Method and device for determining electricity price combination, storage medium and computer program product

Abstract

The application discloses a method and device for determining an electricity price combination, a storage medium and a computer program product, and relates to the electric power field. The method comprises the following steps: sending first ciphertext obtained by encrypting an initial electricity price combination to a server; obtaining second ciphertext returned by the server, and obtaining first electricity price combination information according to the second ciphertext; inputting the first electricity price combination and a reference information set into a target self-loop network to obtain target electricity price information, wherein the target electricity price information comprises at least one second power generation amount and a clearing electricity price probability distribution respectively matched with the at least one second power generation amount; and determining a target electricity price combination according to the target electricity price information, wherein the target electricity price combination comprises a plurality of target power generation amounts and target electricity prices respectively matched with the plurality of target power generation amounts. The above technical scheme solves the problem of inaccurate electricity price prediction in the prior art.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electric power, in particular to a method and device for determining an electricity price combination, a storage medium and a computer program product. BACKGROUND

[0002] In the current electricity price prediction process, the generation and sales data are used to fit the generation price curve, and the market equilibrium point is solved according to the predicted generation price curve and the predicted value of system load, and then the predicted value of the time-of-use spot electricity price is obtained.

[0003] However, as the core secret of each company, it is very difficult to share the generation and sales data and break down the barriers between data sources. For a long time, the data of each company exists in the form of an island. In this case, the electricity price prediction model can only predict the electricity price based on the data of the company, but the data of the company cannot cover all the generation and sales data in the current electricity market. Therefore, the predicted result lacks accuracy. That is, in the prior art, there is a technical problem of inaccurate electricity price prediction.

[0004] At present, there is no effective solution to the above technical problems. SUMMARY

[0005] The embodiments of the present application provide a method and device for determining an electricity price combination, a storage medium and a computer program product to at least solve the problem of inaccurate electricity price prediction in the related art.

[0006] According to an aspect of the embodiments of the present application, a method for determining an electricity price combination is provided, comprising: sending a first ciphertext obtained by encrypting an initial electricity price combination to a server, wherein the initial electricity price combination comprises a plurality of initial generation capacities and initial electricity prices respectively matched with the plurality of initial generation capacities; obtaining a second ciphertext returned by the server, and obtaining first electricity price combination information according to the second ciphertext, wherein the server is configured to receive a plurality of first ciphertexts respectively sent by a plurality of nodes, and model based on the plurality of first ciphertexts, the second ciphertext is determined by the server based on the modeling result, and the first electricity price combination comprises a first output curve and first electricity prices respectively corresponding to a plurality of generation capacities in the first output curve; inputting the first electricity price combination and a reference information set into a target self-loop network to obtain target electricity price information, wherein the target electricity price information comprises at least one second generation capacity and a clearing electricity price probability distribution respectively matched with the at least one second generation capacity; determining a target electricity price combination according to the target electricity price information, wherein the target electricity price combination comprises a plurality of target generation capacities and target electricity prices respectively matched with the plurality of target generation capacities.

[0007] In an example embodiment, before sending the first ciphertext obtained by encrypting the initial electricity price combination to the server, further comprising: determining, according to the at least one generator set and the start-stop plan of each of the at least one generator set, a unit operation capacity curve of each of the at least one generator set matching the target period; determining, according to the at least one generator set, reference electricity price information corresponding to the target period from the historical electricity price combination corresponding to each of the at least one historical period, wherein the reference electricity price information comprises reference electricity prices corresponding to a plurality of unit load rates; and determining the initial electricity price combination based on the unit operation capacity curve and the reference electricity price information.

[0008] In an example embodiment, before determining, according to the at least one generator set, the reference electricity price information corresponding to the target period from the historical electricity price combination corresponding to each of the at least one historical period, further comprising: obtaining period description characteristics of the target period, wherein the period description characteristics are used to indicate period description information of the target period, and the period description information comprises at least one of the following: week type information, geographical environment information, weather information, lunar solar term information, and holiday information; and determining the at least one reference historical period as the at least one historical period in a case where a feature similarity between the period description characteristics and reference period description characteristics of each of the at least one reference historical period satisfies a similarity condition.

[0009] In an example embodiment, determining, according to the at least one generator set, the reference electricity price information corresponding to the target period from the historical electricity price combination corresponding to each of the at least one historical period, comprises: obtaining the historical electricity price combination corresponding to each of the at least one historical period for the at least one generator set; determining a weighted summation result between the at least one historical electricity price combination according to a reference weight matched by each of the at least one historical period; and determining the weighted summation result as the reference electricity price information corresponding to the target period.

[0010] In an example embodiment, sending the first ciphertext obtained by encrypting the initial electricity price combination to the server comprises: obtaining second ciphertext returned by the server, and obtaining the first electricity price combination information according to the second ciphertext, comprising:

[0011] In an example embodiment, inputting the first electricity price combination and the reference information set into the target self-loop network to obtain target electricity price information includes: obtaining the reference information set, wherein the reference information set includes at least one of the following: reference historical electricity price data, historical disclosure data, weather information, holiday information, coal price information, current node historical unit operation capacity information, and supply and demand description information, wherein the supply and demand description information is used to indicate the predicted tension degree of power supply; determining a target feature vector sequence according to the first electricity price combination and the reference information set; inputting the target feature vector sequence into the target self-loop network, wherein the target self-loop network includes a plurality of sequentially connected feature processing modules, a current feature processing module in the plurality of feature processing modules is used to receive historical output results of a previous feature processing module as a first input, and receive a current feature vector matched with a current time node in the target feature vector sequence as a second input, and the current feature processing module is further used to perform likelihood processing on a reference result determined based on the first input and the second input to obtain a current output result; and determining the target electricity price information according to an output result of a last feature processing module in the plurality of feature processing modules.

[0012] In an example embodiment, determining the target electricity price combination according to the target electricity price information includes: obtaining a current power generation amount from at least one second power generation amount, and a current electricity price probability distribution matched with the current power generation amount; determining a reference electricity price matched with the current power generation amount according to a distribution mean of the current electricity price probability distribution; and adjusting the reference electricity price according to a distribution feature of the current electricity price probability distribution to obtain a target electricity price in the target electricity price combination.

[0013] According to another aspect of the embodiments of the present application, a device for determining a power price combination is provided, comprising: a sending unit configured to send a first ciphertext obtained by encrypting an initial power price combination to a server, wherein the initial power price combination comprises a plurality of initial power generation amounts and initial power prices respectively matched with the initial power generation amounts; an analyzing unit configured to obtain a second ciphertext returned by the server, and analyze the first power price combination information according to the second ciphertext, wherein the server is configured to receive a plurality of first ciphertexts respectively sent by a plurality of nodes, and model based on the plurality of first ciphertexts, the second ciphertext is determined by the server according to a modeling result, and the first power price combination comprises a first output curve and first power prices respectively corresponding to a plurality of power generation amounts in the first output curve; an input unit configured to input the first power price combination and a reference information set into a target self-loop network to obtain target power price information, wherein the target power price information comprises at least one second power generation amount and a clearing power price probability distribution respectively matched with the at least one second power generation amount; and a determining unit configured to determine a target power price combination according to the target power price information, wherein the target power price combination comprises a plurality of target power generation amounts and target power prices respectively matched with the target power generation amounts. According to still another aspect of the embodiments of the present application, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program, wherein the computer program is configured to execute the above method for determining a power price combination when running.

