Hybrid energy storage system and scheduling method therefor

Abstract

The present invention discloses a scheduling system and method for a hybrid energy storage system, including step S1: detecting whether a scheduling command has been received by a CPU in real time and detecting and obtaining the maximum power of a power-type energy storage medium; step S2: turning on and replenishing a capacitive energy storage medium if the power value required by the scheduling command is greater than the maximum power of the power-type energy storage medium; step S3: simultaneously outputting and predicting the amplitude and duration of the next scheduling command using a convolutional neural network; and step S4: adjusting both energy storage media based on the predicted amplitude and duration to respond to the scheduling command. The scheduling method for a hybrid energy storage system predicts the amplitude and duration of the next scheduling power command based on the statistical characteristics of historical data, significantly improving the controllability of power scheduling and enabling application public power to be pre-configured, thereby improving power scheduling efficiency.

Description

[Technical Field]

[0001] The present invention relates to the technical field of scheduling of energy storage systems, and in particular to a hybrid energy storage system and a scheduling method thereof. [Background technology]

[0002] With the continuous development of energy storage technology, smoothing out fluctuations in power, voltage, and current by deploying hybrid energy storage systems is an important solution in current energy storage technology, and hybrid energy storage systems can be divided into electric energy storage and capacitive energy storage. The former, represented by supercapacitors, has advantages such as a long service life and high power density, and by combining it with battery-type capacitive energy storage, which has disadvantages such as a short service life and long charge / discharge cycle, it can achieve optimized utilization of stored energy.

[0003] However, the energy storage technologies available in the prior art lack sufficient foresight, and only adaptively adjust when environmental conditions change, rather than being able to predict in advance and make early scheduling adjustments, resulting in low controllability. Furthermore, while the prior art employs training based on artificial intelligence algorithms to obtain results, the raw data trained based on the artificial intelligence algorithms is relatively rough, resulting in inaccurate output results and relatively backward adjustment methods. Furthermore, the environmental temperature has a significant impact on the energy storage system, shortening its service life, and there is no relevant technology to extend the service life of the energy storage system. Summary of the Invention

[0004] For the above problems mentioned in the prior art, in order to solve the above technical problems, the present invention provides: Step S1: detecting whether a scheduling command is received by a CPU in real time, and detecting and obtaining the maximum power of the power-type energy storage medium; Step S2: if the power value required by the scheduling command is greater than the maximum power of the power-type energy storage medium, turn on and replenish the capacitive energy storage medium; otherwise, respond to the scheduling command using the power-type energy storage medium; At the same time as performing step S2, input the current scheduling command request parameters, the operation parameters of the electric energy storage medium, and the operation parameters of the capacitive energy storage medium into the trained convolutional neural network, and output and predict the amplitude and duration of the next scheduling command through the convolutional neural network in step S3; Responding to the scheduling command by the electric energy storage medium when the amplitude value of the command output and predicted by the convolutional neural network is smaller than the maximum power of the electric energy storage medium and the duration is shorter than the discharge duration of the electric energy storage medium, and responding to the scheduling command by the direct capacitive energy storage medium when the predicted duration is longer than the discharge duration of the electric energy storage medium; and step S4 of responding to the scheduling command commonly by both the capacitive and electric energy storage media if the predicted amplitude value exceeds the maximum power to which the capacitive or electric energy storage media respond individually.

[0005] Preferably, the electric energy storage medium includes the use of a nickel-metal hydride battery, a lithium titanate battery, or a supercapacitor, and the capacitive energy storage medium includes the use of a lithium iron phosphate battery, a lead acid battery, or a lead carbon battery.

[0006] Preferably, the scheduling command request parameters include a maximum power value, duration, command issuance time, and command issuance button duration required by the scheduling command; the operating parameters of the power-type energy storage medium include a maximum power of the power-type energy storage medium, a charge capacity of the power-type energy storage medium, an operating duration of the power-type energy storage medium, and a maximum current value of the power-type energy storage medium; and the operating parameters of the capacitive energy storage medium include a maximum power of the capacitive energy storage medium, a charge capacity of the capacitive energy storage medium, an operating duration of the capacitive energy storage medium, and a maximum current value of the capacitive energy storage medium.

