Power-saving excitation method, system and equipment considering user feedback and medium

Abstract

The invention discloses a power-saving excitation method, system and device considering user feedback, and a medium. The method comprises the following steps: collecting user electrical load data in a target area to generate an electrical load sequence; inputting the electrical load sequence into a preset clustering algorithm to determine clustering cluster centers, and taking each clustering cluster center as a corresponding benchmark user; extracting multi-dimensional power consumption characteristics of each benchmark user, and constructing a power-saving incentive measure library according to the multi-dimensional power consumption characteristics; inputting the power utilization characteristics of a target user into a neural network model obtained by training the multi-dimensional power utilization characteristics and a power-saving excitation measure library, and outputting a target power-saving excitation measure matched with the target user; and dynamically adjusting and implementing the target power-saving incentive measure according to the feedback information of the user and the quantified evaluation result so as to improve the power-saving effect of the residential user.

Description

Technical Field

[0001] The present invention relates to the technical field of power saving, and in particular to a power saving incentive method, system, device and medium considering user feedback. Background Art

[0002] To ensure the sustainable development of the power system, energy conservation is one of the most economical and effective ways. With the improvement of residents' living standards, the access of household electrical equipment has increased, and the electricity consumption growth rate is significant.

[0003] Household electricity consumption has the characteristics of random use, huge data, and small single-user scale, which is significantly different from industrial and commercial electricity loads. At present, the existing power saving incentive measures are not perfect and comprehensive for the energy conservation analysis of households, and it is difficult to meet the user needs, resulting in poor power saving effects.

[0004] Therefore, how to formulate reasonable and effective power saving incentive measures for the user side to improve the power saving effect of households has become a technical problem that needs to be urgently solved by those skilled in the art. Summary of the Invention

[0005] The present invention provides a power saving incentive method, system, device and medium considering user feedback, which solves the problem of how to formulate comprehensive power saving incentive measures by analyzing the electricity consumption characteristics of households, and dynamically adjusts and implements the incentive measures by fully considering the power saving achievements of user feedback, so as to significantly improve the power saving efficiency.

[0006] To solve the above technical problems, an embodiment of the present invention provides a power saving incentive method considering user feedback, including:

[0007] Collecting the user electricity load data in the target area to generate an electricity load sequence;

[0008] Inputting the electricity load sequence into a preset clustering algorithm to determine the clustering cluster centers, and taking each clustering cluster center as the corresponding benchmark user;

[0009] Extracting the multi-dimensional electricity consumption characteristics of each benchmark user, and constructing a power saving incentive measure library according to the multi-dimensional electricity consumption characteristics;

[0010] Inputting the electricity consumption characteristics of the target user into a neural network model trained with the multi-dimensional electricity consumption characteristics and the power saving incentive measure library, and outputting a target power saving incentive measure matching the target user;

[0011] Dynamically adjusting and implementing the target power saving incentive measure through the feedback information and quantified evaluation results of the user.

[0012] Further, the step of inputting the electricity consumption load sequence into a preset clustering algorithm to determine the cluster centers and using each cluster center as a corresponding benchmark user respectively includes:

[0013] Determining the target cumulative distance between every two electricity consumption load sequences through a dynamic programming algorithm;

[0014] Using the target cumulative distance as the input of the preset clustering algorithm, performing clustering analysis on the electricity consumption load sequences through a defined action space and reward function, and performing iterative update to determine the cluster centers;

[0015] Using each cluster center as a corresponding benchmark user respectively.

[0016] Further, the construction process of the power-saving incentive measure library includes:

[0017] Standardizing a plurality of the multi-dimensional electricity consumption characteristics into score data within a preset range;

[0018] Grouping the multi-dimensional electricity consumption characteristics based on user social factor categories, electricity price factor categories, environmental perception factor categories, and user behavior factor categories, and calculating the sensitivity of each benchmark user to each category of factors according to the score data;

[0019] Generating corresponding power-saving incentive measures according to the sensitivity and constructing a power-saving incentive measure library.

