation characteristics of the energy storage unit correction coefficient;

[0011] Figure 4 A schematic diagram of the model prediction energy management algorithm flow;

[0012] Figure 5 Schematic diagram of frequency modulation performance for different networking solutions;

[0013] Figure 6 It is a schematic diagram of system state changes in transient processes;

[0014] In the figure: (a) system state quantity and output power of each unit, (b) motor opening and energy storage reference power;

[0015] Figure 7 This is a schematic diagram showing the performance comparison of different networking solutions when wind and wave power is insufficient;

[0016] In the figure: (a) system frequency, (b) energy storage output power;

[0017] Figure 8 It is a schematic diagram for comparing the frequency modulation performance of different MPC methods;

[0018] In the figure: (a) wind and wave power is sufficient, (b) wind and wave power is insufficient;

[0019] Fig. 9 This is a schematic diagram of the wind-wave-load time distribution of an island over 72 hours;

[0020] In the figure: (a) wind and wave resources, (b) source load power;

[0021] Fig.10 This is a schematic diagram of the energy storage SOC change of an island over 72 hours;

[0022] Fig.11 It is a schematic diagram of the operating performance of different control strategies;

[0023] In the figure: In the figure: (a) system frequency, (b) energy storage SOC. DETAILED DESCRIPTION

[0024] like Figure 1 As shown, this embodiment involves a multi-time scale model prediction energy management method for a wind-wave-storage combined system. According to the current system frequency, energy storage SOC and wind and wave power margin, the output power is quickly adjusted by actively adjusting the motor opening of the wind and wave combined power generation unit and the reference power response of the energy storage unit, while frequency optimization and energy storage state adjustment are performed, thereby achieving system energy management optimization.

[0025] like Figure 2 As shown, the wind-wave combined power generation unit comprises: a bidirectional hydraulic cylinder ①, a one-way valve ②, a pneumatic accumulator ③, a hydraulic tank ④ and a pressure relief valve ⑤ located on the wave energy unit side, a variable displacement hydraulic motor ⑥ and a synchronous generator ⑨ respectively connected to both ends of the coupling, a fixed displacement hydraulic pump ⑦ and a wind turbine ⑧ located on the wind power unit side, wherein: the wind turbine drives the fixed displacement hydraulic pump to rotate and pump out hydraulic oil, which drives the variable displacement motor after passing through the hydraulic circuit; the wave energy unit captures wave energy through a bidirectional hydraulic cylinder, and drives the hydraulic motor through a one-way valve rectification circuit. The pneumatic accumulator in the circuit compresses the internal air to smooth the rapid changes in the flow and pressure in the circuit, thereby reducing the output power fluctuation. When the pressure in the circuit exceeds the threshold set by the pressure relief valve, the excess energy is discharged from the system in the form of hydraulic energy.

[0026] The coupling is equivalent to an inertia-spring-damping system, which realizes the speed synchronization and mechanical power superposition of the two variable displacement hydraulic motors.

[0027] The system energy management refers to: power management at the ms level and energy storage SOC management at the h level according to the time scale, wherein: power management refers to the real-time coordinated allocation of the output power of wind, wave, and storage power sources according to the microgrid frequency deviation to achieve system frequency optimization; energy storage SOC management refers to the use of surplus wind and wave power to participate in energy storage state regulation by adjusting the energy storage reference power on the basis of achieving frequency stability control.

[0028] The power management is specifically: P wd , P we , P L and f are the output power of the wind and wave combined power generation unit, the output power of the wave energy unit, the load demand power and the system frequency, respectively. f0 is the rated frequency of the system, where: the output power of the wind and wave combined power generation unit η is the energy transfer efficiency of the hydraulic motor, p is the hydraulic circuit pressure, ω is the motor speed, V max is the maximum displacement of the hydraulic motor, S is the opening of the hydraulic motor, the subscript i=1, 2 corresponds to the wind-wave combined power generation unit and the wave energy unit, and the output power of the energy storage unit P b =P0-m(f-f0), P0 is the reference power of the energy storage unit, m is the droop coefficient, and the system moment of inertia is the sum of the moments of inertia of the rotating equipment directly connected to the synchronous machine, specifically: J1, J2 and J e They are the moments of inertia of the wind turbine motor, wave power motor and synchronous machine rotor, N and P respectively. N is the transmission ratio of the coupling and the number of magnetic pole pairs of the synchronous generator.

