Electric power parameter prediction method and device of photovoltaic power generation system

Abstract

The invention relates to the technical field of quantum computers, in particular to an electric power parameter prediction method and device for a photovoltaic power generation system, and the method comprises the steps: carrying out the sequential transmission of amplitude information between quantum time sequence convolutional networks corresponding to all moments; the electric power parameter prediction model has more excellent performance, and the performance of the electric power parameter prediction model on electric power parameter prediction is improved; the target bits in a plurality of quantum time sequence convolutional networks are measured at the same time, so that the quantum state of the target bits is projected to a specific ground state, complex gate operation is avoided, coherence and error accumulation are reduced, and the depth of a quantum circuit is reduced.

Description

Technical Field

[0001] This invention relates to the field of quantum computer technology, and in particular to a method and apparatus for predicting the power parameters of a photovoltaic power generation system. Background Technology

[0002] As the penetration rate of photovoltaic (PV) power generation continues to increase, the periodicity and randomness of its output fluctuations are also increasing, leading to more serious problems such as real-time power balance and power quality issues in the power system. Therefore, in order to adapt to the needs of high-proportion PV grid integration and meet the requirements of safe and stable operation of the distribution network, it is necessary to conduct accurate and reliable PV power forecasting for PV power plants, and then flexibly adjust their output to achieve grid supply and demand balance.

[0003] With the development of quantum technology, neural network models based on quantum technology have been applied. However, in the existing technology, the quantum-classical neural network prediction model built with quantum technology mainly uses a single feature input method to predict power parameters. The quantum circuit is deep and does not consider the relationship between local features among multiple features, resulting in the poor performance of the quantum time-series convolutional network model dominated by pure quantum parameters in power parameter prediction. Summary of the Invention

[0004] This invention provides a method and apparatus for predicting power parameters of a photovoltaic power generation system, which addresses the problem that existing technologies often fail to adequately consider the relationship between local features among multiple features, resulting in poor performance of quantum time-series convolutional networks dominated by pure quantum parameters in power parameter prediction and high depth of quantum circuits.

[0005] This specification provides an embodiment of a method for predicting the power parameters of a photovoltaic power generation system, including:

[0006] The power parameter prediction model and power parameters at several intervals in the photovoltaic power generation system are obtained; the power parameter prediction model includes a linear layer and a quantum temporal convolutional network corresponding to each time step, and each quantum temporal convolutional network includes an auxiliary bit for loading the running result and a target bit different from the auxiliary bit;

[0007] The electrical parameters at the first time step are input into the corresponding quantum sequential convolutional network; the probability of the target state of the target bit and the amplitude information of the auxiliary bit are measured; and after encoding the amplitude information into the auxiliary bit at the next time step, the electrical parameters at the next time step are input into the corresponding quantum sequential convolutional network.

[0008] After the probability of the target state of the target bit is returned until the probability of the target state of the target bit at all times is measured, the linear layer is used to map the probabilities of all the target states to the power parameters at the target time.

[0009] Optionally, the power parameters include multiple feature parameters, and the power parameter prediction model further includes several bits; the number of bits is greater than or equal to the number of types of feature parameters, and the number of bits satisfies 2. n Where n is a positive integer greater than 1.

[0010] Optionally, when the number of types of feature parameters is less than the number of bits, the target type feature parameter among the multiple feature parameters is multiplexed so that each bit encodes one feature parameter; wherein, the target type feature parameter is one of the multiple feature parameters.

[0011] Optionally, the quantum temporal convolutional network includes a coding layer, several subsequent processing units, and a variational entanglement layer connected in sequence. Each subsequent processing unit comprises a quantum convolutional layer connected to the coding layer and a quantum pooling layer connected to the quantum convolutional layer. The number of subsequent processing units increases logarithmically with the number of bits, as shown in the following formula:

[0012] y = log2x;

[0013] Where y is the number of subsequent processing units in the quantum temporal convolutional network; and x is the number of bits.

[0014] Optionally, the coding layer, the quantum convolutional layer, the quantum pooling layer, and the variational entanglement layer operate on all bits.

