Machine learning-based smart park device energy consumption optimization method, system and medium

Abstract

The application relates to the technical field of energy consumption optimization, in particular to a smart park equipment energy consumption optimization method and system based on machine learning and a medium, the method comprising the following steps: S1, a monitoring module takes real monitoring data of different days in history as sending data, and performs confusion processing on the sending data to obtain a plurality of confusion data blocks; S2, an optimization module recovers different partial sending data from the confusion data blocks, and performs improvement processing on a trained analysis algorithm; S3, the optimization module takes various types of possible condition data of a target day as input data of the improved analysis algorithm to obtain output data of the improved analysis algorithm, and formulates an optimization plan according to the output data. The application can improve the intelligence of the energy consumption optimization method.

Description

Technical Field

[0001] The present application relates to the field of energy consumption optimization technology, and in particular to a method, system, and medium for optimizing energy consumption of smart park equipment based on machine learning. Background Art

[0002] Equipment energy consumption optimization can effectively reduce energy consumption and operating costs, while contributing to energy conservation, emission reduction and sustainable development.

[0003] The Chinese patent application with publication number CN108509019A provides a data center energy consumption optimization method and device. The method obtains the current working status information of the execution equipment and refrigeration equipment in the data center, obtains the corresponding current PUE coefficient based on the current working status information, and uses a preset optimization strategy to adjust the current working status information to obtain adjusted working status information. The corresponding adjusted PUE coefficient is obtained based on the adjusted working status information. When it is determined that the adjusted PUE coefficient is less than the current PUE coefficient, the execution equipment and refrigeration equipment are adjusted according to the adjusted working status information. In addition, a Chinese patent application with publication number CN114169056A discloses an energy consumption optimization method based on green buildings, which includes the following steps: S1: obtaining a three-dimensional model of the building to be designed; S2: designing the relevant layout of the enclosure system, HVAC system, and lighting system in the building based on the three-dimensional model; enclosure system: designing the proportion and position distribution of transparent areas and non-transparent areas in the enclosure structure based on the three-dimensional model of the building and the climate of the area where the building is located, and then determining the thickness of each transparent area and non-transparent area; HVAC system: determining the layout and power of each HVAC equipment in the HVAC system based on the three-dimensional model of the building and the climate of the area where the building is located; S3: obtaining the energy consumption of each energy consumption system of the building, as well as the overall energy consumption of the building; comparing each energy consumption data with each data in the big data, and reducing energy consumption by rationally arranging the enclosure system, HVAC system, and lighting system in the building.

[0004] However, the energy consumption optimization methods in the above two patent applications are relatively complex and lack intelligence. Summary of the Invention

[0005] This application employs a monitoring module to obfuscate transmitted data, generating several obfuscated data blocks. An optimization module is then configured to recover different portions of transmitted data from each obfuscated data block, thereby obtaining complete transmitted data. This module then improves the trained analysis algorithm, using various possible conditional data from the target day as input to the improved analysis algorithm, and developing an optimization plan based on the output data. This application aims to enhance the intelligence of energy consumption optimization methods.

[0006] This application provides a method for optimizing energy consumption of smart park equipment based on machine learning, including the following steps:

[0007] S1. The monitoring module uses real monitoring data from different days in history as sending data, performs obfuscation processing on the sending data to obtain several obfuscated data blocks, and transmits the several obfuscated data blocks to the optimization module through the network module;

[0008] S2. The optimization module recovers different partial transmission data from each obfuscated data block, obtains transmission data based on all the partial transmission data, obtains various types of real condition data from different historical days, determines a number of observation days from the different historical days, and obtains a trained analysis algorithm and performs improvement processing on the trained analysis algorithm.

[0009] S3. The optimization module calculates various types of possible condition data for the target day based on various types of real condition data for different days in history, and the optimization module uses the various types of possible condition data for the target day as input data for the improved analysis algorithm to obtain output data for the improved analysis algorithm, and formulates an optimization plan based on the output data.