[0029] The opening of the hydraulic motor can be adjusted arbitrarily within (0-1] to achieve stepless transmission and ensure smooth and continuous power conversion.

[0030] When the system frequency is disturbed, the system coordinates and regulates Pwd, Pwe, and Pb to quickly respond to frequency changes and achieves rapid frequency recovery through ms-level power management.

[0031] The energy storage SOC management is specifically as follows: Where: W b For the energy storage capacity, the reference power P0 of the energy storage unit is set as the controlled quantity, and the droop coefficient m is set as a constant. When the system frequency is stable at the rated value, the energy storage output power is equal to the reference power P0. The reference power P0 of the energy storage unit is actively adjusted according to the output power margin of the wind and wave unit to achieve energy storage SOC regulation.

[0032] The multi-time scale model is obtained based on the system frequency-motor opening-energy storage reference power state space equation, specifically: Where: I is the identity matrix, T s is the sampling period, k is the sampling time, Y(k+n|k) and ΔU(k) ​​are the system output prediction sequence and control quantity increment sequence respectively, S x , S u , S d They represent the state prediction matrix, control quantity prediction matrix and non-control quantity prediction matrix respectively, Δ represents the difference of state quantity between two sampling times, and x(k)y(k) are the system state quantity vector and output quantity vector at time k respectively.

[0033] The system frequency-motor opening-energy storage reference power state space equation is: Where: Pm is the combined wind and wave output power, T m is the time constant of the hydraulic motor response, ω c is the speed of the coupling. The coupling is equivalent to an inertia-spring-damper unit, which connects the hydraulic motors of the wind and wave combined power generation unit and the wave energy unit to achieve speed synchronization and power superposition.

[0034] The active adjustment of the motor opening and the reference power response to quickly adjust the output power while performing frequency optimization and energy storage state adjustment is achieved in the following ways:

[0035] Step 1: Construct the cost function of system operation, and take the minimum cost function as the control target, establish the system energy management objective function, punish the system output deviation and the controlled quantity increment, and comprehensively consider the influence of frequency deviation, energy storage SOC deviation, motor opening increment, and energy storage reference power increment on the comprehensive operation performance of the system, so that the system can achieve frequency regulation and energy storage state adjustment in energy management with minimal action, specifically: Among them: α, β, λ, μ and are penalty coefficients, P b0 and R b0 is the rated output power and rated ramp rate of the energy storage unit, and It is a function of the SOC of the energy storage unit, which are the correction coefficients of discharge power, charging power, discharge ramp rate and charging ramp rate respectively. The subscript i represents the sampling point, ref represents the reference value, k represents the sampling time, and P represents the control time domain.

[0036] Step 2: Construct the actual operation restrictions of the system, including constraints on the motor opening range and speed, reference power range and speed, energy storage output power range and ramp rate, specifically:

[0037] Step 3: Convert the control problem into a quadratic optimization problem with the control quantity increment sequence ΔU(k) ​​as the optimization variable, and use the quadratic programming solver (Quadprog) function to quickly solve it, specifically: Among them: U is the independent variable, H is the quadratic term coefficient, G is the linear term coefficient, Q and R are the weight matrices of the system output deviation and the control amount increment in the cost function respectively, and Ref is the output reference value matrix.

[0038] like Figure 4 As shown, the energy management optimization includes: power management mode, that is, only optimizing and controlling the frequency, and SOC management mode, that is, adjusting the energy storage state while maintaining frequency stability, specifically: taking frequency deviation as the basis for mode switching, by adjusting the weight β of SOC deviation: when the frequency deviation exceeds the limit, set β=0, the system works in frequency optimization mode, at this time the deviation between SOC and reference value is not punished, and the system does not actively adjust SOC; when the frequency deviation does not exceed the limit, set β≠0, the working mode is switched to SOC adjustment mode, at this time the controller will actively adjust the energy storage state to achieve the goal of minimizing the cost function.