[0014] According to still another aspect of the embodiments of the present application, a computer readable storage medium is provided, and the computer readable storage medium stores a program, wherein the program is configured to execute the steps of the method in the embodiments of the present application when running.

[0015] According to still another aspect of the embodiments of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the above method for determining a power price combination through the computer program.

[0016] According to still another aspect of the embodiments of the present application, a computer program product is provided, comprising a computer program, wherein the computer program is configured to execute the steps of the method in the embodiments of the present application when executed by a processor.

[0017] By the present application, the first ciphertext obtained by encrypting the initial electricity price combination can be sent to the server, wherein the initial electricity price combination includes a plurality of initial electricity generation amounts and initial electricity prices matched with the plurality of initial electricity generation amounts respectively; then the second ciphertext returned by the server is obtained, and the first electricity price combination information is obtained according to the second ciphertext, wherein the server is used for receiving the first ciphertext sent by a plurality of nodes respectively, and modeling based on the plurality of first ciphertexts, the second ciphertext is determined by the server according to the modeling result, and the first electricity price combination includes a first output curve and a first electricity price corresponding to each of a plurality of electricity generation amounts in the first output curve; further, the first electricity price combination and the reference information set are input into the target self-loop network to obtain target electricity price information, wherein the target electricity price information includes at least one second electricity generation amount and a clearing electricity price probability distribution matched with the at least one second electricity generation amount respectively; then the target electricity price combination can be determined according to the target electricity price information, wherein the target electricity price combination includes a plurality of target electricity generation amounts and target electricity prices matched with the plurality of target electricity generation amounts respectively, thereby solving the problem of inaccurate electricity price prediction in the prior art. BRIEF DESCRIPTION OF DRAWINGS

[0018] The accompanying drawings, which are incorporated into and form a part of the specification, illustrate an embodiment consistent with the present application and, together with the description, serve to explain the principles of the application.

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, for those skilled in the art, other drawings can also be obtained based on these drawings without any creative labor.

[0020] Figure 1 is a hardware structure block diagram of a computer terminal of a method for determining an electricity price combination according to an embodiment of the present application;

[0021] Figure 2 is a flowchart of a method for determining an electricity price combination according to an embodiment of the present application;

[0022] Figure 3 is a flowchart of another method for determining an electricity price combination according to an embodiment of the present application;

[0023] Figure 4 is a schematic diagram of a method for determining an electricity price combination according to an embodiment of the present application;

[0024] Figure 5 is a flowchart of still another method for determining an electricity price combination according to an embodiment of the present application;

[0025] Figure 6 is a flowchart of still another method for determining an electricity price combination according to an embodiment of the present application;

[0026] Figure 7 is a schematic diagram of another method for determining an electricity price combination according to an embodiment of the present application;

[0027] Figure 8 This is a structural block diagram of a device for determining an electricity price combination according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0029] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0030] The method embodiments provided in the embodiments of the present application can be executed in a computer terminal or similar computing device. Taking running on a computer terminal as an example, Figure 1 This is a hardware structure block diagram of a computer terminal for determining a power price combination according to an embodiment of the present application. Figure 1 As shown, the computer terminal may include one or more ( Figure 1 Only one is shown) a processor 102 (the processor 102 may include but is not limited to a microprocessor (Central Processing Unit, MCU) or a programmable logic device (Field Programmable Gate Array, FPGA) and a memory 104 for storing data. The computer terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It will be understood by those skilled in the art that Figure 1The illustrated structure is only a schematic, which does not limit the structure of the computer terminal. For example, the computer terminal can further include more or less components than those shown, or have a different configuration or arrangement of the components. Figure 1 The illustrated structure is only a schematic, which does not limit the structure of the computer terminal. For example, the computer terminal can further include more or less components than those shown, or have a different configuration or arrangement of the components. Figure 1 The illustrated structure is only a schematic, which does not limit the structure of the computer terminal. For example, the computer terminal can further include more or less components than those shown, or have a different configuration or arrangement of the components.

[0031] The memory 104 can be used to store computer programs, such as software programs of application software and modules, such as the computer program corresponding to the method for determining the electricity price combination in the embodiments of the present application. The processor 102 executes various functional applications and data processing by running the computer programs stored in the memory 104, that is, implements the above method. The memory 104 can include a high-speed random access memory, and can further include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some examples, the memory 104 can further include a memory remotely arranged with respect to the processor 102, which can be connected to the computer terminal through a network. Examples of the above network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0032] The computer terminal can be connected to the wireless network provided by a communication service provider. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC for short), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (Radio Frequency, RF for short) module, which is used to communicate with the Internet in a wireless manner.

[0033] In the embodiments of the present application, a method for determining an electricity price combination is provided, Figure 2 The flow chart of the method for determining an electricity price combination according to the embodiments of the present application is shown in FIG. 2, which includes the following steps: Figure 2

[0034] S202, sending a first ciphertext obtained by encrypting an initial electricity price combination to a server, wherein the initial electricity price combination includes a plurality of initial power generation amounts and initial electricity prices respectively matched with the plurality of initial power generation amounts;

[0035] S204, obtaining a second ciphertext returned by the server, and obtaining first electricity price combination information according to the second ciphertext, wherein the server is used to receive the first ciphertext respectively sent by a plurality of nodes, and model based on the plurality of first ciphertexts, the second ciphertext is determined by the server according to a modeling result, and the first electricity price combination includes a first output curve and first electricity prices respectively corresponding to a plurality of power generation amounts in the first output curve;

[0036] ​S206, input the first electricity price combination and the reference information set into the target self-loop network to obtain target electricity price information, wherein the target electricity price information comprises at least one second generation amount and a clearing electricity price probability distribution respectively matched with the at least one second generation amount;

[0037] S208, determine the target electricity price combination according to the target electricity price information, wherein the target electricity price combination comprises a plurality of target generation amounts and target electricity prices respectively matched with the plurality of target generation amounts.

[0038] It should be noted that in the above step S202, the initial electricity price combination is a plurality of initial generation amounts and corresponding electricity prices. It can be understood that different generation amounts correspond to different prices for electricity. In actual power generation, the power generation capacity and load of a generator set are usually constant within a period of time. Further, the maximum power generation capacity of a generator set is constant, but the generator set can work at different load rates, for example, at a load rate of 0.4, the actual power generation capacity of the set is 40% of the maximum power generation capacity. Different load rates correspond to different electricity prices. The initial electricity price combination is a combination of a plurality of different load rates and corresponding electricity prices.

[0039] Optionally, the initial electricity price combination can be obtained through the following steps S2001 to S2003.

[0040] S2001, determine the unit operation capacity curve of each of the at least one generator set matched with the target period according to the at least one generator set and the start-stop plan of each of the at least one generator set;

[0041] S2002, determine the reference electricity price information corresponding to the target period according to the at least one generator set and the historical electricity price combination corresponding to each of the at least one historical period, wherein the reference electricity price information comprises reference electricity prices respectively corresponding to a plurality of unit load rates;

[0042] S2003, determine the initial electricity price combination based on the unit operation capacity curve and the reference electricity price information.