[0015] Optionally, when the number of subsequent processing units in the quantum temporal convolutional network is greater than 1, the coding layer, the variational entanglement layer, the quantum convolutional layer and the quantum pooling layer in the current subsequent processing unit connected to the coding layer act on all bits, and the number of bits acted on by the quantum convolutional layer and the quantum pooling layer in the next subsequent processing unit connected to the current subsequent processing unit is half the number of bits acted on by the quantum pooling layer in the current subsequent processing unit; wherein, the quantum convolutional layer and the quantum pooling layer act on at least two bits.

[0016] Optionally, before inputting the electrical parameters at the first time step into the corresponding quantum temporal convolutional network, the method further includes:

[0017] The quantum states of each quantum temporal convolutional network in the power parameter prediction model are initialized.

[0018] This specification also provides an embodiment of a power parameter prediction device for a photovoltaic power generation system, comprising:

[0019] An information acquisition module is used to acquire power parameter prediction models and power parameters at several intervals in a photovoltaic power generation system; the power parameter prediction model includes a linear layer and a quantum temporal convolutional network corresponding to each time step, and each quantum temporal convolutional network includes auxiliary bits for loading the running results and target bits different from the auxiliary bits;

[0020] The measurement module is used to input the power parameters at the first moment into the corresponding quantum temporal convolutional network; measure the probability of the target state of the target bit and the amplitude information of the auxiliary bit, and after encoding the amplitude information into the auxiliary bit at the next moment, input the power parameters at the next moment into the corresponding quantum temporal convolutional network.

[0021] The power parameter prediction module is used to return the probability of the target state of the target bit until the probability of the target state of the target bit at all times is measured, and then use the linear layer to map the probabilities of all the target states to the power parameters at the target time.

[0022] A quantum control system that uses the power parameter prediction method of the photovoltaic power generation system to perform quantum computing tasks, or the power parameter prediction device of the photovoltaic power generation system described above.

[0023] A quantum computer, comprising a quantum control system as described above.

[0024] A readable storage medium having a computer program stored thereon, which, when executed by a processor, enables the above-described method for predicting the power parameters of a photovoltaic power generation system.

[0025] Its beneficial effects are as follows: This application can not only process local time features by sequentially passing amplitude information between quantum temporal convolutional networks corresponding to each time step, but also make the power parameter prediction model have better performance and improve the performance of the power parameter prediction model in power parameter prediction; at the same time, the target bit in multiple quantum temporal convolutional networks is measured, so that the quantum state of the target bit is projected onto a specific ground state, avoiding complex gate operations, reducing coherence and error accumulation, and reducing the depth of quantum circuits. Attached Figure Description

[0026] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0027] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0028] Figure 1 A flowchart illustrating a method for predicting power parameters of a photovoltaic power generation system, as provided in the embodiments of this specification.

[0029] Figure 2 A schematic diagram of the relationships between multiple quantum temporal convolutional networks used in the power parameter prediction method for photovoltaic power generation systems, as provided in the embodiments of this specification.

[0030] Figure 3 A schematic diagram of the quantum circuitry of the coding layer in a four-bit quantum temporal convolutional network provided in the embodiments of this specification;

[0031] Figure 4 A schematic diagram of the quantum circuitry of the quantum convolutional layer in a four-qubit quantum temporal convolutional network provided in the embodiments of this specification;

[0032] Figure 5 A schematic diagram of the quantum circuitry for the quantum pooling layer in a four-qubit quantum temporal convolutional network provided in the embodiments of this specification;

[0033] Figure 6 A schematic diagram of the relationship between multiple quantum temporal convolutional networks for another method of predicting power parameters in a photovoltaic power generation system, provided in an embodiment of this specification.

[0034] Figure 7 This is a schematic diagram of the structure of a power parameter prediction device for a photovoltaic power generation system provided in the embodiments of this specification;

[0035] Figure 8 This is a schematic diagram of a computer-readable medium provided for embodiments of this specification. Detailed Implementation

[0036] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0037] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0038] It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of the invention.

[0039] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use.

[0040] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0041] In all the examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0042] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0043] Reference Figure 1 The flowchart of a method for predicting power parameters of a photovoltaic power generation system provided in the embodiments of this specification includes: S101: obtaining a power parameter prediction model and power parameters of the photovoltaic power generation system at several intervals; the power parameter prediction model includes a linear layer and a quantum temporal convolutional network corresponding to each time step, and each quantum temporal convolutional network includes an auxiliary bit for loading the running result and a target bit different from the auxiliary bit.