[0039] The reference value of SOC is calculated based on the current wind and wave margin, specifically: Among them: P1 and P2 are the output powers of the wind-wave combined power generation unit and the wave energy unit respectively.

[0040] After specific actual experiments, based on the equipment parameters of the Wenzhou Nanji Island microgrid project, as shown in Table 1, the RT-LAB semi-physical simulation model of a typical wind-wave-storage island microgrid system including a wind-wave complement device, a hydraulic transmission device, a coupling, energy storage, a typical load and a control unit was taken as the object. The simulation platform used included the RT-LAB OP7000 hardware simulation real-time simulator, the AN706 controller, the XLINX FPGA ZYNQ7020 development board and the Tektronix MSO44 oscilloscope. The controller samples the state quantity signals such as system frequency and power supply power, and the FPGA quickly solves the optimal solution of the model predictive control, realizing the semi-physical simulation of frequency optimization and energy storage state management, so as to verify the effectiveness of the proposed system and method.

[0041] Table 1 Island microgrid power system parameters

[0042] This embodiment uses FPGA to solve the optimal solution of model predictive control online. To ensure the real-time and accuracy of the calculation, the sampling step size T s It is set to 0.02s and the step size is 5. The system operation constraints and state constraints related parameters are selected according to the real microgrid data, as shown in Table 2.

[0043] Table 2 System state constraints and weight coefficients

[0044] Comparison of frequency regulation performance of different networking schemes: In order to compare the frequency regulation performance of different networking schemes, three schemes were compared and analyzed under the same operating conditions and constraints. The results are as follows: Figure 5 and as shown in Table 3. Among them: The environmental input and load demand are set as follows: The input of the wind and wave combined power generation unit is a mean of 8m / s and a variance of 0.25m 2 / s 2 The wave energy unit input is a periodic wave force with a period of 2πs and an amplitude of 400kN; the load demand is a load increase of 20kW within 1s.

[0045] Table 3 Comparison of frequency characteristics of different networking solutions

[0046] All three networking schemes can achieve frequency regulation: the system composed only of wind-wave networking achieves frequency regulation by controlling the speed of the generator, but due to the large rotational inertia of the system and the slow speed regulation, the transient time and overshoot of the frequency response exceed the limit; the system composed only of energy storage networking responds quickly, but there is a steady-state deviation, and the requirements for energy storage capacity are relatively high; in comparison, the wind-wave-storage combined networking system couples the advantages of the two and performs superior in frequency optimization, reducing the maximum frequency deviation by 87.6% and 44.4% respectively, and the frequency recovery time is also reduced by 74.4% and 16.7% respectively, which is conducive to the system reaching stability faster. When the frequency drops rapidly, the system switches to the frequency optimization mode. At this time, the energy storage unit quickly outputs power through droop control to ensure that the lowest frequency point does not exceed the limit. At the same time, the wind, wave, and storage units respond to frequency changes and increase the system power output by adjusting the motor opening and reference power respectively until the system returns to a steady state. After that, the system resumes the SOC adjustment mode, and the surplus wind and wave energy is used to adjust the energy storage state. The characteristics of the system state changes at each stage are as follows: Figure 6 The power supply output power characteristics are shown in Table 4.

[0047] Table 4 Power output power variation characteristics at each stage

[0048] In addition, changes in ocean meteorology may lead to insufficient wind and wave power generation. In order to verify the frequency regulation performance of this method under this working condition, the experiment simulates a scenario in which the wind and wave input is reduced by 50% at a certain moment. At this time, the wind and wave output power is less than the load demand, and the wind and wave combined power generation unit operates at maximum power, and the power gap is supported by the energy storage unit.

[0049] The operation results of the three networking solutions are as follows: Figure 7 As shown. The wind-wave networking system includes a hydraulic energy storage structure, which can quickly release power when power is insufficient to slow down the frequency drop, but its output power is uncontrollable; after the wind and wave output decreases, the energy storage network needs to bear more power output, resulting in a larger frequency deviation; in contrast, the wind-wave-storage joint networking system has a stronger output capacity, with a maximum frequency of only 0.06Hz, which is 57.1% and 97.2% lower than the first two solutions, respectively. At the same time, the time to restore the steady state is also reduced by 65.5% and 89.4%, respectively. Although the energy storage unit is still outputting power at this time, its reference power is also adjusted during the frequency modulation process, which is equivalent to secondary frequency modulation of the system, so there is no frequency deviation in the steady state.