[0043] It should be noted that in the above steps S2001 to S2003, the unit operation capacity refers to the power generation capacity of the unit during operation. For the above step S2001, in the electricity market, each electricity price prediction period includes a plurality of sub-periods, each of which can have different predicted power generation capacities. The specific predicted power generation capacity can be determined according to whether the generator set has a start-stop plan.

[0044] For example, a prediction day of an electricity price contains 96 sub-periods, each of which is 15 minutes. In this case, in the process of determining the initial electricity price combination, if the unit has a start-up and shutdown plan, the real-time power generation of the running and standby unit is set to increase the adjustable total power by 1 / 8 every 15 minutes from the start-up time until the predicted power generation reaches the maximum power generation of the unit. If the unit has no start-up and shutdown plan, the unit will maintain the current running power. Then the unit running capacity of the unit in the 96 sub-periods can be determined, and the unit running capacity of the 96 sub-periods and the 96 time sub-periods form the above-mentioned unit capacity operation curve, which can take time as the horizontal coordinate and power as the vertical coordinate.

[0045] It can be understood that the determined unit running capacity of the unit in the 96 sub-periods is the maximum power generation of the unit in the 96 sub-periods. After obtaining the unit running capacity curve, the above-mentioned step S2002 is performed, and reference price information corresponding to the target period is determined according to at least one power generation unit and at least one historical period corresponding to the historical electricity price combination of the respective reference historical period, wherein the reference price information includes reference prices corresponding to the load rates of the plurality of units respectively.

[0046] Specifically, the historical period in the above-mentioned step S2002 can be determined by the following steps S3001 to S3002.

[0047] S3001, obtaining the period description characteristics of the target period, wherein the period description characteristics are used to indicate the period description information of the target period, and the period description information includes at least one of the following: week type information, geographical environment information, weather information, lunar solar term information, and holiday information;

[0048] S3002, in the case that the feature similarity between the period description characteristics and the respective reference period description characteristics of the at least one reference historical period satisfies the similarity condition, the at least one reference historical period is determined as the at least one historical period.

[0049] It can be understood that the above-mentioned target period is a time period to be predicted for the electricity price, for example, when the period to be predicted for the electricity price is tomorrow 0 o'clock to 24 o'clock, each 0 o'clock to 24 o'clock in the past day is a reference historical period, the period description characteristics of the above-mentioned target period and the reference historical period are obtained, including week type information, geographical environment information, weather information, lunar solar term information, and holiday information, and different period description characteristics correspond to a weight.

[0050] For example, for holiday information, if the reference historical period is the historical period of the same day of the target period last year, the weight of this item is 1, and if it is different, the weight is 0. Finally, the total weight of each item is added, and the five reference historical periods with the highest total weight are determined as the historical periods. It should be noted that this is only an example, and in practice, more or fewer reference historical periods can be determined as historical periods.

[0051] After that, the reference price information can be determined by weighting the historical price combination corresponding to each of the determined at least one historical period. For specific weighting process, refer to steps S4001 to S4003.

[0052] S4001, obtaining at least one generator set, in the historical price combination corresponding to each of the at least one historical period;

[0053] S4002, determining the weighted sum result between the at least one historical price combination according to the reference weight matched by each of the at least one historical period;

[0054] S4003, determining the weighted sum result as the reference price information corresponding to the target period.

[0055] For example, the reference price information can be determined by average weighting. In this case, assuming there are three historical periods, and the respective prices of the three historical periods are 700 yuan, 800 yuan and 900 yuan respectively under the load rate of 0.4, the average weighted price corresponding to the load rate of 0.4 is 800 yuan. The average weighted prices under other load rates can be calculated in the same way, which are the reference prices corresponding to the load rates of the generator sets in step S2002.

[0056] In addition, non-average weighting can also be performed. For example, the above three historical periods are sorted in order of similarity from small to large and are respectively assigned weights of 0.5, 0.8 and 1.0. The respective prices of the three historical periods are 700 yuan, 800 yuan and 900 yuan respectively under the load rate of 0.4. The non-average weighted price corresponding to the load rate of 0.4 is (700*0.5+800*0.8+900*1) / 3=630 yuan. The non-average weighted prices under other load rates can be calculated in the same way, which are the reference prices corresponding to the load rates of the generator sets in step S2002.

[0057] It is worth noting that at this time, the reference price information obtained has multiple continuous load rates with the same price, for example, when the load rate is 0.4, the price is 800, and the price of the load rate is still 800 thereafter, until the load rate reaches 0.5, in other words, the correspondence between the load rate and the premium is the correspondence between the load interval and the specific price.

[0058] Further, step S2003 is performed to determine an initial price combination based on the unit operation capacity curve and the reference price information.

[0059] Specifically, on the basis of the above reference price information, the correspondence between the price and the load rate can be further refined, for example, in the load rate interval [0.4, 0.5), the corresponding price is 800, further, two subintervals [0.4, 0.45) and [0.45, 0.5) are divided, in the subinterval [0.4, 0.45), the price is 800, and in the subinterval [0.4, 0.45), the price is increased to 820 yuan.

[0060] It should be noted that the above subinterval division is only an example, and alternatively, the number of load rate subintervals can be divided according to the load rate, for example, in the part with low load rate, such as the load rate interval [0.4, 0.5), the number of segmentation blocks can be relatively small, such as the aforementioned two subintervals, while in the part with high load rate, such as the load rate interval [0.8, 0.9), more precise segmentation is required, such as dividing into ten load rate subintervals [0.80, 0.81), [0.81, 0.82), … [0.89, 0.9).

[0061] It can be understood that the product of the segmented load rate and the rated capacity (maximum power) of the unit is the power of the increase based on the previous price segment. Each point of power added corresponds to a new price increase. For example, the price of the subinterval [0.45, 0.5) is 820, and the previous price stage is 0.4, which corresponds to a price of 800, the product of the load rate value in the subinterval [0.45, 0.5) and the rated capacity corresponding to the unit is the power of the increase in the power corresponding to the load rate 0.4, and the price 820 corresponding to the subinterval [0.45, 0.5) is the increase in the price of the price 800 corresponding to the previous price stage load rate 0.4.

[0062] It should be pointed out that during the actual operation of the unit, it is not always operated at full load, that is, the actual operating load rate is not always 1. When the load rate is not 1, the unit's generated power may reach the planned power (the unit capacity determined by the aforementioned steps S2001 to S2003). In this case, there is no need to consider the prices of all situations from the lowest load rate to the load rate of 1. Instead, it is only necessary to consider the lowest load rate to the highest load rate that can meet the planned power. Specifically, the highest load rate that can meet the planned power can be determined by the following mathematical model.

[0063] Operating capacity ≤∑ i＝0 Load rate 最小出力点 *Rated capacity of the unit + load rate i *Rated capacity of the unit

[0064] Among them, the operating capacity in the above formula is the unit capacity determined by the aforementioned steps S2001 to S2003. When the right value in the above formula is basically equal to the left value, the load rate corresponding to the largest i at this time is determined as the maximum load rate, and the load rates from the minimum load rate to the maximum load rate and their respective corresponding prices are used as the initial electricity price combination in the above step S202.