[0010] As a preferred technical solution of the present application, the optimization module performs an improvement process on the trained analysis algorithm, including the following steps:

[0011] S211. For each observation day, the optimization module calculates various types of possible condition data for the observation day based on various types of real condition data from different days in history;

[0012] S212. For each observation day, use various types of possible condition data on the observation day as input data of the trained analysis algorithm to obtain output data of the trained analysis algorithm, and calculate the absolute value of the difference between the output data and the actual monitoring data of the observation day;

[0013] S213. The optimization module calculates a first ratio of the sum of all absolute values ​​to the total number of all absolute values, and the optimization module determines whether the first ratio is less than a preset first ratio threshold. If it is less than, all steps are terminated. If it is greater than or equal to, a first improvement process is performed on the trained analysis algorithm, and the process jumps to S212.

[0014] As a preferred technical solution of the present application, the optimization module performs improvement processing on the trained analysis algorithm and further includes the following steps:

[0015] S221, the optimization module calculates a second ratio of a sum of all the real condition data to a total number of all the real condition data, and calculates a square value of a difference between each real condition data and the second ratio, and further calculates a third ratio of a sum of all the square values to a total number of all the square values, and a square root value of the third ratio;

[0016] S222, the optimization module calculates the possible condition data of each type for the target day according to the real condition data of each type for different days in history;

[0017] S223, the optimization module calculates an absolute value of a difference between the possible condition data and the corresponding second ratio for each type of possible condition data for the target day, and further calculates a fourth ratio of the absolute value to the corresponding square root value;

[0018] S224, the fourth ratio corresponding to the possible condition data of each type for the target day is used for a second improvement of the trained analysis algorithm.

[0019] As a preferred technical solution of the present application, the monitoring module performs confusion processing on the sending data, including the following steps:

[0020] S11, the monitoring module divides the sending data into a plurality of sending data blocks, and for each sending data block, generates secret data corresponding to the sending data block, and further takes the data size of the sending data block as representative data for each sending data block;

[0021] S12, for each sending data block, the monitoring module uses the secret data corresponding to the sending data block to perform secret processing on the sending data block to obtain an intermediate data block;

[0022] S13, for each intermediate data block, the monitoring module adds the secret data corresponding to the previous intermediate data block to the intermediate data block based on the representative data corresponding to the intermediate data block to obtain a confusion data block.

[0023] As a preferred technical solution of the present application, the monitoring module adds the secret data corresponding to the previous intermediate data block to the intermediate data block based on the representative data corresponding to the intermediate data block, including the following steps:

[0024] S131, the monitoring module performs a preset operation on the secret data corresponding to the previous intermediate data block to obtain target data, and performs a preset operation on the representative data corresponding to the intermediate data block to obtain process data;

[0025] S132: The monitoring module determines whether the process data is larger than the data size of the intermediate data block. If so, the monitoring module subtracts the data size of the intermediate data block from the process data and repeats this step. If less than or equal to the process data, a portion of the target data is added to the intermediate data block according to the process data.

[0026] S133, the monitoring module determines whether the complete target data has been added to the intermediate data block. If not, the process data is updated using the result data of the preset operation on the process data, and the process jumps to S132. If yes, all steps are completed.

[0027] As a preferred technical solution of the present application, the optimization module recovers part of the transmitted data from the obfuscated data block, including the following steps:

[0028] S231, the optimization module determines the data size of the obfuscated data block, subtracts the preset data from the data size as the representative data corresponding to the obfuscated data block, and performs a preset operation on the representative data to obtain process data;

[0029] S232: The optimization module determines whether the process data is greater than the representative data. If so, the representative data is subtracted from the process data, and this step is repeated. If less than or equal to the representative data, data is removed from the obfuscated data block based on the process data.

[0030] S233, the optimization module determines whether the stop condition is met. If not, the process data is updated using the result data of the preset operation on the process data, and the process jumps to S232. If the condition is met, the obfuscated data is restored to obtain part of the sent data, and all steps are completed.