[0050] Comparison of energy storage state management performance: Under the same working conditions, the control performance differences of the two MPC methods at two different time scales, frequency optimization and SOC regulation, are compared. The frequency optimization performance of the two methods is investigated in a short time scale. The conditions where the island load increases by 20kW within 1s are simulated in two scenarios: sufficient wind and wave power and insufficient wind and wave power (the same as Example 1). The results are as follows: Figure 8 shown.

[0051] When the wind and wave power is sufficient, both MPC methods can respond to frequency fluctuations in real time, adjust the power output of each unit, and achieve frequency optimization. In contrast, the MPC method taking into account the energy storage state can adjust the reference power according to the frequency deviation on the basis of energy storage droop control, which is conducive to the rapid recovery of frequency, and can reduce the maximum frequency deviation by 21.1%, while reducing the frequency recovery time by 37.5%.

[0052] When wind and wave power is insufficient, the energy storage unit needs to output to supplement the power gap. Since the droop control that only considers the primary frequency regulation will cause a stable deviation in the system frequency, the MPC method that does not take into account the energy storage status cannot restore the frequency to the rated value when the system is stable, but changes with the fluctuation of wind and wave output power. In contrast, the MPC method that takes into account the energy storage status can actively adjust the reference power point in response to frequency fluctuations, perform secondary frequency regulation on the system, and the adjustment speed is greater than the speed of change of wind and wave power, ensuring that the frequency can be restored to the rated value when the system is stable.

[0053] The energy storage SOC regulation performance of the two methods was investigated over a long time scale, and the operating conditions of an island for 3 days and 72 hours were simulated. The power fluctuations of the source and load during this period and the SOC operation results were shown in Figure 2. Fig. 9 and 10 shown.

[0054] The wind and wave resources on the island changed constantly within 72 hours, and there were multiple cases of insufficient wind and wave power. At this time, the energy storage unit needed to continuously output to supplement the power gap. The MPC method that did not take into account the energy storage status could not convert the excess power into electrical energy storage when the power was sufficient, and therefore could not restore the energy storage status. In contrast, the MPC method that took into account the energy storage status could convert the excess wind and wave power into electrical energy storage by adjusting the reference power point. The energy storage SOC decline under the control of this method was reduced by 88.5%.

[0055] Performance comparison of different control strategies: The comparison strategy constructs a two-layer control system with top-level energy management and bottom-level power supply regulation. Based on the minimum usage cost and energy storage state balance, the voltage control and SOC regulation of the DC microgrid are realized through PI control. The example sets the same system operation and constraints, and compares the system frequency and energy storage SOC change characteristics under different PI parameters.

[0056] Table 5 PI parameter settings

[0057] The comparison strategy uses the cost function to calculate the penalty item of each unit operation, taking into account the overall system benefit. This strategy uses different control functions in the two scenarios of frequency optimization and energy storage status management, and can switch quickly according to the system operation status, which is more targeted, so it also shows better performance in specific scenarios. Fig.11 As shown in Table 6, specifically, the proposed strategy has a faster frequency recovery speed and a smaller maximum frequency deviation during frequency transient optimization; in energy storage state management, the power response is faster and the energy conversion is more complete.

[0058] Table 6 Comparison of frequency and SOC of different control strategies

[0059] When examining the frequency optimization performance, for the two-layer control strategy based on PI regulation, as the proportional gain increases, the system response speed continues to improve, the maximum frequency deviation and frequency recovery time decrease, but too large a proportional gain will cause frequency overshoot, which is not conducive to system stability; compared with PI 1 with the best comprehensive performance, the proposed strategy shortens the frequency recovery time by 43.3% and the maximum frequency deviation by 46.4%.

[0060] When examining the SOC regulation performance, the comparison strategy calculates the power distribution of each power source through the top-level energy management system, and then realizes power output through PI control. The power supply does not respond quickly enough, resulting in the inability of energy storage to efficiently absorb excess wind and wave power. In contrast, the proposed strategy updates the SOC reference value in real time according to the current power margin, directly adjusts control quantities such as motor opening and energy storage reference power, and maximizes the use of excess wind and wave power to adjust the energy storage state while ensuring frequency stability, thereby improving the energy storage energy absorption efficiency by 34.7%.