[0020] Further, the step of inputting the electricity consumption characteristics of a target user into a neural network model trained with the multi-dimensional electricity consumption characteristics and the power-saving incentive measure library and outputting a target power-saving incentive measure matching the target user includes:

[0021] Inputting the multi-dimensional electricity consumption characteristics and the power-saving incentive measure library into a preset neural network model for training;

[0022] Inputting the obtained actual electricity consumption characteristics of the target user into the trained neural network model, and calculating the similarity probability between the target user and each benchmark user;

[0023] Based on a preset probability threshold, determining a similar benchmark user corresponding to the target user;

[0024] Matching the multi-dimensional electricity consumption characteristics corresponding to the similar benchmark user in the power-saving incentive measure library to obtain the target power-saving incentive measure.

[0025] Further, the step of dynamically adjusting and implementing the target power-saving incentive measure through the feedback information of the user and the quantified evaluation result includes:

[0026] Construct a user feedback path through the user-side mobile device, and collect user feedback information from the user feedback path;

[0027] By comparing and analyzing the total actual electricity consumption in the target area after implementing the corresponding target power-saving incentive measures with the pre-calculated theoretical electricity consumption, determine the area evaluation result;

[0028] Quantitatively analyze the power-saving effects before and after implementing the corresponding target power-saving incentive measures for each user in the target area to determine the individual evaluation result;

[0029] Dynamically adjust the target power-saving incentive measures according to the feedback information, the area evaluation result, and the individual evaluation result.

[0030] Further, the quantitative analysis of the power-saving effects before and after implementing the corresponding target power-saving incentive measures for each user in the target area includes:

[0031] Collect the subjective power-saving behavior data and objective environment data of each user in the target area, use the entropy weight method to analyze the weights of the subjective power-saving behavior data and the objective environment data, and determine the user's power-saving potential according to the weight analysis results;

[0032] Real-time collect the load data before and after the implementation of the target power-saving incentive measures, and calculate the electricity consumption difference according to the load data;

[0033] Determine the actual power-saving effect of the user according to the electricity consumption difference and the user's power-saving potential;

[0034] Combine the multi-dimensional electricity consumption characteristics and the user's power-saving potential for baseline analysis to determine the user's theoretical power-saving effect.

[0035] Further, the process of implementing the target power-saving incentive measures includes:

[0036] Use the user-side mobile device as the push medium to push personalized power-saving suggestions composed of the target power-saving incentive measures to the users in the target area.

[0037] Another embodiment of the present invention provides a power-saving incentive system considering user feedback, including:

[0038] A data collection module for collecting the user electricity load data in the target area to generate an electricity load sequence;

[0039] A clustering analysis module for inputting the electricity load sequence into a preset clustering algorithm to determine the clustering cluster centers, and taking each clustering cluster center as the corresponding benchmark user;

[0040] A measure library construction module, configured to extract multi-dimensional power consumption characteristics of each benchmark user, and construct a power-saving incentive measure library according to the multi-dimensional power consumption characteristics;

[0041] A measure matching module, configured to input the power consumption characteristics of a target user into a neural network model trained with the multi-dimensional power consumption characteristics and the power-saving incentive measure library, and output a target power-saving incentive measure matching the target user;

[0042] A measure optimization module, configured to dynamically adjust and implement the target power-saving incentive measure through user feedback information and a quantified evaluation result.

[0043] Another embodiment of the present invention provides a computer device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the power-saving incentive method considering user feedback as described above is implemented.

[0044] Another embodiment of the present invention provides a computer-readable storage medium. The computer-readable storage medium stores a computer program. When the device where the computer-readable storage medium is located executes the computer program, the power-saving incentive method considering user feedback as described above is implemented.

[0045] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following:

[0046] In the embodiments of the present invention, by using the dynamic time rule enhanced clustering method to cluster the load curves of users, the calculation efficiency can be improved while avoiding the problem of being sensitive to the offset of the user load time series; according to the clustering results, the power consumption characteristics of typical users are extracted to generate diversified power-saving incentive measures, thereby realizing the construction of a multi-dimensional, wide-coverage and accurate power-saving incentive measure library. Using the CNN model to determine the optimal measures corresponding to similar benchmark users to improve the power-saving effect of residential users, the flexibility of selecting power-saving measures is improved, and the power-saving measures are dynamically adjusted in combination with user feedback information and effectiveness evaluation to fully meet user needs, thereby improving the power-saving effect of residential users. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a schematic flowchart of a power-saving incentive method considering user feedback in one embodiment of the present invention;

[0048] Figure 2 is an operation flowchart of a clustering algorithm in one embodiment of the present invention;