[0031] This application also provides a smart park equipment energy consumption optimization system based on machine learning, including the following modules:

[0032] The monitoring module is used to use the real monitoring data of different days in history as the sending data, perform obfuscation processing on the sending data to obtain a number of obfuscated data blocks, and transmit the several obfuscated data blocks to the optimization module through the network module;

[0033] The network module is used to transmit several obfuscated data blocks generated by the monitoring module to the optimization module;

[0034] The optimization module is used to recover different partial transmission data from each obfuscated data block, obtain transmission data based on all partial transmission data, obtain various types of real condition data for different historical days, determine a number of observation days among different historical days, obtain a trained analysis algorithm, and perform improvement processing on the trained analysis algorithm; and is used to calculate various types of possible condition data for the target day based on various types of real condition data for different historical days, use the various types of possible condition data for the target day as input data of the improved analysis algorithm to obtain output data of the improved analysis algorithm, and formulate an optimization plan based on the output data.

[0035] The present application also provides a medium, wherein the medium stores program instructions, wherein when the program instructions are executed, the device where the medium is located is controlled to execute any one of the above methods.

[0036] Compared with the prior art, the present invention has at least the following advantages:

[0037] In the technical solution provided by the present application, first, the monitoring module uses the real monitoring data of different days in history as the transmission data, performs obfuscation processing on the transmission data to obtain several obfuscated data blocks, and transmits the several obfuscated data blocks to the optimization module through the network module; secondly, the optimization module recovers different partial transmission data from each obfuscated data block, obtains transmission data based on all partial transmission data, and the optimization module obtains various types of real condition data of different days in history, determines several observation days in different days in history, and the optimization module also obtains a trained analysis algorithm and performs improvement processing on the trained analysis algorithm; finally, the optimization module calculates various types of possible condition data of the target day based on various types of real condition data of different days in history, and the optimization module uses various types of possible condition data of the target day as input data of the improved analysis algorithm to obtain output data of the improved analysis algorithm, and formulates an optimization plan based on the output data. Through the present application, not only can the optimization plan be formulated simply and quickly, but also the intelligence level of the energy consumption optimization method can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0039] Figure 1 A flowchart of the method for optimizing energy consumption of smart park equipment based on machine learning in this application;

[0040] Figure 2 A flowchart for improving the processing of the trained analysis algorithm in this application;

[0041] Figure 3 A flowchart of the process of continuing to improve the trained analysis algorithm in this application;

[0042] Figure 4 Schematic diagram of the smart park equipment energy consumption optimization system based on machine learning in this application. DETAILED DESCRIPTION

[0043] The embodiments of the present application provide a method, system and medium for optimizing energy consumption of smart park equipment based on machine learning. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0044] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 The method for optimizing energy consumption of smart park equipment based on machine learning in the embodiment of the present application includes the following main steps:

[0045] S1. The monitoring module uses real monitoring data from different days in history as sending data, performs obfuscation processing on the sending data to obtain several obfuscated data blocks, and transmits the several obfuscated data blocks to the optimization module through the network module;

[0046] S2. The optimization module recovers different partial transmission data from each obfuscated data block and obtains transmission data based on all the partial transmission data. The optimization module also obtains various types of real condition data from different historical days and determines a number of observation days from different historical days. The optimization module also obtains a trained analysis algorithm and performs improvements on the trained analysis algorithm.

[0047] S3. The optimization module calculates various types of possible condition data for the target day based on various types of real condition data on different days in history, and the optimization module uses the various types of possible condition data for the target day as input data for the improved analysis algorithm to obtain output data of the improved analysis algorithm, and formulates an optimization plan based on the output data.