[0007] Preferably, the trained convolutional neural network is completed through the following training steps: step S21 of dividing 24 hours a day into 288 5-minute time slots; step S22 of performing FFT transformation on the power command for each time slot; step S23 of training the convolutional neural network on the amplitude-frequency characteristics of the power command for each time slot; and step S24 of completing the training of the convolutional neural network.

[0008] Preferably, performing an FFT frequency transformation on the power command includes: a step S221 of determining a total power composition of the hybrid storage system from an amplitude value of the power command; S222: calculating a duration period of the low frequency command and determining a storage duration of the capacitive energy storage medium from the period statistical data; S223: statistically calculating period characteristics of the high frequency command and determining a storage duration of the power type energy storage medium based on period statistical data of the high frequency command; S224: calculating the amplitude value characteristics of the low-frequency command and determining the power distribution ratio of the capacitive energy storage medium; and step S225 of calculating the amplitude value characteristics of the high frequency command and determining the power distribution ratio of the power type energy storage medium.

[0009] Preferably, the convolutional neural network employs an improved loss function as V:

[0010]

number

[0011] The present invention provides a hybrid energy storage scheduling system including a control system, a power conversion system, an electric energy storage medium, and a capacitive energy storage medium, The electric energy storage medium and the capacitive energy storage medium are connected to a power grid via independent power conversion systems, and the electric energy storage medium and the capacitive energy storage medium are connected to a control system, and the control system controls the electric energy storage medium, the capacitive energy storage medium, and the power converter through communication connections and selector switches; The control system includes two sets of virtual energy storage units, each of which stores protection parameters and operation parameters of an electric energy storage medium and a capacitive energy storage medium; When the storage-type energy storage medium is connected to the power conversion system, the control system enables a protection operation parameter of the virtual storage-type energy storage medium to realize operation scheduling and protection of the storage-type energy storage medium; When the capacitive energy storage medium is connected to the power conversion system, the control system enables the operation protection parameters of the virtual capacitive energy storage medium, and realizes the operation scheduling and protection of the capacitive energy storage medium, wherein the control system is configured to execute the method of claim 1 by a processor of the control system. A hybrid energy storage scheduling system is also provided.

[0012] Preferably, the electric energy storage medium includes the use of a nickel-metal hydride battery, a lithium titanate battery, or a supercapacitor, and the capacitive energy storage medium includes the use of a lithium iron phosphate battery, a lead acid battery, or a lead carbon battery.

[0013] Preferably, the electric energy storage medium employs a high-power nickel-metal hydride battery suitable for low-temperature operation, and the capacitive energy storage medium employs a lithium iron phosphate battery to form a hybrid energy storage system.

[0014] Preferably, the nickel-metal hydride battery pack and the lithium iron phosphate battery pack are mounted in an enclosed space within the container and spaced apart, and charging and discharging of the nickel-metal hydride battery is started first at low temperatures, and heat generated by the charging and discharging losses is used to heat the enclosed space and the lithium iron phosphate battery; After the ambient temperature in the sealed space has risen, charging and discharging of the lithium iron phosphate battery is resumed.

[0015] The present invention provides a hybrid energy storage system and a scheduling method thereof, and has the following beneficial technical effects:

[0016] 1. By predicting the amplitude and duration of the next scheduled power command based on the statistical characteristics of historical data, the controllability of power scheduling is greatly improved, and advance configuration of application exposure power is possible, thereby improving the efficiency of power scheduling.

[0017] 2. By constructing training data, the accuracy of the algorithm is significantly improved. Specifically, 24 hours a day is divided into 288 five-minute time slots, FFT analysis is performed on the power commands of each time slot, and the amplitude-frequency characteristics of the power commands of each time slot are studied and summarized to construct a convolutional neural network algorithm, which significantly improves the prediction accuracy.