[0049] Figure 3 is a flowchart for constructing a power-saving incentive measure library in one embodiment of the present invention;

[0050] Figure 4 It is a schematic structural diagram of a power-saving incentive system considering user feedback in one embodiment of the present invention;

[0051] Figure 5 It is a structural block diagram of a preferred embodiment of a computer device provided by the present invention. Detailed implementation manners

[0052] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0053] In the description of the present application, the terms "first", "second", "third", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first", "second", "third", etc. may explicitly or implicitly include one or more of such features. In the description of the present application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0054] In the description of the present application, it should be noted that unless otherwise clearly defined and limited, the terms "installed", "connected", "connected" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected, or indirectly connected through an intermediate medium, and it may be the communication inside two elements. The terms "vertical", "horizontal", "left", "right", "up", "down" and similar expressions used herein are only for the purpose of illustration, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation of the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0055] In the description of the present application, it should be noted that unless otherwise defined, all the technical and scientific terms used in the present invention have the same meanings as those commonly understood by those of ordinary skill in the technical field to which the present invention belongs. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0056] An embodiment of the present invention provides a power-saving incentive method considering user feedback. Specifically, please refer to Figure 1 , Figure 1 which shows a schematic flow chart of the power-saving incentive method considering user feedback in one embodiment of the present invention, including the following steps:

[0057] S1. Collect user power consumption load data in the target area to generate a power consumption load sequence.

[0058] In the target area, through the smart meters deployed on the residential user side, the user power consumption load curve data is collected in real time, including the hourly / daily power consumption, the power consumption ratio during peak and valley periods, etc. The collected data is subjected to data cleaning and normalization. For example, by removing abnormal data (such as zero values or extreme values caused by equipment failures) and normalizing the user load data to eliminate the dimension difference, it is convenient for subsequent clustering analysis to form a standardized time series, that is, the power consumption load sequence.

[0059] S2. Input the power consumption load sequence into a preset clustering algorithm to determine the clustering cluster centers, and use each clustering cluster center as the corresponding benchmark user.

[0060] Since the residential user data is huge and complex, in order to effectively promote power saving on the user side, clustering to obtain similar user groups and extracting typical user power consumption curves and their power consumption characteristics have become the basis for power-saving work. And because the traditional Kmeans algorithm has a problem of decreasing calculation efficiency when dealing with large data samples, and is sensitive to the offset of the user load time series, it cannot align similar but out-of-phase load curves, and the number of clustering clusters also needs to be given in advance. In order to adapt to the characteristics of the user load curve, this embodiment preferably uses the dynamic time rule enhanced clustering method (DTW-KMeans-RL) to perform clustering analysis on the user's load sequence data.

[0061] It can be understood that using the cumulative distance after dynamic time warping (DTW) to replace the traditional Euclidean distance can "elastically align" the time axes of different sequences and solve the time series alignment problem. In this way, the time series characteristics of the load can be taken into account to obtain a higher-quality clustering result. After replacing with the DTW cumulative distance, the clustering misjudgment caused by time offset can be avoided.

[0062] In order to dynamically adapt to the change of data distribution. The embodiment of the present invention combines the way of reinforcement learning Q-learning. By designing a reward function and taking into account both the clustering quality index and the power-saving potential index, it is possible to classify and obtain a user group that more meets the actual needs, that is, the benchmark user. For the specific clustering analysis process, please refer to Figure 2 as shown, Figure 2It is shown as the operation flow chart of the clustering algorithm in one embodiment of the present invention.

[0063] Combined with Figure 2 It can be seen that in this embodiment, the target cumulative distance between every two electrical load sequences is determined by the dynamic programming algorithm. In some embodiments of the present invention, the Euclidean distance between each pair of load points in every two sequences is calculated to obtain the distance matrix M. Based on the determined distance matrix M, starting from the starting point (1, 1) of the matrix until the end point (m, n), the sum of the distances of all points on the path is minimized. The DTW algorithm is used to calculate the minimum cumulative distance between every two electrical load sequences, that is, the target cumulative distance, and the corresponding alignment path.