[0048] Specifically, in order to improve the intelligence of the energy consumption optimization method, S1 to S3 are mainly proposed. In S1, the monitoring module collects real monitoring data from different days in history and uses it as the transmission data to be sent to the optimization module. The real monitoring data refers to the monitoring data on the electricity consumption of the smart park equipment. The monitoring module performs obfuscation processing on the transmission data to obtain several obfuscated data blocks, and transmits the several obfuscated data blocks to the optimization module through the network module. This is done to ensure the security of the transmission data during transmission and prevent the transmission data from being tampered with, so as to ensure the correctness of the final optimization plan. In S2, the optimization module recovers the corresponding partial transmission data from each obfuscated data block, and then obtains the transmission data based on all the partial transmission data. The optimization module obtains various types of real condition data from different days in history, determines several observation days in different days in history, and also obtains the trained analysis algorithm and performs improvement processing on the trained analysis algorithm. It should be noted that although the analysis algorithm has been trained, if the performance of the trained analysis algorithm is not good, it needs to be further improved. For ease of understanding, the analysis algorithm can be, for example. In S3, the optimization module calculates the corresponding type of possible condition data for the target day based on the same type of real condition data on different days in history. The target day refers to a day that has not yet arrived. It can be calculated using the machine learning algorithm in the existing technology. I will not go into details about this. The optimization module uses various types of possible condition data for the target day as input data for the improved analysis algorithm. The purpose is to obtain the output data of the improved analysis algorithm. The output data is the estimated data on the electricity consumption of the smart park equipment. Then, an optimization plan is formulated based on the output data. The optimization plan can, for example, shorten the working time of the equipment.

[0049] Furthermore, the optimization module improves the trained analysis algorithm, including the following steps:

[0050] S211. For each observation day, the optimization module calculates various types of possible condition data for the observation day based on various types of real condition data from different days in history;

[0051] S212, for each observation day, the optimization module takes the various types of possible condition data of the observation day as input data of the trained analysis algorithm to obtain output data of the trained analysis algorithm, calculates absolute values of differences between the output data and the real monitoring data of the observation day;

[0052] S213, the optimization module calculates a first ratio of a sum of all absolute values to a total number of all absolute values, and judges whether the first ratio is less than a preset first ratio threshold value, in the case of less than, ends all steps, in the case of greater than or equal to, carries out first improvement processing for the trained analysis algorithm, and jumps to S212.

[0053] Specifically, please refer to Figure 2 , the process of the optimization module for improving the trained analysis algorithm is introduced, in S211, for each observation day, the optimization module uses the same type of real condition data of different days in history to calculate the possible condition data of the corresponding type of the observation day, which can be calculated using the machine learning algorithm in the prior art, which will not be described here. In fact, the optimization module uses the same type of real condition data of different days in history before the observation day. In S212, the optimization module takes the various types of possible condition data of the observation day as input data of the trained analysis algorithm to obtain output data of the trained analysis algorithm, and calculates absolute values of differences between the output data and the real monitoring data of the observation day. In S213, the optimization module calculates a first ratio of a sum of all absolute values to a total number of all absolute values, judges whether the first ratio is less than a preset first ratio threshold value, and the first ratio threshold value is set according to the actual application scenario. If it is less than, end all steps, if it is greater than or equal to, carry out first improvement processing for the trained analysis algorithm, and jump to S212 for continuous execution, and then the above examples of the analysis algorithm, the first improvement processing refers to changing at least one of the coefficients to one of the coefficients.

[0054] Further, the optimization module for improving the trained analysis algorithm further comprises the following steps:

[0055] S221, for each type of real condition data of different days in history, the optimization module calculates a second ratio of a sum of all real condition data to a total number of all real condition data, and calculates square values of differences between each real condition data and the second ratio, and also calculates a square root value of a third ratio of a sum of all square values to a total number of all square values;

[0056] S222, the optimization module calculates the various types of possible condition data of the target day according to the various types of real condition data of different days in history;

[0057] S223. For each type of possible condition data on the target day, the optimization module calculates the absolute value of the difference between the possible condition data and the corresponding second ratio, and then calculates a fourth ratio of the absolute value to the corresponding square root value.

[0058] S224 , performing a second improvement process on the trained analysis algorithm based on the fourth ratios corresponding to various types of possible conditional data of the target day.