[0065] It is worth noting that the above initial electricity price combination is the core confidential data of each company and cannot be directly shared. Instead, it needs to be encrypted and the encrypted data (first ciphertext) is sent to the server for electricity price modeling. Specifically, the encryption process includes the following steps S5001 to S5004.

[0066] S5001, performing data conversion on the initial electricity price combination according to a first truth table to obtain a first ciphertext;

[0067] S5002, encrypt the first truth table according to the first secret key to obtain a third ciphertext;

[0068] S5003, performing a scramble operation on the third ciphertext to obtain a fourth ciphertext;

[0069] S5004, sending the first ciphertext and the fourth ciphertext to the server;

[0070] The above-mentioned first truth table is used to perform data conversion on the above-mentioned initial electricity price combination to encrypt the initial electricity price combination. At the same time, the first truth table is encrypted according to the first secret key and the order is scrambled, and the encrypted initial electricity price combination and the first truth table are sent to the server.

[0071] It should be noted that the server side has the same first secret key, and the server side will decrypt the above first truth table according to the first secret key, and parse the initial electricity price combination according to the decrypted first truth table. It is worth noting that the initial electricity price combination parsed by the server side is only data for calculation, and the server side itself cannot parse the specific meaning of the data, therefore, the privacy of each company is protected, so that each company can also perform electricity price estimation modeling through the server without disclosing the core data of the company.

[0072] Further, the server calculates the corresponding clearing price according to the initial electricity price combination after obtaining the initial electricity price combination. Specifically, the server adds up the unit output (planned output power) according to the initial electricity price combination sent by all companies after receiving the first ciphertext. Specifically, for the server, there is a total power generation demand of the power generation company. In order to minimize the power generation cost, the server will find the lowest price of the power generation load rate.

[0073] For example, assuming that the initial electricity price combination of company A is 0.4 load rate corresponding to 700 yuan, 0.5 load rate corresponding to 800, and the initial electricity price combination of company B is 0.3 load rate corresponding to 650 yuan, 0.4 load rate corresponding to 750 yuan, the server will preferentially find the lowest price of company B 0.3 load rate. Assuming that the power generation power of the unit of company B is 3 gigawatts at 0.3 load rate.

[0074] It can be understood that the load rate and the output power satisfy the following mathematical model:

[0075] Unit output = load rate * unit rated capacity

[0076] The above unit output is the actual output power of the unit at the corresponding load rate, and the above unit rated capacity is the maximum output power of the unit under full load.

[0077] Further assuming that the total output power of each power generation company is 5 gigawatts at this time, the second lowest price of company A 0.4 load rate is found at this time, and the power generation power of the unit of company A is 3 gigawatts at 0.4 load rate. At this time, the power generation of the two companies A and B can meet the demand, and the power generation power of each unit is determined as the bid output at this time, and the output of each power generation company is determined through the above-mentioned manner for 96 15-minute power generation demands in a day. The corresponding bid output curve of each company can be obtained, and the clearing price corresponding to each bid output in the bid output curve can also be obtained according to different bid processing, and then each bid output and the corresponding clearing price are taken as the first electricity price combination.

[0078] It should be noted that the total power generation demand of the power generation company is calculated by the following formula:

[0079] Bidding space = direct load - total power generation of local power plants - load on interconnection lines - wind and solar output

[0080] -Nuclear power output-Captive unit output-Pumped storage output

[0081] The above direct regulation meets the total power generation demand, and the total thermal power generation demand of each power generation company is the total power generation demand minus the power generation power of each local power plant, external power, wind power generation power, standby unit power (self-provided unit output) and hydropower generation power.

[0082] Then, when the final winning output curves and corresponding clearing prices of each unit are obtained, the obtained winning output curves and corresponding clearing prices of each unit are returned to the corresponding power generation company as the second ciphertext according to the sending ID of the first ciphertext.

[0083] It should be noted that the encryption process of the above-mentioned second ciphertext can refer to the aforementioned steps S5001 to S5004. The server will encrypt the winning output curve and the corresponding clearing price based on the second truth table. In addition, the second truth table is encrypted and scrambled using the second secret key. The above-mentioned second secret key and the above-mentioned first key can be the same key or different keys, and there is no specific limitation here. Finally, the encrypted second truth table and the encrypted winning output curve and the corresponding clearing price are sent back to each power generation company respectively.

[0084] Then execute the above step S204, obtain the second ciphertext returned by the server, and parse the second ciphertext to obtain the first electricity price combination information, wherein the server is used to receive the first ciphertexts sent by multiple nodes respectively, and perform modeling based on multiple first ciphertexts. The second ciphertext is determined by the server according to the modeling results. The first electricity price combination includes the first output curve and the first electricity prices corresponding to the multiple power generation amounts in the first output curve.

[0085] Specifically, the parsing process may include the following steps S5005 to S5006.

[0086] S5005: Obtain the second ciphertext and the fifth ciphertext returned by the server, wherein the fifth ciphertext is the encrypted second truth table, and the second truth table is used to generate the second ciphertext;

[0087] S5006, decrypt the fifth ciphertext using the second secret key to obtain a second truth table;

[0088] S5007: Parse the second ciphertext according to the second truth table to obtain a first electricity price combination.

[0089] It should be noted that the power generation company itself holds the second secret key. The power generation company parses the second truth table according to the second secret key it holds, and parses out the first electricity price combination through the parsed second truth table.

[0090] Further, execute the above step S206, input the first electricity price combination and the reference information set into the target self-circulating network to obtain target electricity price information, wherein the target electricity price information includes at least one second power generation and a clearing electricity price probability distribution respectively matching the at least one second power generation.

[0091] Specifically, the above step S206 includes the following steps S206-1 to S206-4:

[0092] S206-1, obtaining a reference information set, wherein the reference information set includes at least one of the following: reference historical electricity price data, historical disclosure data, weather information, holiday information, coal price information, historical unit operating capacity information at the current node, and supply and demand description information, wherein the supply and demand description information is used to indicate the expected tightness of the power supply;

[0093] The simulated winning output curve and simulated clearing electricity price obtained in step S204 above, as well as real historical electricity price data, historical disclosure data (load, etc.), weather information, holiday information, coal price information, and the company's historical unit operating capacity data are used as a reference information set. In addition, the reference information set also includes supply and demand description information to reflect the expected degree of power supply and demand tension. The formula is as follows:

[0094]

[0095] S206-2, determining a target feature vector sequence according to the first electricity price combination and the reference information set;

[0096] Specifically, the above reference set is cleaned and processed so that all data are presented as a characteristic sequence of 96 time points per day.

[0097] S206-3: Inputting the target feature vector sequence into the target self-circulating network, wherein the target self-circulating network includes a plurality of feature processing modules connected in sequence, wherein a current feature processing module among the plurality of feature processing modules is configured to receive a historical output result of a previous feature processing module as a first input, and receive a current feature vector in the target feature vector sequence that matches the current time node as a second input, and the current feature processing module is further configured to perform likelihood processing on a reference result determined based on the first input and the second input to obtain a current output result;

[0098] The above target self-loop network is as follows Figure 7 As shown, Figure 7 In, z i，t-2and x i，t-1 is the input of the first feature processing module in the target self-loop network, and the output h i，t-1 is obtained through the first feature processing module i，t-1 , the result z i，t-1 of the entire model output at the last moment (the real result at the last moment), and the independent variable x i，t at the next moment are taken as the input of the next feature processing module.