[0044] In one optional embodiment, before predicting the power parameters of a photovoltaic power generation system, a trained power parameter prediction model needs to be obtained. Taking power generation as an example, an initial power parameter prediction model is first constructed. Specifically, the number of bits in the initial power parameter prediction model is determined based on the number of types of feature parameters in the power parameters, ensuring that each feature parameter can be encoded onto one bit. Assuming that the types of feature parameters in the power parameters include four types—local total radiation, local diffuse radiation, local temperature, and power generation—then the number of bits in the initial power parameter prediction model is 4. Simultaneously, the number of bits should be greater than or equal to the number of types of feature parameters, and the number of bits should satisfy 2. n Where n is a positive integer greater than 1.

[0045] Assuming that the power parameters include three types of characteristic parameters: local total radiation, local diffuse radiation, and power generation, the initial power parameter prediction model should also have 4 bits to meet the bit requirement. Since the three characteristic parameters only need to be encoded into three bits, to ensure the number of bits is met, the target type of characteristic parameter among the multiple characteristic parameters is multiplexed so that each bit encodes one characteristic parameter. For example, local total radiation, local diffuse radiation, and power generation are encoded into three bits, and power generation is multiplexed once into the remaining bit. That is, when the number of characteristic parameter types is less than the number of bits, the target type of characteristic parameter among the multiple characteristic parameters is multiplexed so that each bit encodes one characteristic parameter; where the target type of characteristic parameter is one of the multiple characteristic parameters. Feature completion can integrate more feature information, allowing the quantum convolutional layer in the initial power parameter prediction model to learn local features.

[0046] After determining the number of bits in the initial power parameter prediction model, it is necessary to determine the number of time points with the same time interval. Taking four time points as an example, an initial power parameter prediction model is constructed, which includes a linear layer and a quantum temporal convolutional network corresponding to each time point. Figure 2 As shown, each of the quantum temporal convolutional networks includes a coding layer, several subsequent processing units, and a variational entanglement layer connected in sequence. Each subsequent processing unit comprises a quantum convolutional layer connected to the coding layer and a quantum pooling layer connected to the quantum convolutional layer. The number of subsequent processing units increases logarithmically with the number of bits, as shown by the following formula:

[0047] y = log2x;

[0048] Where y is the number of subsequent processing units in the quantum temporal convolutional network; and x is the number of bits.

[0049] like Figure 2 As shown, taking power generation and four time points as examples in power parameters, the encoding layer, quantum convolutional layer, quantum pooling layer, and variational entanglement layer act on all bits. This layered approach helps avoid getting trapped in local minima during the optimization of the initial power parameter prediction model, thus reducing the impact of low efficiency. Measurements of auxiliary bits q2 and q3 provide amplitude information h1 to the auxiliary bits in the quantum temporal convolutional network corresponding to the next time point, enabling quantum state initialization in the network. Measurements of the target bit q0 at the current time yield the probability y1 that the quantum state of the target bit q0 is the |1> state, providing data support for subsequent power generation prediction.

[0050] like Figure 3As shown, each bit of the encoding layer in a quantum temporal convolutional network has an H logic gate and an RY logic gate to encode various feature parameters.

[0051] like Figure 4 As shown, the quantum convolutional layer in the quantum temporal convolutional network contains several quantum circuit layers. Each quantum circuit layer consists of RZ logic gates, RY logic gates, and CNOT logic gates, which are used to extract the local feature relationships of various feature parameters at the current time.

[0052] like Figure 5 As shown, the quantum pooling layer in the quantum sequential convolutional network also contains several quantum circuit layers. Each quantum circuit layer consists of RZ logic gates, RY logic gates, and CNOT logic gates, which are used to merge the information on four bits into two bits, reducing the measurement of subsequent bits.