[0064] Next, clustering analysis is performed on the electrical load sequences with the target cumulative distance as the input of the preset clustering algorithm. In this embodiment, first, the number of clusters needs to be optimized by reinforcement learning. Specifically:

[0065] Exemplarily, this embodiment can select the elbow method to preselect the initial value of the current number of clusters K, such as K = 3. Randomly select or select representative sequences from the DTW distance matrix as the cluster centers to construct a set of cluster centers {C1, C2,..., C K}. And determine the corresponding clustering quality indicators (including silhouette coefficient, Davies - Bouldin coefficient, Calinski - Harabasz coefficient, etc.).

[0066] Then, define the action space and design a reward function based on the power saving potential of the user: Set three actions in the action space, namely increasing the number of clusters (K ← K + 1), decreasing the number of clusters (K ← K - 1), and maintaining the current number of clusters. In this embodiment, the reward function is designed by weighing the clustering quality and the power saving potential, and the calculation formula is:

[0067] R = αI + βD P-V + γP e (0.1)

[0068] Where α, β, γ are weight coefficients, satisfying α + β + γ = 1; I, D P-V , P e are the mean values of the within - cluster silhouette coefficient (or DB coefficient, CH coefficient), the mean value of the peak - valley difference, and the mean value of the power saving potential respectively. (The process of obtaining the power saving potential will be elaborated in the following step S5)

[0069] Iterative update is performed through reinforcement learning decision - making (Q - learning). After each round of iteration, an action (adjust K) is selected according to the current state, and the Q - table or the policy network is updated. When the reward function has no significant change for N consecutive rounds (such as ΔR < 10 -4) or when the preset maximum number of iterations is reached, stop the iteration, determine the optimal number of clusters K and the cluster centers, and use the cluster centers of each cluster as the corresponding benchmark users respectively.

[0070] In this embodiment, clustering analysis of DTW alignment collaborative reinforcement learning is performed on the load data of users, so as to ensure that the clustering results not only conform to the data distribution characteristics but also serve the energy-saving goal, laying a foundation for the formulation of subsequent personalized incentive measures.

[0071] S3. Extract the multi-dimensional electricity consumption characteristics of each benchmark user, and construct an electricity-saving incentive measure library according to the multi-dimensional electricity consumption characteristics.

[0072] In order to guide the construction of the electricity-saving incentive measure library and the selection of the optimal strategy, it is also necessary to have practical significance and reflect the characteristics of users' electricity consumption behaviors. Based on this, after determining the benchmark users in this embodiment, the multi-dimensional electricity consumption characteristics of various benchmark users will be extracted. Exemplarily, this embodiment will extract 17-dimensional features, including monthly average electricity consumption, annual maximum utilization hours, residential load density, typical daily average load rate, quarterly imbalance coefficient, and peak-valley electricity ratio. They can reflect the social factors, electricity price level factors, environmental perception factors, and electricity consumption behavior factors of residential electricity consumption. Combining these characteristic data can comprehensively obtain the annual electricity consumption characteristics of users. Some electricity consumption characteristics can be referred to Table 1 below:

[0073] Table 1 User Load Characteristic Index

[0074]

[0075] For the above multi-dimensional electricity consumption characteristics, various different incentive strategies can be formed from the above four factor perspectives. Exemplarily: For users with a high typical daily average load rate but a low load density on national holidays and other typical days, such users are sensitive to social factors and should adopt electricity-saving incentive measures for holiday demand response. For users with a high peak-valley electricity ratio and a high load density, such users are sensitive to electricity price level factors and should adopt electricity-saving incentive measures to adjust the time-of-use electricity price. For users with significantly higher loads in summer and winter, such users are sensitive to environmental perception factors, and electricity-saving suggestions for adjusting the temperature and usage time of air conditioners and electric heating can be proposed. Specifically, please refer to Figure 3 as shown, Figure 3 shows the flow chart for constructing the electricity-saving incentive measure library in one embodiment of the present invention.

[0076] From Figure 3It can be seen that in this embodiment, several of the multi-dimensional power consumption characteristics will be scored, and the power consumption characteristic data will be standardized into score data within a preset range. Exemplarily, the maximum value of a certain characteristic in the multi-dimensional power consumption characteristics among all benchmark users is denoted as 1, and the minimum value is denoted as 0. Further, if the "peak-valley electricity ratio" range of the benchmark users is 1.5 to 4.0, then a user with a peak-valley ratio of 4.0 gets 1 point, a user with 1.5 gets 0 points, and a user with a peak-valley ratio of 2.75 gets 0.5 points. It should be noted that for other users who are not benchmark users in this implementation, the characteristic scores are calculated by linear interpolation.