[0099] The hidden layer update process of the above target autoregressive network in the training stage is as follows:

[0100]

[0101] The above refers to a given parameter of the LSTM cell state used in the autoregressive recurrent network. H represents an autoregressive recurrent network, and h is the output result of the autoregressive recurrent network hidden layer.

[0102] It is worth noting that in the training process, since all data is known, z i，t-1 uses the real result at the last moment, while in the prediction part after the model training, z i，t-1 needs to be taken from the result at the last moment.

[0103] As shown in Figure 7 , a likelihood function is added to the autoregressive recurrent network to make the output prediction value a probability distribution, which can generally be selected as a Gaussian distribution (suitable for real continuous distribution data) or a negative binomial distribution (suitable for integer data), without specific limitation here.

[0104] It should be noted that the above is the neural network model in the training process. Here, the likelihood function is taken as a Gaussian distribution for illustration, and it can be understood that the electricity price prediction result is a price, which is a continuously changing number, and is approximated to have 3 digits after the decimal point. The Gaussian distribution formula is as follows:

[0105]

[0106] Where μ is the mathematical expectation, σ 2 is the variance, w is the weight obtained in the likelihood function training, and b is the error term.

[0107] It should be noted that since there is no negative variance, in order to ensure that the variance is positive, a softplus activation function is added, which can be understood as a smoothed version of the relu function. The softplus activation function makes the variance change continuous and non-negative.

[0108] Further, the final DeepAR overall formula is as follows:

[0109]

[0110] wherein, is a probability product form of the joint probability distribution, is to express the autoregressive result in the form of a likelihood function.

[0111] is a likelihood function. In addition, As described in the foregoing, this will not be repeated here, and only H is emphasized again as an autoregressive recurrent network.

[0112] The target autoregressive recurrent network can be trained by a plurality of historical day reference information sets on the autoregressive recurrent network shown in Figure 7 The trained network is used as the target autoregressive network described above, and the daily updated reference information set is input into the trained target autoregressive network, and then step S206-4 is performed to determine the target electricity price information according to the output result of the last feature processing module in the plurality of feature processing modules. To obtain the probability distribution of the optimized clearing electricity price.

[0113] After obtaining the target electricity price information, step S208 is performed to determine the target electricity price combination according to the target electricity price information, wherein the target electricity price combination includes a plurality of target power generation amounts and a plurality of target electricity prices matched with the plurality of target power generation amounts. Specifically, step S208 can include the following steps S208-1 to S208-3.

[0114] S208-1, obtaining a current power generation amount from at least one second power generation amount, and a current electricity price probability distribution matched with the current power generation amount;

[0115] S208-2, determining a reference electricity price matched with the current power generation amount according to the distribution mean of the current electricity price probability distribution;

[0116] S208-3, adjusting the reference electricity price according to the distribution characteristics of the current electricity price probability distribution to obtain a target electricity price in the target electricity price combination.

[0117] It can be understood that the step S208-1 described above is to obtain the target electricity price information through the step S206 described above, and then the step S208-2 is performed to determine the reference electricity price matched with the current power generation amount according to the distribution mean of the current electricity price probability distribution. Specifically, the mean of the predicted probability distribution is calculated first, and the result is determined as the clearing electricity price predicted by each node.

[0118] After that, the step S208-3 is performed, the reference electricity price is adjusted according to the distribution characteristics of the current electricity price probability distribution, and one target electricity price in the target electricity price combination is obtained. Specifically, when the prediction certainty is relatively low, a conservative bidding strategy is adopted, and the prices lower than the predicted value in the bidding table are set more; when the certainty is relatively high, an aggressive strategy is adopted, and the output points close to the predicted value in the bidding are increased.

[0119] It should be noted that the above certainty can be determined by statistical characteristics or by shape. Further, for the way of determining the certainty by statistical characteristics, the standard deviation and the 95% confidence interval of the prediction distribution need to be calculated. When the standard deviation is large and the confidence interval is wide, it is considered that the certainty is low.

[0120] Alternatively, for the way of determining the certainty by shape, the kurtosis and skewness of the prediction distribution need to be checked. Lower kurtosis indicates that the prediction value distribution is wider. The skewness of the distribution may affect risk assessment. The longer the right tail of the distribution means that the possibility of higher extreme value increases, and higher upward risk may be faced. The longer the left tail is the higher downward risk, that is, the lower certainty.

[0121] Through the above embodiments of the present application, the first ciphertext obtained by encrypting the initial electricity price combination can be sent to the server, wherein the initial electricity price combination includes a plurality of initial power generation amounts and initial electricity prices matched with the plurality of initial power generation amounts respectively; then the second ciphertext returned by the server is obtained, and the first electricity price combination information is obtained according to the second ciphertext, wherein the server is used to receive the first ciphertext sent by the plurality of nodes respectively, and modeling is performed based on the plurality of first ciphertexts. The second ciphertext is determined by the server according to the modeling result. The first electricity price combination includes a first output curve and first electricity prices corresponding to a plurality of power generations in the first output curve. Further, the first electricity price combination and the reference information set are input into the target self-loop network to obtain target electricity price information, wherein the target electricity price information includes at least one second power generation and a clearing electricity price probability distribution matched with the at least one second power generation respectively. Then, the target electricity price combination can be determined according to the target electricity price information, wherein the target electricity price combination includes a plurality of target power generations and target electricity prices matched with the plurality of target power generations respectively, thereby solving the problem of inaccurate electricity price prediction in the prior art.

[0122] The following will be described in combination with Figure 3 A complete electricity price combination determination method is described.

[0123] S302, align the data used for simulation by each company.

[0124] It should be noted that, as before, the power production and sales data are the core private data of each company and cannot be disclosed externally, and in theory only the power grid company can obtain all the information. In order to improve the prediction accuracy of each power generation company, it is necessary to use the private data of other companies under the condition of ensuring data security (under the condition that each company has reached an agreement) through federated learning, to solve the problem of data isolation while ensuring that the data will not be leaked.

[0125] In the above step S302, N data owners {F1,...,F N} and the respective data to be combined {D1,...,D N} are defined, and F i will not expose data D i to others during model training.

[0126] It is worth noting that the feature space of the shared data for each company's own offer, unit operation mode and the like is the same, the business is similar, but the users of each company are the power generation and power selling institutions under their own flag, and the user intersection is small, so horizontal federated learning can be selected.

[0127] Specifically, the feature space is represented as X, the label space is represented as Y, and I is used to represent the sample ID space, and the relationship of the data set features is as follows:

[0128]

[0129] As shown in Figure 4 , Figure 4 , the power offer and unit operation mode are features, the target price is a label, and the power selling company of each company itself is a sample ID.

[0130] S304, the winning price pair is calculated by the data of the company itself (the offer table is obtained by similar days).

[0131] Specifically, the clearing algorithm is simulated based on the historical similar day information, and the winning power curve of each unit and the clearing price based on the simulation are obtained, as shown in Figure 5 , and the simulation process details of the local part are referred to Figure 6 .