[0061] Compared with the existing technology, this method monitors the system status in real time, takes whether the frequency deviation exceeds the limit as the switching basis, and actively adjusts the weight coefficient of the cost function to achieve flexible switching of the dual operating modes of power management and SOC management, thereby realizing multi-objective optimization energy management of system operation.

[0062] The above-mentioned specific implementation can be partially adjusted in different ways by those skilled in the art without departing from the principle and purpose of the present invention. The protection scope of the present invention shall be based on the claims and shall not be limited by the above-mentioned specific implementation. Each implementation scheme within its scope shall be subject to the constraints of the present invention.

Claims

1. A multi-time scale model prediction energy management method for a wind-wave-storage combined system, characterized in that: According to the current system frequency, energy storage SOC and wind and wave power margin, the output power is quickly adjusted by actively adjusting the motor opening of the wind and wave combined power generation unit and the reference power response of the energy storage unit, while frequency optimization and energy storage status adjustment are performed, thereby achieving system energy management optimization.

2. The multi-time scale model prediction energy management method for the wind-wave-storage combined system according to claim 1 is characterized in that: The wind and wave combined power generation unit comprises: a bidirectional hydraulic cylinder, a one-way valve, an air-filled accumulator, a hydraulic tank and a pressure relief valve located on the wave energy unit side, a variable displacement hydraulic motor and a synchronous generator respectively connected to the two ends of the coupling, a fixed displacement hydraulic pump and a wind turbine located on the wind power unit side, wherein: the wind turbine drives the fixed displacement hydraulic pump to rotate and pump out hydraulic oil, which drives the variable displacement motor after passing through the hydraulic circuit; the wave energy unit captures wave energy through a bidirectional hydraulic cylinder, drives the hydraulic motor through a one-way valve rectification circuit, and the air-filled accumulator in the circuit compresses the internal air to smooth the rapid changes in the flow and pressure in the circuit, thereby reducing the output power fluctuation. When the pressure in the circuit exceeds the threshold set by the pressure relief valve, the excess energy is discharged from the system in the form of hydraulic energy.

3. The multi-time scale model prediction energy management method for the wind-wave-storage combined system according to claim 1 is characterized in that: The system energy management refers to: power management at the ms level and energy storage SOC management at the h level according to the time scale, wherein: power management refers to the real-time coordinated allocation of the output power of wind, wave, and storage power sources according to the microgrid frequency deviation to achieve system frequency optimization; energy storage SOC management refers to the use of surplus wind and wave power to participate in energy storage state regulation by adjusting the energy storage reference power on the basis of achieving frequency stability control.

4. The multi-time scale model prediction energy management method for the wind-wave-storage combined system according to claim 1 is characterized in that: The power management is specifically: P wd , P we , P L and f are the output power of the wind and wave combined power generation unit, the output power of the wave energy unit, the load demand power and the system frequency, respectively. f0 is the rated frequency of the system, where: the output power of the wind and wave combined power generation unit η is the energy transfer efficiency of the hydraulic motor, p is the hydraulic circuit pressure, ω is the motor speed, V max is the maximum displacement of the hydraulic motor, S is the opening of the hydraulic motor, subscript i=1, 2 corresponds to the wind and wave combined power generation unit and the wave energy unit, and the output power of the energy storage unit P b =P0-m(f-f0), P0 is the reference power of the energy storage unit, m is the droop coefficient, and the system moment of inertia is the sum of the moments of inertia of the rotating equipment directly connected to the synchronous machine, specifically: J1, J2 and J e They are the moments of inertia of the wind turbine motor, wave power motor and synchronous machine rotor, N and P respectively. N is the transmission ratio of the coupling and the number of magnetic pole pairs of the synchronous generator.