[0053] like Figure 6 As shown, taking the power generation and four time points in the power parameters as an example, when the number of subsequent processing units in the quantum temporal convolutional network is greater than 1, the coding layer, the variational entanglement layer, the quantum convolutional layer and the quantum pooling layer in the current subsequent processing unit connected to the coding layer act on all bits. The number of bits acted on by the quantum convolutional layer and the quantum pooling layer in the next subsequent processing unit connected to the current subsequent processing unit is half the number of bits acted on by the quantum pooling layer in the current subsequent processing unit, that is... Figure 6 Bits q2 and q3 are defined in the quantum convolutional network; wherein the quantum convolutional layer and the quantum pooling layer operate on at least two bits. For example, if the quantum convolutional layer and quantum pooling layer in the current subsequent processing unit operate on four bits, then the quantum convolutional layer and quantum pooling layer in the next subsequent processing unit will operate on two bits. By applying quantum convolutional layers and quantum pooling layers multiple times, the local feature relationships of the extracted multiple feature parameters become more stable and the information becomes more concentrated, reducing the number of measurements for subsequent bits. At the same time, the parameterized logic gates of the above-mentioned quantum temporal convolutional network have fewer parameters, which helps to alleviate the plateau problem. Fewer parameters mean a smaller optimization space and a lower probability of gradient vanishing, enabling the power parameter prediction model to have a stable convergence process.

[0054] After constructing an initial power parameter prediction model, we will take the power generation and four characteristic parameters at four different times as examples to illustrate the process. We will obtain multiple sets of historical power parameters with the same time interval, such as the historical power generation at the first set of four times: p1, p2, p3, p4, and the historical direct irradiance at the four times: I1, I2, I4. 3. I4, historical scattered irradiance at four times: J1, J2, J3, J4; historical local temperatures at four times: m1, m2, m3, m4; and historical power generation at the fifth time: p5. Then, the initial power parameter prediction model is initialized with quantum states, and the parameters of the parameterized logic gates on the variational entanglement layer, quantum convolution layer, and quantum pooling layer in the initial power parameter prediction model are initialized. During initialization, the historical power generation, historical direct irradiance, historical scattered irradiance, and historical local temperature at the first time of the first group are input into the quantum temporal convolutional network corresponding to the first time of the initial power parameter prediction model for encoding and quantum state evolution. The probability of the target state of the target bit is measured, and the amplitude information of the auxiliary bit is measured. The amplitude information is encoded and then sent to the next... After the auxiliary bit at one time step, the historical power generation, historical direct irradiance, historical diffuse irradiance, and historical local temperature at the next time step are input into the corresponding quantum temporal convolutional network. The network returns the probability of the target state of the target bit until the probability of the target state of the target bit at all corresponding time steps is measured. Then, the linear layer maps the probabilities of all the target states to the predicted power generation at the fifth time step. Finally, the loss function is calculated based on the predicted power generation at the fifth time step and the historical power generation at the fifth time step p5. Backpropagation is performed based on the loss function value to determine the parameter gradient of the parameterized quantum logic gates in the initial power parameter prediction model. Based on the parameter gradient, the parameters of the parameterized quantum logic gates in the initial power parameter prediction model are updated using a classical optimization algorithm, thus completing one model training. The model is trained using multiple sets of historical power generation, historical direct irradiance, historical diffuse irradiance, and historical local temperature data at four time steps until the preset number of iterations is reached or the loss function converges, at which point the model training stops, resulting in a trained power parameter prediction model. A well-trained power parameter prediction model is used to predict future power generation. This model considers various time-series characteristic parameters, exhibiting better stability and accuracy compared to existing power generation prediction technologies. Furthermore, it allows for effective model training with limited data, reducing reliance on large datasets during optimization and mitigating the risk of problems related to low-yield power generation. The target state can be defined as the |1> state.

[0055] After obtaining the trained power parameter prediction model, it is also necessary to obtain the target time of the preset power generation and the power generation, direct irradiance, diffuse irradiance, and local temperature at the four times with the same time interval as the target time. Assuming that the time interval of the power parameters used in the initial power parameter prediction model is 1 hour, if the user wants to predict the power generation of the photovoltaic power generation system at 17:00, then it is necessary to obtain the power generation, direct irradiance, diffuse irradiance, and local temperature at 13:00, 14:00, 15:00, and 16:00.

[0056] 102: Input the power parameters at the first time step into the corresponding quantum temporal convolutional network; measure the probability of the target state of the target bit and the amplitude information of the auxiliary bit, and after encoding the amplitude information into the auxiliary bit at the next time step, input the power parameters at the next time step into the corresponding quantum temporal convolutional network; S103: Return the probability of the target state of the target bit until the probability of the target state of the target bit at all time steps is measured, and then use the linear layer to map the probabilities of all the target states into the power parameters at the target time step.