[0132] First, according to the company's unit start-stop machine plan, the 96-point (96 15-minute points in a day) unit operation capacity is updated, wherein the real-time power generation output of the running and standby unit is set to increase the total adjustable capacity by 1 / 8 every 15 minutes from the start time, until the bidding output completely reaches the adjustable output; the shutdown is to reduce the current running capacity by 1 / 8 every 15 minutes, and the output is reduced to 0 in one hour; the unit without start-stop is kept at the current running capacity. The 96-point unit operation capacity obtained is the maximum output of the unit at each time.

[0133] Then, the weighted average historical quotation obtained by the similar day method is used as the preliminary quotation scheme. The quotation is divided into several quotation segments according to the unit load rate, and the same price is quoted within a load rate range, as shown in Table 1,

[0134] Load rate 0.4 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 1.0 Quote 300 400 500 600 700 800 900 1000 1100 1300

[0135] Table 1

[0136] When using the quotation table shown in Table 1, when the called unit output needs to be increased to the next stage, the higher quotation of the next stage is also used.

[0137] Further, in combination with the quotation table and the unit rated capacity, the unit can be divided into multiple segments of bidding output and quotation according to different load rates. On the basis of each quotation segment of the quotation table, the quotation and the load rate are further divided according to the same proportion to refine different output conditions. In the part with low load rate, the number of division blocks can be relatively small, and in the part with high load rate, more refined division is needed. The product of the divided load rate and the unit rated capacity is the output increased on the basis of the previous quotation segment. Each point of increased output corresponds to a new price increase. When the cumulative capacity is substantially equal to the 96-point unit operation capacity, the matching based on the preliminary quantity-price pair is completed, Figure 6 The preliminary quantity-price pair determination formula in the above embodiment is shown.

[0138] Running capacity ≤∑ i＝0 Load rate 最小出力点 Unit rated capacity + load rate i Unit rated capacity

[0139] The above formula has been described in the foregoing embodiment, and will not be repeated here. Further, the step S503 in the above embodiment is executed, Figure 5 The preliminary quantity-price pair is masked by SMC, and the masking result is sent to the server.

[0140] S306, the company data is summarized by federated learning, and the unit bidding curve and the clearing price are simulated.

[0141] Specifically, after receiving the data sent by each company, the server executes Figure 5In step S505, the quantity-price pair of all participating companies is compared with the total thermal power bidding space to simulate the winning power curve of each unit and the clearing price. Specifically, the server sets the clearing price to increase from low to high, and accumulates the power output of each unit according to the initial quantity-price pair of all units. When the quantity-price pair is stacked until the planned cumulative power output exceeds the thermal power bidding space, the final winning power curve of each unit and the corresponding clearing price are obtained.

[0142] Finally, step S507 in the method is executed, and the server feeds back the results to each company according to the corresponding sample ID. Figure 5

[0143] S308, based on the simulation results and other characteristics, the DeepAR optimizes the price prediction results.

[0144] The specific process is described in the foregoing steps S206-1 to S206-4, which will not be repeated here.

[0145] S310, adjust the bid according to the prediction result.

[0146] First, the uncertainty of the prediction result needs to be evaluated, which can be evaluated by statistical characteristics or by shape. Specifically, for the case of evaluation by statistical characteristics, the standard deviation and 95% confidence interval of the prediction distribution are calculated. When the standard deviation is large and the confidence interval is wide, it is considered to have high uncertainty. For the evaluation method by shape, the kurtosis and skewness of the prediction distribution are checked. Lower kurtosis indicates a wider distribution of predicted values. The skewness of the distribution may affect risk assessment. A longer right tail of the distribution means a higher likelihood of extreme values, which may face higher upward risk. A longer left tail means higher downward risk.

[0147] Further, the mean of the prediction probability distribution is calculated, which can be considered as the clearing price of each node prediction. When the prediction certainty is relatively low, a conservative bidding strategy is adopted, and the prices lower than the predicted value in the bid table are set. When the certainty is high, an aggressive strategy is adopted, and the power points close to the predicted value in the bid are increased.

[0148] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be realized by software and the necessary general hardware platform, of course, it can also be realized by hardware, but in many cases the former is a better implementation. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes a plurality of instructions for making a terminal device (which can be a mobile phone, computer, server, or network device, etc.) execute the method of each embodiment of the present application. ​

[0149] A device for determining an electricity price portfolio is also provided in the embodiments, which is used to implement the above embodiments and preferred embodiments, and will not be described again. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, hardware, or a combination of software and hardware implementations are also possible and contemplated.

[0150] Figure 8 A structural block diagram of a device for determining an electricity price portfolio according to an embodiment of the present application is shown in FIG. 1, which includes:

[0151] The sending unit 92 is configured to send the first ciphertext obtained by encrypting the initial electricity price portfolio to the server, wherein the initial electricity price portfolio includes a plurality of initial generation amounts and initial electricity prices respectively matched with the plurality of initial generation amounts.

[0152] The parsing unit 94 is configured to obtain the second ciphertext returned by the server, and parse the first electricity price portfolio information according to the second ciphertext, wherein the server is configured to receive the first ciphertext sent by each of the plurality of nodes, and model based on the plurality of first ciphertexts, the second ciphertext is determined by the server according to the modeling result, and the first electricity price portfolio includes a first output curve and first electricity prices respectively corresponding to a plurality of generation amounts in the first output curve.

[0153] The input unit 96 is configured to input the first electricity price portfolio and the reference information set into the target self-loop network to obtain target electricity price information, wherein the target electricity price information includes at least one second generation amount and a clearing electricity price probability distribution respectively matched with the at least one second generation amount.

[0154] The determining unit 98 is configured to determine a target electricity price portfolio according to the target electricity price information, wherein the target electricity price portfolio includes a plurality of target generation amounts and target electricity prices respectively matched with the plurality of target generation amounts.

[0155] The first ciphertext obtained by encrypting the initial electricity price combination can be sent to a server, wherein the initial electricity price combination includes a plurality of initial electricity generation amounts and initial electricity prices matched with the plurality of initial electricity generation amounts respectively; then, the second ciphertext returned by the server is obtained, and the first electricity price combination information is obtained according to the second ciphertext, wherein the server is used for receiving the first ciphertext sent by the plurality of nodes respectively, and modeling based on the plurality of first ciphertexts, the second ciphertext is determined by the server according to the modeling result, the first electricity price combination includes a first output curve and first electricity prices corresponding to a plurality of electricity generation amounts in the first output curve respectively; further, the first electricity price combination and the reference information set are input into the target self-loop network to obtain target electricity price information, wherein the target electricity price information includes at least one second electricity generation amount and a clearing electricity price probability distribution matched with the at least one second electricity generation amount respectively; then, the target electricity price combination can be determined according to the target electricity price information, wherein the target electricity price combination includes a plurality of target electricity generation amounts and target electricity prices matched with the plurality of target electricity generation amounts respectively, thereby solving the problem of inaccurate electricity price prediction in the prior art.

[0156] In an exemplary embodiment, the electricity price combination determination device further includes an electricity price combination determination unit configured to determine, according to the at least one generator set and the start-stop plan of each of the at least one generator set, a unit operation capacity curve of each of the at least one generator set matched with the target period; determine, according to the at least one generator set and the historical electricity price combination corresponding to each of the at least one historical period, reference electricity price information corresponding to the target period, wherein the reference electricity price information includes reference electricity prices corresponding to a plurality of unit load rates respectively; and determine the initial electricity price combination based on the unit operation capacity curve and the reference electricity price information.