[0018] 3. Dynamically respond to scheduling commands based on a comparison of the predicted amplitude value with the maximum power of the electric energy storage medium and the power of the capacitive energy storage medium, and a comparison of the predicted duration with the discharge duration of the electric energy storage medium and the discharge duration of the capacitive energy storage medium, thereby greatly improving the flexibility and scalability of the system.

[0019] 4. In application operating conditions where the ambient temperature is below 0 degrees for a long time, the heating loss of the lithium iron phosphate energy storage system is large, and the low temperature environment affects the cycle life of the battery. Therefore, a hybrid energy storage system is constructed using high-power nickel-metal hydride batteries and lithium iron phosphate batteries suitable for low-temperature operation, which greatly extends the battery life.

[0020] 5. An improved loss function is adopted as V, where ln used in the improved loss function is the natural logarithm, and the convolutional neural network is adjusted by the tangent value of the i-th time period sample vector and the historical average standard feature vector, which greatly improves the self-learning ability and prediction accuracy. [Brief explanation of the drawings]

[0021] In order to more clearly describe the embodiments of the present invention or the technical solutions of the prior art, the following briefly introduces the drawings used in the description of the embodiments or the prior art. Of course, the drawings described below are only some embodiments of the present invention, and those skilled in the art can further obtain other drawings based on these drawings without any creative work. [Figure 1]FIG. 1 is a schematic diagram of a hybrid energy storage system of the present invention. [Figure 2] FIG. 2 is an FFT frequency analysis diagram of the power command of the hybrid energy storage system of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0022] The following provides a clear and complete description of the technical solutions in the embodiments of the present invention, in conjunction with the drawings in the embodiments of the present invention, but of course, the described embodiments are only a part of the embodiments of the present invention, and are not all of the embodiments. All other embodiments that can be obtained by those skilled in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of the claims of the present invention.

[0023] Example 1 Compared with traditional energy storage systems, embodiments of the present invention provide a scheduling method for a hybrid energy storage system that predicts the amplitude and duration of the next scheduled power command based on statistical characteristics of historical data, and further provides scheduling flexibility for advance scheduling; Step S1: detecting whether a scheduling command is received by a CPU in real time, and detecting and obtaining the maximum power of the power-type energy storage medium; Step S2: if the power value required by the scheduling command is greater than the maximum power of the power-type energy storage medium, turn on and replenish the capacitive energy storage medium; otherwise, respond to the scheduling command using the power-type energy storage medium; At the same time as performing step S2, input the current scheduling command request parameters, the operation parameters of the electric energy storage medium, and the operation parameters of the capacitive energy storage medium into the trained convolutional neural network, and output and predict the amplitude and duration of the next scheduling command through the convolutional neural network in step S3; Step S4: responding to the scheduling command by the power-type energy storage medium if the amplitude value of the command output and predicted by the convolutional neural network is smaller than the maximum power of the power-type energy storage medium and the duration is shorter than the discharge duration of the power-type energy storage medium, and responding to the scheduling command by the direct-capacitive energy storage medium if the predicted duration is longer than the discharge duration of the power-type energy storage medium; If the predicted amplitude value exceeds the maximum power to which either the capacitive or the electric energy storage medium responds individually, the scheduling command is responded to jointly by both energy storage media.

[0024] In some embodiments, the electric energy storage medium includes the use of a nickel metal hydride battery, a lithium titanate battery, or a supercapacitor, and the capacitive energy storage medium includes the use of a lithium iron phosphate battery, a lead acid battery, or a lead carbon battery.