[0059] Specifically, see Figure 3 , introduces the process of the optimization module continuing to improve the analysis algorithm for the training. In S221, for each type of real condition data on different days in history, the optimization module calculates the second ratio of the sum of all real condition data to the total number of all real condition data, calculates the square value of the difference between each real condition data and the second ratio, and also calculates the square root value of the third ratio of the sum of all square values ​​to the total number of all square values. In S222, the optimization module calculates the corresponding type of possible condition data of the target day based on the real condition data of the same type on different days in history. The calculation can be performed using the machine learning algorithm in the prior art, which will not be described in detail. In S223, for each type of possible condition data on the target day, the optimization module calculates the absolute value of the difference between the possible condition data and the corresponding second ratio. The corresponding second ratio refers to the second ratio of the corresponding type that is the same as the type of the possible condition data. The fourth ratio of the absolute value to the corresponding square root value is further calculated. The corresponding square root value also refers to the square root value of the corresponding type that is the same as the type of the possible condition data. In S224, a second improvement process is performed on the trained analysis algorithm based on the fourth ratios corresponding to the various types of possible conditional data for the target day. Continuing with the example of the analysis algorithm above, for example, if the fourth ratio corresponding to a certain type of possible conditional data is , then the second improvement process involves updating with the coefficients corresponding to that type of possible conditional data in the analysis algorithm. By continuing to perform improvement processes on the trained analysis model, the output accuracy of the trained analysis model can be improved.

[0060] Furthermore, the monitoring module performs obfuscation processing on the sent data, including the following steps:

[0061] S11. The monitoring module divides the transmitted data into a plurality of transmitted data blocks, and for each transmitted data block, generates secret data corresponding to the transmitted data block, and further, for each transmitted data block, uses the data size of the transmitted data block as representative data;

[0062] S12. For each transmitted data block, the monitoring module performs secret processing on the transmitted data block using the secret data corresponding to the transmitted data block to obtain an intermediate data block;

[0063] S13. For each intermediate data block, the monitoring module adds the secret data corresponding to the previous intermediate data block to the intermediate data block based on the representative data corresponding to the intermediate data block to obtain an obfuscated data block.

[0064] Specifically, the process of the monitoring module performing obfuscation processing on the transmitted data is introduced. In S11, the monitoring module divides the transmitted data into several transmitted data blocks. The data sizes of the several transmitted data blocks may be different. For each transmitted data block, secret data corresponding to the transmitted data block is generated. The secret data can be understood as the key used by the symmetric encryption algorithm in the prior art. For each transmitted data block, the data size of the transmitted data block is also used as representative data. The representative data is, for example, the decimal number 8, corresponding to 8 bytes. In S12, for each transmitted data block, the monitoring module uses the secret data corresponding to the transmitted data block to perform secret processing on the transmitted data block in order to obtain an intermediate data block. The secret processing can be understood as encryption processing performed using the symmetric encryption algorithm in the prior art. It should be noted that the data size of the intermediate data block is the same as the data size of the corresponding transmitted data block. In S13, for each intermediate data block, the monitoring module adds the secret data corresponding to the previous intermediate data block to the intermediate data block based on the representative data corresponding to the intermediate data block, thereby obtaining an obfuscated data block. For ease of understanding, for example, there are sending data block D1, sending data block D2, sending data block D3, secret data K1 corresponding to sending data block D1, secret data K2 corresponding to sending data block D2, secret data K3 corresponding to sending data block D3, intermediate data block T1 corresponding to sending data block D1, intermediate data block T2 corresponding to sending data block D2, and intermediate data block T3 corresponding to sending data block D3. Then, the secret data K3 corresponding to intermediate data block T3 is added to intermediate data block T1, the secret data K1 corresponding to intermediate data block T1 is added to intermediate data block T2, and the secret data K2 corresponding to intermediate data block T2 is added to intermediate data block T3.

[0065] Furthermore, the monitoring module adds the secret data corresponding to the previous intermediate data block to the intermediate data block based on the representative data corresponding to the intermediate data block, including the following steps:

[0066] S131. The monitoring module performs a preset operation on the secret data corresponding to the previous intermediate data block to obtain target data, and performs a preset operation on the representative data corresponding to the intermediate data block to obtain process data.