[0157] In an exemplary embodiment, the electricity price combination determination device further includes a historical period determination unit configured to obtain a period description feature of the target period, wherein the period description feature is used to indicate period description information of the target period, and the period description information includes at least one of the following: week type information, geographical environment information, weather information, lunar solar term information, and holiday information; and in a case where a feature similarity between the period description feature and a reference period description feature of each of the at least one reference historical period satisfies a similarity condition, the at least one reference historical period is determined as the at least one historical period.

[0158] In an exemplary embodiment, the electricity price combination determination device further includes an electricity price information determination unit configured to obtain, according to the at least one generator set, historical electricity price combinations corresponding to each of the at least one historical period; determine a weighted summation result between the at least one historical electricity price combination according to reference weights matched with each of the at least one historical period; and determine the weighted summation result as the reference electricity price information corresponding to the target period.

[0159] In an example embodiment, the input unit 96 is further configured to obtain a reference information set, wherein the reference information set comprises at least one of the following: reference historical electricity price data, historical disclosure data, weather information, holiday information, coal price information, current node historical unit operation capacity information, and supply-demand description information, wherein the supply-demand description information is used to indicate the predicted tension degree of power supply; determine a target feature vector sequence according to the first electricity price combination and the reference information set; input the target feature vector sequence into a target self-loop network, wherein the target self-loop network comprises a plurality of sequentially connected feature processing modules, a current feature processing module in the plurality of feature processing modules is configured to receive a historical output result of a previous feature processing module as a first input, and receive a current feature vector in the target feature vector sequence that matches a current time node as a second input, and the current feature processing module is further configured to perform likelihood processing on a reference result determined based on the first input and the second input to obtain a current output result; and determine the target electricity price information according to an output result of a last feature processing module in the plurality of feature processing modules.

[0160] In an example embodiment, the determination unit 98 is further configured to obtain a current power generation capacity from the at least one second power generation capacity, and a current power generation capacity matching current electricity price probability distribution; determine a reference electricity price matching the current power generation capacity according to a distribution mean of the current electricity price probability distribution; and adjust the reference electricity price according to a distribution feature of the current electricity price probability distribution to obtain one target electricity price in the target electricity price combination.

[0161] Embodiments of the present application also provide a computer readable storage medium, which stores a computer program, wherein the computer program is configured to execute the steps in any of the above method embodiments when running.

[0162] Optionally, in the present embodiment, the storage medium can be configured to store a computer program for executing the following steps:

[0163] S1, sending a first ciphertext obtained by encrypting an initial electricity price combination to a server, wherein the initial electricity price combination comprises a plurality of initial power generation capacities and initial electricity prices respectively matching the plurality of initial power generation capacities;

[0164] S2, obtaining a second ciphertext returned by the server, and analyzing the first electricity price combination information according to the second ciphertext, wherein the server is configured to receive a plurality of first ciphertexts respectively sent by a plurality of nodes, and model based on the plurality of first ciphertexts, the second ciphertext is determined by the server according to a modeling result, and the first electricity price combination comprises a first output curve and first electricity prices respectively corresponding to a plurality of power generation capacities in the first output curve;

[0165] S3, inputting the first electricity price combination and the reference information set into the target self-circulating network to obtain target electricity price information, wherein the target electricity price information includes at least one second power generation amount and a clearing electricity price probability distribution respectively matching the at least one second power generation amount;

[0166] S4, determining a target electricity price combination according to the target electricity price information, wherein the target electricity price combination includes multiple target power generation amounts and target electricity prices that match the multiple target power generation amounts.

[0167] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0168] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail here.

[0169] An embodiment of the present application further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0170] Optionally, in this embodiment, the processor may be configured to execute the following steps through a computer program:

[0171] S1, sending a first ciphertext obtained by encrypting an initial electricity price combination to a server, wherein the initial electricity price combination includes multiple initial power generation amounts and initial electricity prices matching each of the multiple initial power generation amounts;

[0172] S2, obtaining a second ciphertext returned by the server, and parsing the second ciphertext to obtain first electricity price combination information, wherein the server is configured to receive first ciphertexts respectively sent by multiple nodes and perform modeling based on the multiple first ciphertexts, the second ciphertext being determined by the server based on the modeling results, and the first electricity price combination including the first output curve and first electricity prices corresponding to multiple power generation amounts in the first output curve;

[0173] S3, inputting the first electricity price combination and the reference information set into the target self-circulating network to obtain target electricity price information, wherein the target electricity price information includes at least one second power generation amount and a clearing electricity price probability distribution respectively matching the at least one second power generation amount;

[0174] S4, determine a target electricity price combination according to the target electricity price information, wherein the target electricity price combination comprises a plurality of target power generation amounts and target electricity prices respectively matched with the plurality of target power generation amounts.

[0175] In one example embodiment, the electronic device described above can further comprise a transmission device connected to the processor and an input / output device connected to the processor.

[0176] Embodiments of the present application also provide a computer program product comprising a non-volatile computer readable storage medium storing a computer program product, the computer program being executed by a processor to implement the steps of the method in various embodiments of the present application.

[0177] Optionally, in the present embodiment, the computer program described above can be configured to be executed by the processor to implement the following steps:

[0178] S1, send a first ciphertext obtained by encrypting an initial electricity price combination to a server, wherein the initial electricity price combination comprises a plurality of initial power generation amounts and initial electricity prices respectively matched with the plurality of initial power generation amounts;

[0179] S2, obtain a second ciphertext returned by the server, and obtain first electricity price combination information according to the second ciphertext, wherein the server is configured to receive a plurality of first ciphertexts respectively sent by a plurality of nodes, and model based on the plurality of first ciphertexts, the second ciphertext is determined by the server according to a modeling result, and the first electricity price combination comprises a first output curve and first electricity prices respectively corresponding to a plurality of power generation amounts in the first output curve;

[0180] S3, input the first electricity price combination and a reference information set into a target self-loop network to obtain target electricity price information, wherein the target electricity price information comprises at least one second power generation amount and a clearing electricity price probability distribution respectively matched with the at least one second power generation amount;

[0181] S4, determine a target electricity price combination according to the target electricity price information, wherein the target electricity price combination comprises a plurality of target power generation amounts and target electricity prices respectively matched with the plurality of target power generation amounts.

[0182] The specific examples in the present embodiment can refer to the examples described in the above embodiments and example embodiments, which will not be described here again.