5. The multi-time scale model prediction energy management method for wind-wave-storage combined system according to claim 1 is characterized in that: The energy storage SOC management is specifically as follows: Where: W b For the energy storage capacity, the reference power P0 of the energy storage unit is set as the controlled quantity, and the droop coefficient m is set as a constant. When the system frequency is stable at the rated value, the energy storage output power is equal to the reference power P0. The reference power P0 of the energy storage unit is actively adjusted according to the output power margin of the wind and wave unit to achieve energy storage SOC regulation.

6. The multi-time scale model prediction energy management method for wind-wave-storage combined system according to claim 1 is characterized in that: The multi-time scale model is obtained based on the system frequency-motor opening-energy storage reference power state space equation, specifically: Where: I is the identity matrix, T s is the sampling period, k is the sampling time, Y(k+n|k) and ΔU(k) ​​are the system output prediction sequence and control quantity increment sequence respectively, S x , S u , S d They represent the state prediction matrix, control quantity prediction matrix and non-control quantity prediction matrix respectively, Δ represents the difference of state quantity between two sampling times, x(k)y(k) are the system state quantity vector and output quantity vector at time k respectively, 7. The multi-time scale model prediction energy management method for wind-wave-storage combined system according to claim 1 is characterized in that: The system frequency-motor opening-energy storage reference power state space equation is: Where: Pm is the combined wind and wave output power, T m is the time constant of the hydraulic motor response, ω c is the speed of the coupling. The coupling is equivalent to an inertia-spring-damper unit, which connects the hydraulic motors of the wind and wave combined power generation unit and the wave energy unit to achieve speed synchronization and power superposition.

8. The multi-time scale model prediction energy management method for wind-wave-storage combined system according to claim 1 is characterized by: The active adjustment of the motor opening and the reference power response to quickly adjust the output power while performing frequency optimization and energy storage state adjustment is achieved in the following ways: Step 1: Construct the cost function of system operation, and take the minimum cost function as the control target, establish the system energy management objective function, punish the system output deviation and the controlled quantity increment, and comprehensively consider the influence of frequency deviation, energy storage SOC deviation, motor opening increment, and energy storage reference power increment on the comprehensive operation performance of the system, so that the system can achieve frequency regulation and energy storage state adjustment in energy management with minimal action, specifically: Among them: α, β, λ, μ and are penalty coefficients, P b0 and R b0 is the rated output power and rated ramp rate of the energy storage unit, and is a function of the SOC of the energy storage unit, which are the correction coefficients of discharge power, charge power, discharge ramp rate and charge ramp rate, respectively. The subscript i represents the sampling point, ref represents the reference value, k represents the sampling time, and P represents the control time domain; Step 2: Construct the actual operation restrictions of the system, including constraints on the motor opening range and speed, reference power range and speed, energy storage output power range and ramp rate, specifically: Step 3: Convert the control problem into a quadratic optimization problem with the control quantity increment sequence ΔU(k) ​​as the optimization variable, and use the quadratic programming solver (Quadprog) function to quickly solve it, specifically: stcU(k)≥b(k+1|k) Among them: U is the independent variable, H is the quadratic term coefficient, G is the linear term coefficient, Q and R are the weight matrices of the system output deviation and the control amount increment in the cost function respectively, and Ref is the output reference value matrix.

9. The multi-time scale model prediction energy management method for wind-wave-storage combined system according to claim 1 is characterized in that: The energy management optimization includes: a power management mode, that is, only optimizing and controlling the frequency, and an SOC management mode, that is, adjusting the energy storage state while maintaining the frequency stability. Specifically, the frequency deviation is used as the basis for mode switching, and the weight β of the SOC deviation is adjusted to achieve: when the frequency deviation exceeds the limit, β=0 is set, and the system works in the frequency optimization mode. At this time, the deviation between the SOC and the reference value is not punished, and the system does not actively adjust the SOC; when the frequency deviation does not exceed the limit, β≠0 is set, and the working mode is switched to the SOC adjustment mode. At this time, the controller will actively adjust the energy storage state to achieve the goal of minimizing the cost function.

10. The multi-time scale model prediction energy management method for the wind-wave-storage combined system according to claim 9 is characterized in that: The reference value of SOC is calculated based on the current wind and wave margin, specifically: Among them: P1 and P2 are the output powers of the wind-wave combined power generation unit and the wave energy unit respectively.

Images