[0025] In some embodiments, the scheduling command request parameters include a maximum power value, duration, command issuance time, and command issuance button duration required by the scheduling command; the operating parameters of the power-type energy storage medium include a maximum power of the power-type energy storage medium, a charge capacity of the power-type energy storage medium, an operating duration of the power-type energy storage medium, and a maximum current value of the power-type energy storage medium; and the operating parameters of the capacitive energy storage medium include a maximum power of the capacitive energy storage medium, a charge capacity of the capacitive energy storage medium, an operating duration of the capacitive energy storage medium, and a maximum current value of the capacitive energy storage medium.

[0026] In some embodiments, the trained convolutional neural network is completed through the following training steps: step S21: dividing 24 hours a day into 288 5-minute time slots; step S22: performing FFT transformation on the power command for each time slot; step S23: training the convolutional neural network on the amplitude-frequency characteristics of the power command for each time slot; and step S24: completing the training of the convolutional neural network.

[0027] In some embodiments, as shown in FIG. 2 , performing an FFT frequency transformation on the power command includes: S221, determining a total power composition of the hybrid storage system from the amplitude value of the power command; In some embodiments, calculating a duration period of the low frequency command and determining a storage duration of the capacitive energy storage medium from the period statistics; In some embodiments, calculating a period characteristic of the high frequency command and determining a storage duration of the power type energy storage medium from the period statistical data of the high frequency command; In some embodiments, calculating the amplitude characteristic of the low frequency command to determine the power allocation ratio of the capacitive energy storage medium; In some embodiments, the amplitude value characteristics of the high frequency command are statistically analyzed to determine the power distribution ratio of the electrically powered energy storage medium.

[0028] In some embodiments, the convolutional neural network employs an improved loss function as V:

[0029]

number

[0030] The present invention further provides a hybrid energy storage scheduling system, in which energy storage media share a set of power conversion systems, and different energy storage media achieve time-sharing sharing for the power conversion systems through switching circuits and control and protection systems. As shown in Figure 1, the system includes a control system, a power conversion system, an electric energy storage medium, and a capacitive energy storage medium, Energy storage 1 and energy storage 2 are composed of electric and capacitive energy storage media, respectively. The electric energy storage media mainly use nickel-metal hydride batteries, lithium titanate batteries, supercapacitors, etc., while the capacitive energy storage media mainly use lithium iron phosphate batteries, lead-acid batteries, lead-carbon batteries, etc., and the power conversion system is a general-purpose power converter.

[0031] The DC buses of energy storage 1 and energy storage 2 are connected to the DC bus of the power converter via a transfer switch, and the DC voltage ranges of energy storage 1 and energy storage 2 are within the voltage range of the DC bus of the power converter. Energy storage 1 and energy storage 2 are each connected to a control system, and their protection parameters are stored in the control system. The control system controls the connection between the energy storage and the power converter via a transfer switch. The control system has two sets of virtual energy storage units that store the protection parameters and operation parameters of energy storage 1 and energy storage 2, respectively. When energy storage 1 is connected to the power conversion system, the control system enables the protection operation parameters of virtual energy storage unit 1 and realizes operation scheduling and protection of energy storage 1. When energy storage 2 is connected to the power conversion system, the control system enables the operation protection parameters of virtual energy storage unit 2 and realizes operation scheduling and protection of energy storage 2.

[0032] The electric energy storage medium and the capacitive energy storage medium are each connected to a power grid via an independent power conversion system, the electric energy storage medium and the capacitive energy storage medium are each connected to a control system, and the control system controls the electric energy storage medium, the capacitive energy storage medium, and the power converter through a communication connection and a changeover switch, the electric energy storage medium and the capacitive energy storage medium are connected to a DC bus of the power converter through a changeover switch, the DC voltage ranges of the electric energy storage medium and the capacitive energy storage medium are within the DC bus voltage range of the power converter, the electric energy storage medium and the energy storage medium are each connected to the control system, and the control system controls the electric energy storage medium, the capacitive energy storage medium, and the power converter through a switch, The control system includes two sets of virtual energy storage units, each of which stores protection parameters and operation parameters of an electric energy storage medium and a capacitive energy storage medium; When the storage-type energy storage medium is connected to the power conversion system, the control system enables a protection operation parameter of the virtual storage-type energy storage medium to realize operation scheduling and protection of the storage-type energy storage medium; When a capacitive energy storage medium is connected to the power conversion system, the control system enables the operation protection parameters of the virtual capacitive energy storage medium and realizes operation scheduling and protection of the capacitive energy storage medium, wherein the control system is configured to execute the method described in claim 1 by a processor of the control system.