[0183] Obviously, those skilled in the art should understand that the modules or steps of the present application described above can be implemented using a general-purpose computing device, they can be concentrated on a single computing device, or distributed across a network composed of multiple computing devices, they can be implemented using program code executable by the computing device, and thus, they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be performed in a different order than herein, or they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0184] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for determining an electricity price portfolio, characterized in that The method comprises the steps of: sending a first ciphertext obtained by encrypting an initial electricity price combination to a server, wherein the initial electricity price combination comprises a plurality of initial power generation amounts and initial electricity prices respectively matched with the initial power generation amounts; obtaining second ciphertext returned by the server and first electricity price combination information parsed according to the second ciphertext, wherein the server is configured to receive the first ciphertext respectively sent by a plurality of nodes and model based on the first ciphertext, the second ciphertext is determined by the server according to a modeling result, the first electricity price combination comprises a first output curve and first electricity prices respectively corresponding to a plurality of power generation amounts in the first output curve; inputting the first electricity price combination and a reference information set into a target self-loop network to obtain target electricity price information, wherein the target electricity price information comprises at least one second power generation amount and a clearing electricity price probability distribution respectively matched with the at least one second power generation amount; determining a target electricity price combination according to the target electricity price information, wherein the target electricity price combination comprises a plurality of target power generation amounts and target electricity prices respectively matched with the target power generation amounts; and the step of sending the first ciphertext obtained by encrypting the initial electricity price combination to the server comprises the steps of: performing data conversion on the initial electricity price combination according to a first truth table to obtain the first ciphertext; encrypting the first truth table according to a first secret key to obtain third ciphertext; performing a disorder operation on the third ciphertext to obtain fourth ciphertext; and sending the first ciphertext and the fourth ciphertext to the server. The step of obtaining the second ciphertext returned by the server and the first electricity price combination information parsed according to the second ciphertext comprises the steps of: obtaining second ciphertext and fifth ciphertext returned by the server, wherein the fifth ciphertext is a second truth table after encryption, and the second truth table is used to generate the second ciphertext; decrypting the fifth ciphertext according to a second secret key to obtain a second truth table; and parsing the second ciphertext according to the second truth table to obtain the first electricity price combination. The first electricity price combination and the reference information set are input into a target self-loop network to obtain target electricity price information, including: obtaining the reference information set, wherein the reference information set includes: reference historical electricity price data, historical disclosure data, weather information, holiday information, coal price information, current node historical unit operation capacity information, supply and demand description information, and the supply and demand description information is used to indicate the predicted tension degree of power supply; determining a target feature vector sequence according to the first electricity price combination and the reference information set; inputting the target feature vector sequence into the target self-loop network, wherein the target self-loop network includes a plurality of sequentially connected feature processing modules, a current feature processing module in a plurality of the feature processing modules is used to receive a historical output result of a previous feature processing module as a first input, and receive a current feature vector matched with a current time node in the target feature vector sequence as a second input, and the current feature processing module is further used to perform likelihood processing on a reference result determined based on the first input and the second input to obtain a current output result; and determining the target electricity price information according to the output result of a last feature processing module in a plurality of the feature processing modules.

2. The method of claim 1, wherein, Before the first ciphertext obtained by encrypting the initial electricity price combination is sent to the server, the method further includes: determining, according to at least one generator set and a start-stop plan of each of the at least one generator set, a unit operation capacity curve of each of the at least one generator set matched with a target period; determining, according to at least one historical electricity price combination corresponding to each of at least one historical period and each of the at least one generator set, reference electricity price information corresponding to the target period, wherein the reference electricity price information includes reference electricity prices corresponding to a plurality of unit load rates, respectively; determining the initial electricity price combination based on the unit operation capacity curve and the reference electricity price information.

3. The method of claim 2, wherein, Before the determining, according to at least one historical electricity price combination corresponding to each of at least one historical period and each of the at least one generator set, reference electricity price information corresponding to the target period, the method further includes: obtaining period description features of the target period, wherein the period description features are used to indicate period description information of the target period, and the period description information includes at least one of the following: week type information, geographical environment information, weather information, lunar solar term information, and holiday information; in a case where a feature similarity between the period description features and reference period description features of each of at least one reference historical period satisfies a similarity condition, determining the at least one reference historical period as the at least one historical period.

4. The method of claim 2, wherein, The determining, according to at least one historical electricity price combination corresponding to each of at least one historical period and each of the at least one generator set, reference electricity price information corresponding to the target period, includes: obtaining at least one historical electricity price combination corresponding to each of the at least one generator set and each of the at least one historical period; determining a weighted summation result between the at least one historical electricity price combination according to a reference weight matched with each of the at least one historical period; and The weighted sum result is determined as the reference electricity price information corresponding to the target period.

5. The method of claim 1, wherein, The target electricity price combination is determined according to the target electricity price information, including: Obtaining a current power generation from at least one of the second power generations, and a current electricity price probability distribution matched with the current power generation; Determining a reference electricity price matched with the current power generation according to a distribution mean of the current electricity price probability distribution; Adjusting the reference electricity price according to a distribution feature of the current electricity price probability distribution to obtain one of the target electricity prices in the target electricity price combination.

6. An apparatus for determining a combination of electricity prices, characterized in that Including: The sending unit is configured to send a first ciphertext obtained by encrypting an initial electricity price combination to a server, wherein the initial electricity price combination includes a plurality of initial power generations and initial electricity prices respectively matched with the initial power generations; The parsing unit is configured to obtain a second ciphertext returned by the server and parse the first electricity price combination information according to the second ciphertext, wherein the server is configured to receive the first ciphertext sent by a plurality of nodes respectively, and model based on the first ciphertext, the second ciphertext is determined by the server according to a modeling result, the first electricity price combination includes a first output curve and a first electricity price respectively corresponding to a plurality of power generations in the first output curve; The input unit is configured to input the first electricity price combination and a reference information set into a target self-loop network to obtain target electricity price information, wherein the target electricity price information includes at least one second power generation and a clearing electricity price probability distribution respectively matched with the at least one second power generation; The determining unit is configured to determine a target electricity price combination according to the target electricity price information, wherein the target electricity price combination includes a plurality of target power generations and target electricity prices respectively matched with the target power generations; The electricity price combination determination apparatus is further configured to: perform data conversion on the initial electricity price combination according to a first truth table to obtain the first ciphertext; encrypt the first truth table according to a first secret key to obtain a third ciphertext; perform a disordering operation on the third ciphertext to obtain a fourth ciphertext; and send the first ciphertext and the fourth ciphertext to the server; Obtain the second ciphertext and a fifth ciphertext returned by the server, wherein the fifth ciphertext is a second truth table after encryption, the second truth table is used to generate the second ciphertext; decrypt the fifth ciphertext according to a second secret key to obtain a second truth table; parse the second ciphertext according to the second truth table to obtain the first electricity price combination; obtaining the reference information set, wherein the reference information set comprises reference historical electricity price data, historical disclosure data, weather information, holiday information, coal price information, current node historical unit operation capacity information, and supply and demand description information used to indicate the predicted tightness of power supply; determining a target feature vector sequence according to the first electricity price combination and the reference information set; inputting the target feature vector sequence into the target self-loop network, wherein the target self-loop network comprises a plurality of sequentially connected feature processing modules, a current feature processing module in the plurality of feature processing modules is configured to receive historical output results of a previous feature processing module as first input, and receive a current feature vector matched with a current time node in the target feature vector sequence as second input, and the current feature processing module is further configured to perform likelihood processing on a reference result determined based on the first input and the second input to obtain a current output result; and determining the target electricity price information according to the output result of a last feature processing module in the plurality of feature processing modules.

7. A computer readable storage medium, characterized in that, The computer-readable storage medium comprises a stored program, wherein the program, when executed, performs the method of any one of claims 1 to 5.

8. A computer program product comprising a computer program, characterized in that, The computer program, when executed by a processor, implements the steps of the method of any one of claims 1 to 5.

Images