[0067] S132, the monitoring module judges whether the process data is greater than the data size of the intermediate data block, in the case of greater, the process data is subtracted by the data size of the intermediate data block, the step is repeated, in the case of less than or equal, a part of the target data is added to the intermediate data block according to the process data;

[0068] S133, the monitoring module judges whether the complete target data has been added to the intermediate data block, in the case of no, the process data is updated by the result data of the preset operation on the process data, and jumps to S132, in the case of yes, the whole step is ended.

[0069] Specifically, how the monitoring module adds the secret data corresponding to the last intermediate data block to the intermediate data block based on the representative data corresponding to the intermediate data block is introduced, in S131, the monitoring module performs a preset operation on the secret data corresponding to the last intermediate data block, the purpose is to obtain the target data, the preset operation can be an operation using a hash function, the meaning of the preset operation appearing multiple times in this embodiment is the same, the data size of the target data is a fixed value set in advance, for example, a decimal number 2, corresponding to 2 bytes, and the preset operation is also performed on the representative data corresponding to the intermediate data block, the purpose is to obtain the process data. In S132, the monitoring module judges whether the process data is greater than the data size of the intermediate data block, in order to facilitate understanding, for example, the process data is a binary number 10001, corresponding to a decimal number 18, the data size of the intermediate data block is a decimal number 8, corresponding to 8 bytes, then the process data 18 is greater than the data size 8 of the intermediate data block, in the case of greater, the process data is subtracted by the data size of the intermediate data block, the judgment processing is repeated until the process data is less than or equal to the data size of the intermediate data block, continue to assume that the process data is reduced to 2, in the case of less than or equal, a part of the target data is added to the intermediate data block according to the process data, for example, the first byte data of the target data from the left is added to the data after the second byte data from the left of the intermediate data block, it should be noted that when the byte data of the target data is continuously added to the intermediate data block, the original intermediate data block is still taken as a reference, instead of the intermediate data block after the addition processing. In S133, the monitoring module judges whether the complete target data has been added to the intermediate data block, that is, whether each byte data of the target data has been added to the intermediate data block, if no, the process data is updated by the result data of the preset operation on the process data, that is, the result data of the preset operation on the process data is taken as the new process data, and jumps to S132 for continuous execution, if yes, the whole step is ended.

[0070] Further, the optimization module recovers part of the sending data from the confusion data block, including the following steps:

[0071] S231, the optimization module determines the data size of the confusion data block, subtracts the result data of the preset data from the data size as the representative data corresponding to the confusion data block, and performs a preset operation on the representative data to obtain process data;

[0072] S232, the optimization module determines whether the process data is greater than the representative data, in the case of greater, subtracts the representative data from the process data, and repeats the step, in the case of less than or equal to, removes data from the confusion data block based on the process data;

[0073] S233, the optimization module determines whether the stop condition is met, in the case of not meeting, updates the process data using the result data of the preset operation on the process data, jumps to S232, in the case of meeting, restores the confusion data to obtain partial sending data, and ends all steps.

[0074] Specifically, the process of the optimization module recovering partial sending data from the confusion data block is introduced. Before the introduction, it needs to be explained that the monitoring module sends different confusion data blocks in different orders of generating different confusion data blocks, for example, the sending data is divided into several sending data blocks in the order from left to right, the order of generating different confusion data blocks is the same as this order, and the order of the optimization module receiving different confusion data blocks is the same as the order of the monitoring module sending different confusion data blocks. Therefore, after the optimization module recovers partial sending data from each confusion data block, it can obtain complete sending data according to all the partial sending data.