[0033] In some embodiments, the electric energy storage medium includes the use of a nickel metal hydride battery, a lithium titanate battery, or a supercapacitor, and the capacitive energy storage medium includes the use of a lithium iron phosphate battery, a lead acid battery, or a lead carbon battery.

[0034] In some embodiments, the electric energy storage medium employs a high-power nickel-metal hydride battery suitable for low-temperature operation, and the capacitive energy storage medium employs a lithium iron phosphate battery to form a hybrid energy storage system.

[0035] In some embodiments, in application operating conditions where the ambient temperature is below 0°C for a long period of time, the heating loss of the lithium iron phosphate energy storage system is large, and the low temperature environment affects the cycle life of the battery. A hybrid energy storage system is constructed using high-power nickel-metal hydride batteries and lithium iron phosphate batteries suitable for low-temperature operation. The nickel-metal hydride battery pack and the lithium iron phosphate battery pack are installed in an enclosed space within a container and spaced apart. At low temperatures, the nickel-metal hydride battery pack is first charged and discharged, and the heat generated by the charge and discharge loss is used to heat the enclosed space and the lithium iron phosphate battery pack. After the ambient temperature in the sealed space has risen, charging and discharging of the lithium iron phosphate battery is resumed.

[0036] The present invention provides a hybrid energy storage system and a scheduling method thereof, and has the following beneficial technical effects:

[0037] 1. The amplitude and duration of the next scheduled power command are predicted based on the statistical characteristics of historical data, which greatly improves the controllability of power scheduling and allows the application's public power to be pre-configured, thereby improving the efficiency of power scheduling.

[0038] 2. By constructing training data, the accuracy of the algorithm is significantly improved. Specifically, 24 hours a day is divided into 288 five-minute time slots, FFT analysis is performed on the power commands of each time slot, and the amplitude-frequency characteristics of the power commands of each time slot are studied and summarized to construct a convolutional neural network algorithm, which significantly improves the prediction accuracy.

[0039] 3. Dynamically respond to scheduling commands based on a comparison of the predicted amplitude value with the maximum power of the electric energy storage medium and the power of the capacitive energy storage medium, and a comparison of the predicted duration with the discharge duration of the electric energy storage medium and the discharge duration of the capacitive energy storage medium, thereby greatly improving the flexibility and scalability of the system.

[0040] Although the method for immobilizing and associating electronic data evidence has been described in detail above, the present specification uses specific examples to explain the principles and embodiments of the present invention, and the description of the above examples is only used to help understand the core idea of ​​the present invention, and at the same time, those skilled in the art can make any changes in the specific embodiments and application scope according to the idea and method of the present invention. In short, the contents of the present specification should not be understood as limiting the present invention.

Claims

1. Step S1: detecting whether a scheduling command is received by a CPU in real time, and detecting and obtaining the maximum power of the power-type energy storage medium; Step S2: if the power value required by the scheduling command is greater than the maximum power of the power-type energy storage medium, turn on and replenish the capacitive energy storage medium; otherwise, respond to the scheduling command using the power-type energy storage medium; At the same time as performing step S2, input the current scheduling command request parameters, the operating parameters of the electric energy storage medium, and the operating parameters of the capacitive energy storage medium into the trained convolutional neural network, and output and predict the amplitude and duration of the next scheduling command through the convolutional neural network in step S3; If the amplitude value of the command output and predicted by the convolutional neural network is smaller than the maximum power of the power-type energy storage medium and the duration is shorter than the discharge duration of the power-type energy storage medium, respond to the scheduling command by the power-type energy storage medium, and if the predicted duration is longer than the discharge duration of the power-type energy storage medium, respond to the scheduling command by the direct capacitive energy storage medium; and step S4 of responding jointly to the scheduling command by both capacitive and electric energy storage media if the predicted amplitude value exceeds the maximum power to which the capacitive or electric energy storage media respond individually. A method for scheduling a hybrid energy storage system.