[0057] In one optional embodiment, assuming that the power generation, direct irradiance, diffuse irradiance, and local temperature at 13:00, 14:00, 15:00, and 16:00 have been obtained, and the user wants to predict the power generation of the photovoltaic power generation system at 17:00, firstly, the power generation, direct irradiance, diffuse irradiance, and local temperature at 13:00 are input into the corresponding quantum temporal convolutional network to measure the probability of the target state of the target bit and the amplitude information of the auxiliary bit. After encoding the amplitude information into the auxiliary bit at 14:00, the power generation, direct irradiance, diffuse irradiance, and local temperature at 14:00 are input into the corresponding quantum temporal convolutional network to return the probability of the |1> state of the target bit until the probability of the |1> state of the target bit at the corresponding four times is measured. Then, the linear layer is used to map the probabilities of all |1> states to the power generation at 17:00.

[0058] This application not only processes local time-series features by sequentially passing amplitude information between quantum temporal convolutional networks corresponding to each time step, but also enables the power parameter prediction model to have better performance and improves the performance of the power parameter prediction model in power parameter prediction. At the same time, the target bit in multiple quantum temporal convolutional networks is measured so that the quantum state of the target bit is projected onto a specific ground state, avoiding complex gate operations, reducing coherence and error accumulation, and reducing the depth of quantum circuits.

[0059] In one optional embodiment, in order to ensure the prediction accuracy of the power parameter prediction model, when using the power parameter prediction model, the quantum states of each quantum temporal convolutional network in the power parameter prediction model need to be initialized before inputting the power parameters at the first time moment into the corresponding quantum temporal convolutional network.

[0060] Reference Figure 7 This specification also provides an embodiment of a power parameter prediction device for a photovoltaic power generation system, comprising:

[0061] The information acquisition module 201 is used to acquire the power parameter prediction model and the power parameters of the photovoltaic power generation system at several intervals. The power parameter prediction model includes a linear layer and a quantum temporal convolutional network corresponding to each time step. Each quantum temporal convolutional network includes an auxiliary bit for loading the running result and a target bit different from the auxiliary bit.

[0062] The measurement module 202 is used to input the power parameters at the first moment into the corresponding quantum temporal convolutional network; measure the probability of the target state of the target bit and measure the amplitude information of the auxiliary bit, and after encoding the amplitude information into the auxiliary bit at the next moment, input the power parameters at the next moment into the corresponding quantum temporal convolutional network.

[0063] The power parameter prediction module 203 is used to return the probability of the target state of the target bit until the probability of the target state of the target bit at all times is measured, and then use the linear layer to map the probability of all the target states to the power parameters at the target time.

[0064] Optionally, the device further includes:

[0065] An initialization module is used to initialize the quantum states of each quantum temporal convolutional network in the power parameter prediction model.

[0066] Regarding the apparatus in the above embodiments, the process of performing each step has been described in detail in the embodiments of the method, and will not be elaborated here.

[0067] A quantum control system that uses the power parameter prediction method of the photovoltaic power generation system to perform quantum computing tasks, or the power parameter prediction device of the photovoltaic power generation system described above.

[0068] A quantum computer, comprising a quantum control system as described above.

[0069] Reference Figure 8 This is a schematic diagram of a computer-readable medium provided for embodiments of this specification.

[0070] accomplish Figure 1The computer instructions of the method shown can be stored on one or more computer-readable media. A computer-readable medium can be a readable signal medium or a readable storage medium. A readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0071] The computer-readable storage medium may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, capable of transmitting, propagating, or transmitting a program for use by or in connection with an instruction execution device, apparatus, or apparatus. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0072] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0073] In summary, this invention can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that in practice, general-purpose data processing devices such as microprocessors or digital signal processors (DSPs) can be used to implement some or all of the functions of some or all of the components according to the embodiments of the invention. The invention can also be implemented as a device or apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such programs implementing the invention can be stored on a computer-readable medium or can take the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0074] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the present invention is not inherently related to any specific computer, virtual device, or electronic device, and various general-purpose devices can also implement the present invention. The above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0075] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0076] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of this application should be included within the scope of the claims of this application.