[0075] In S231, the optimization module determines the data size of the obfuscated data block, subtracts a preset data from the data size, and uses the calculated result data as the representative data corresponding to the obfuscated data block. The preset data is the data size of the target data mentioned above and is a fixed value, such as the decimal number 2, corresponding to 2 bytes. A preset operation is performed on the representative data to obtain the process data. In S232, the optimization module determines whether the process data is greater than the representative data. If so, the representative data is subtracted from the process data, and this step is repeated until the process data is less than or equal to the representative data. If so, data is removed from the obfuscated data block based on the process data. For ease of understanding, for example, if the process data is 2, then the data byte after the second byte from the left of the obfuscated data is deleted. It should be noted that when continuing the deletion process, the obfuscated data that has undergone the deletion process is referenced, not the original obfuscated data. In S233, the optimization module determines whether the stopping condition is met. The stopping condition is, for example, that the data size of the obfuscated data is equal to the representative data. If not, the process data is updated using the result data of the preset operation on the process data. That is, the result data of the preset operation on the process data is regarded as the new process data, and the process is skipped to S232 to continue. If not, the obfuscated data is restored to obtain part of the sent data, and all steps are terminated. The restoration process can be understood as a decryption process using the symmetric encryption algorithm in the prior art.

[0076] According to another aspect of the embodiment of the present application, Figure 4 As shown, the present application also provides a smart park equipment energy consumption optimization system based on machine learning, including a monitoring module, a network module, and an optimization module to implement the smart park equipment energy consumption optimization method based on machine learning described above.

[0077] The functions of each module are as follows:

[0078] The monitoring module is used to use the real monitoring data of different days in history as the sending data, perform obfuscation processing on the sending data to obtain a number of obfuscated data blocks, and transmit the several obfuscated data blocks to the optimization module through the network module;

[0079] The network module is used to transmit several obfuscated data blocks generated by the monitoring module to the optimization module;

[0080] The optimization module is used to recover different partial transmission data from each obfuscated data block, obtain transmission data based on all partial transmission data, obtain various types of real condition data for different historical days, determine a number of observation days among different historical days, obtain a trained analysis algorithm, and perform improvement processing on the trained analysis algorithm; and is used to calculate various types of possible condition data for the target day based on various types of real condition data for different historical days, use the various types of possible condition data for the target day as input data of the improved analysis algorithm to obtain output data of the improved analysis algorithm, and formulate an optimization plan based on the output data.

[0081] According to another aspect of an embodiment of the present application, a medium is further provided, wherein the medium stores program instructions, wherein when the program instructions are executed, the device where the medium is located is controlled to execute any one of the above methods.

[0082] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0083] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program code.

[0084] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for optimizing energy consumption of smart park equipment based on machine learning, characterized in that: The method comprises the following steps: S1. The monitoring module uses real monitoring data from different days in history as sending data, performs obfuscation processing on the sending data to obtain several obfuscated data blocks, and transmits the several obfuscated data blocks to the optimization module through the network module; S2. The optimization module recovers different partial transmission data from each obfuscated data block, obtains transmission data based on all the partial transmission data, obtains various types of real condition data from different historical days, determines a number of observation days from the different historical days, and obtains a trained analysis algorithm and performs improvement processing on the trained analysis algorithm. S3. The optimization module calculates various types of possible condition data for the target day based on various types of actual condition data for different days in history, and uses the various types of possible condition data for the target day as input data for an improved analysis algorithm to obtain output data of the improved analysis algorithm, and formulates an optimization plan based on the output data. The monitoring module performs obfuscation processing on the sent data, including the following steps: S11, the monitoring module divides the transmission data into a plurality of transmission data blocks, and for each transmission data block, generates secret data corresponding to the transmission data block, and further for each transmission data block, uses the data size of the transmission data block as representative data; S12. For each transmitted data block, the monitoring module performs secret processing on the transmitted data block using the secret data corresponding to the transmitted data block to obtain an intermediate data block; S13. For each intermediate data block, the monitoring module adds the secret data corresponding to the previous intermediate data block to the intermediate data block based on the representative data corresponding to the intermediate data block to obtain an obfuscated data block; The monitoring module adds secret data corresponding to the previous intermediate data block to the intermediate data block based on the representative data corresponding to the intermediate data block, including the following steps: S131, the monitoring module performs a preset operation on the secret data corresponding to the previous intermediate data block to obtain target data, and performs a preset operation on the representative data corresponding to the intermediate data block to obtain process data; S132: The monitoring module determines whether the process data is larger than the data size of the intermediate data block. If so, the monitoring module subtracts the data size of the intermediate data block from the process data and repeats this step. If less than or equal to the process data, a portion of the target data is added to the intermediate data block according to the process data. S133, the monitoring module determines whether the complete target data has been added to the intermediate data block. If not, the process data is updated using the result data of the preset operation on the process data, and the process jumps to S132. If yes, all steps are completed.