2. The electric energy storage medium includes the use of a nickel-metal hydride battery, a lithium titanate battery, or a supercapacitor, and the capacitive energy storage medium includes the use of a lithium iron phosphate battery, a lead acid battery, or a lead carbon battery.

10. The method of claim 1, wherein the scheduling of a hybrid energy storage system is performed in accordance with the present invention.

3. The scheduling command request parameters include a maximum power value, duration, and command issuance time required by the scheduling command; the operating parameters of the electric energy storage medium include a maximum power of the electric energy storage medium, a charge capacity of the electric energy storage medium, an operating duration of the electric energy storage medium, and a maximum current value of the electric energy storage medium; and the operating parameters of the capacitive energy storage medium include a maximum power of the capacitive energy storage medium, a charge capacity of the capacitive energy storage medium, an operating duration of the capacitive energy storage medium, and a maximum current value of the capacitive energy storage medium.

3. The method of claim 2, wherein the scheduling of the hybrid energy storage system is performed in accordance with claim 2.

4. The trained convolutional neural network is then completed through the following training steps: step S21: dividing 24 hours a day into 288 five-minute time slots; step S22: performing FFT transformation on the power command for each time slot; step S23: training the convolutional neural network on the amplitude-frequency characteristics of the power command for each time slot; and step S24: completing the training of the convolutional neural network.

3. The method of claim 2, wherein the scheduling of the hybrid energy storage system is performed in accordance with claim 2.

5. The convolutional neural network employs an improved loss function as V: [Equation 1] Among them, p ｉ represents the ratio of the number of times the i-th time period sample appears in the history samples to the total number of times all the history samples appear, and y ｉ is represented as a feature vector consisting of scheduling command request parameters, which include the maximum power value requested by the scheduling command, the scheduling command request duration, the command issuance time, and the command issuance button duration; N=288 represents the division of 24 hours into 288 5-minute time slots; s represents the maximum power value requested by the scheduling command; m represents the scheduling command request duration; and tan θ ｉ represents the tangent value between the i-th time period sample vector and the historical average standard feature vector.

10. The method of claim 1, wherein the scheduling of a hybrid energy storage system is performed in accordance with the present invention.

6. A hybrid energy storage scheduling system including a control system, a power conversion system, an electric energy storage medium, and a capacitive energy storage medium, The electric energy storage medium and the capacitive energy storage medium are connected to a power grid via independent power conversion systems, and the electric energy storage medium and the capacitive energy storage medium are connected to a control system, and the control system controls the electric energy storage medium, the capacitive energy storage medium, and the power converter through communication connections and selector switches; The control system includes two sets of virtual energy storage units, each storing protection parameters and operation parameters of an electric energy storage medium and a capacitive energy storage medium; When the storage-type energy storage medium is connected to the power conversion system, the control system enables a protection operation parameter of the virtual storage-type energy storage medium to realize operation scheduling and protection of the storage-type energy storage medium; When the capacitive energy storage medium is connected to the power conversion system, the control system enables the operation protection parameters of the virtual capacitive energy storage medium, and realizes the operation scheduling and protection of the capacitive energy storage medium, wherein the control system is configured to execute the method of claim 1 by a processor of the control system. A hybrid energy storage scheduling system comprising:

7. The electric energy storage medium includes the use of nickel metal hydride batteries, lithium titanate batteries, or supercapacitors, and the capacitive energy storage medium includes the use of lithium iron phosphate batteries, lead acid batteries, or lead carbon batteries.

7. The hybrid energy storage scheduling system of claim 6.

Images