Claims

1. A method for predicting the power parameters of a photovoltaic power generation system, characterized in that, include: The power parameter prediction model and power parameters at several intervals in the photovoltaic power generation system are obtained; the power parameter prediction model includes a linear layer and a quantum temporal convolutional network corresponding to each time step, and each quantum temporal convolutional network includes an auxiliary bit for loading the running result and a target bit different from the auxiliary bit; The electrical parameters at the first time step are input into the corresponding quantum sequential convolutional network; the probability of the target state of the target bit and the amplitude information of the auxiliary bit are measured; and after encoding the amplitude information into the auxiliary bit at the next time step, the electrical parameters at the next time step are input into the corresponding quantum sequential convolutional network. After the probability of the target state of the target bit is returned until the probability of the target state of the target bit at all times is measured, the linear layer is used to map the probabilities of all the target states to the power parameters at the target time.

2. The method as described in claim 1, characterized in that, The power parameters include multiple feature parameters, and the power parameter prediction model also includes several bits; the number of bits is greater than or equal to the number of types of feature parameters, and the number of bits satisfies 2. n Where n is a positive integer greater than 1.

3. The method as described in claim 1, characterized in that, When the number of types of feature parameters is less than the number of bits, the target type feature parameter among the multiple feature parameters is multiplexed so that each bit encodes one feature parameter; wherein, the target type feature parameter is one of the multiple feature parameters.

4. The method as described in claim 2, characterized in that, The quantum temporal convolutional network comprises a coding layer, several subsequent processing units, and a variational entanglement layer connected in sequence. Each subsequent processing unit includes a quantum convolutional layer connected to the coding layer and a quantum pooling layer connected to the quantum convolutional layer. The number of subsequent processing units increases logarithmically with the number of bits, as shown in the following formula: y = log2x; Where y is the number of subsequent processing units in the quantum temporal convolutional network; and x is the number of bits.

5. The method as described in claim 4, characterized in that, The coding layer, the quantum convolutional layer, the quantum pooling layer, and the variational entanglement layer operate on all bits.

6. The method as described in claim 4, characterized in that, When the number of subsequent processing units in the quantum temporal convolutional network is greater than 1, the coding layer, the variational entanglement layer, the quantum convolutional layer and the quantum pooling layer in the current subsequent processing unit connected to the coding layer act on all bits. The number of bits acted on by the quantum convolutional layer and the quantum pooling layer in the next subsequent processing unit connected to the current subsequent processing unit is half the number of bits acted on by the quantum pooling layer in the current subsequent processing unit; wherein, the quantum convolutional layer and the quantum pooling layer act on at least two bits.

7. The method as described in claim 1, characterized in that, Before inputting the power parameters at the first moment into the corresponding quantum temporal convolutional network, the method further includes: The quantum states of each quantum temporal convolutional network in the power parameter prediction model are initialized.

8. A power parameter prediction device for a photovoltaic power generation system, characterized in that, include: An information acquisition module is used to acquire power parameter prediction models and power parameters at several intervals in a photovoltaic power generation system; the power parameter prediction model includes a linear layer and a quantum temporal convolutional network corresponding to each time step, and each quantum temporal convolutional network includes auxiliary bits for loading the running results and target bits different from the auxiliary bits; The measurement module is used to input the power parameters at the first moment into the corresponding quantum temporal convolutional network; measure the probability of the target state of the target bit and the amplitude information of the auxiliary bit, and after encoding the amplitude information into the auxiliary bit at the next moment, input the power parameters at the next moment into the corresponding quantum temporal convolutional network. The power parameter prediction module is used to return the probability of the target state of the target bit until the probability of the target state of the target bit at all times is measured, and then use the linear layer to map the probabilities of all the target states to the power parameters at the target time.

9. A quantum control system, characterized in that, The execution of quantum computing tasks is carried out using the power parameter prediction method of the photovoltaic power generation system as described in any one of claims 1-7, or includes the power parameter prediction device of the photovoltaic power generation system as described in claim 8.

10. A quantum computer, characterized in that, Including the quantum control system as described in claim 9.

11. A readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it can implement the power parameter prediction method for the photovoltaic power generation system as described in any one of claims 1 to 7.

Images