2. The method according to claim 1, characterized in that The optimization module performs an improvement process on the trained analysis algorithm, including the following steps: S211. For each observation day, the optimization module calculates various types of possible condition data for the observation day based on various types of real condition data from different days in history; S212. For each observation day, use various types of possible condition data on the observation day as input data of the trained analysis algorithm to obtain output data of the trained analysis algorithm, and calculate the absolute value of the difference between the output data and the actual monitoring data of the observation day; S213. The optimization module calculates a first ratio of the sum of all absolute values ​​to the total number of all absolute values, and the optimization module determines whether the first ratio is less than a preset first ratio threshold. If it is less than, all steps are terminated. If it is greater than or equal to, a first improvement process is performed on the trained analysis algorithm, and the process jumps to S212.

3. The method according to claim 2, characterized in that The optimization module improves the trained analysis algorithm and further includes the following steps: S221. For each type of real condition data on different days in history, the optimization module calculates a second ratio of the sum of all real condition data to the total number of all real condition data, calculates the square value of the difference between each real condition data and the second ratio, and further calculates the square root value of a third ratio of the sum of all square values ​​to the total number of all square values. S222, the optimization module calculates various types of possible condition data for the target day based on various types of real condition data for different days in history; S223. For each type of possible condition data on the target day, the optimization module calculates the absolute value of the difference between the possible condition data and the corresponding second ratio, and then calculates a fourth ratio of the absolute value to the corresponding square root value. S224 , performing a second improvement process on the trained analysis algorithm based on the fourth ratios corresponding to various types of possible conditional data of the target day.

4. The method according to claim 1, wherein The optimization module recovers part of the transmitted data from the obfuscated data block, including the following steps: S231, the optimization module determines the data size of the obfuscated data block, subtracts the preset data from the data size as the representative data corresponding to the obfuscated data block, and performs a preset operation on the representative data to obtain process data; S232: The optimization module determines whether the process data is greater than the representative data. If so, the representative data is subtracted from the process data, and this step is repeated. If less than or equal to the representative data, data is removed from the obfuscated data block based on the process data. S233, the optimization module determines whether the stop condition is met. If not, the process data is updated using the result data of the preset operation on the process data, and the process jumps to S232. If the condition is met, the obfuscated data is restored to obtain part of the sent data, and all steps are completed.

5. A smart park equipment energy consumption optimization system based on machine learning, used to implement the method according to any one of claims 1 to 4, characterized in that: Includes the following modules: The monitoring module is used to use the real monitoring data of different days in history as the sending data, perform obfuscation processing on the sending data to obtain a number of obfuscated data blocks, and transmit the several obfuscated data blocks to the optimization module through the network module; The network module is used to transmit several obfuscated data blocks generated by the monitoring module to the optimization module; an optimization module for recovering different partial transmission data from each obfuscated data block, obtaining transmission data based on all partial transmission data, obtaining various types of real condition data for different historical days, determining a number of observation days among the different historical days, obtaining a trained analysis algorithm, and performing improvement processing on the trained analysis algorithm; It is also used to calculate various types of possible condition data for the target day based on various types of real condition data on different days in history, and use the various types of possible condition data for the target day as input data for the improved analysis algorithm to obtain output data of the improved analysis algorithm, and formulate an optimization plan based on the output data.

6. A medium, characterized in that The medium stores program instructions, wherein when the program instructions are executed, the device where the medium is located is controlled to execute the method according to any one of claims 1 to 4.

Images