[0077] Based on user social factors, electricity price factors, environmental perception factors, and user behavior factors as benchmarks, the multi-dimensional power consumption characteristics are grouped. Specifically, the following examples can be referred to:

[0078] Social factors: Typical daily average load rate, social activity correlation.

[0079] Electricity price level: Peak-valley electricity ratio, participation rate in electricity price sensitive periods.

[0080] Environmental perception: Quarterly imbalance coefficient, climate sensitivity coefficient.

[0081] Power consumption behavior: High-power device usage duration, penetration rate of energy-saving devices.

[0082] In some embodiments of the present invention, the sensitivity of each benchmark user to each type of factor is also calculated based on the score data. For example, if the scores of the relevant power consumption characteristics reflecting social factors of user A are 0.8, 0.7, and 0.9, then the social sensitivity can be the average of the three: 0.8. Further, by setting a threshold (such as ≥0.7), users can be marked as highly sensitive to a certain type of factor. Corresponding electricity-saving incentive measures are generated based on the sensitivity, and an electricity-saving incentive measure library is constructed based on the generated electricity-saving incentive measures. Some incentive measures can be referred to as shown in Table 2 below:

[0083] Table 2 Sensitive factors and their corresponding incentive measures

[0084]

[0085] S4. Input the power consumption characteristics of the target user into the neural network model trained with the multi-dimensional power consumption characteristics and the electricity-saving incentive measure library, and output the target electricity-saving incentive measure matching the target user.

[0086] Input the multi-dimensional electricity consumption characteristics and the power-saving incentive measure library into a preset neural network model for training. In this embodiment, a convolutional neural network (CNN) model is preferably used. By establishing a CNN model, based on the actual electricity consumption characteristics of the target user obtained, input them into the trained neural network model to calculate the similarity probability between the actual target user and each benchmark user, and then realize the selection of power-saving incentive measures. Before this, in this embodiment, 100 benchmark users of each type are preferably selected, and 17 of their characteristics are normalized. It should be noted that for a certain benchmark user, the probability of belonging to the corresponding type is 1, and the probability of belonging to other types is 0. By processing the samples, the probability vectors of all users are obtained.

[0087] In the process of metal feature matching through the CNN model, a preset probability threshold is used to determine the similar benchmark user corresponding to the target user. In this embodiment, the probability threshold is preferably set to 0.5, that is, the elements in the probability vector greater than 0.5 are recorded as similar benchmark users. Match the multi-dimensional electricity consumption characteristics corresponding to the similar benchmark user in the power-saving incentive measure library to obtain the target power-saving incentive measure as the incentive measure for this actual user. Example: The CNN output of user B is 0.2, 0.7, 0.1, which is matched to the incentive measure sensitive to electricity price (such as dynamic time-of-use electricity price).

[0088] S5. Dynamically adjust and implement the target power-saving incentive measure based on the feedback information and quantified evaluation results of the user.

[0089] In order to further optimize the power-saving incentive measures so that they can fully meet the needs of the user side. In this embodiment, the feedback of the user is used as an index to dynamically adjust to obtain more user-demand-adapted power-saving incentive measures.

[0090] Specifically, a user feedback path is constructed through the user-side mobile terminal, and user feedback information is collected from the user feedback path. Exemplarily, in this embodiment, the mobile terminal platform (such as a web page, a small program, etc.) is used to collect user feedback on measures such as satisfaction and implementation difficulty. Users can view the electricity consumption situation of their families and the proposed power-saving measures. If users find that the data records are incorrect, they can provide feedback to the power service enterprise through the platform. For different power-saving measures, if users think that they have damaged their electricity consumption satisfaction, they can also feedback to the enterprise and put forward their expected incentive measures. For example, user C feedback: "The time-of-use electricity price period is unreasonable", then mark this measure as needing optimization. In this process, the user-side mobile terminal can be used as a push medium to push personalized power-saving suggestions composed of the target power-saving incentive measures to users in the target area, so as to visually present various data to users / service enterprises, and at the same time provide a real-time feedback channel for users to improve the power-saving effect of personalized incentive measures.

[0091] For each user, since there is a difference between the actual power-saving effect and the theoretical power-saving effect, the smaller the difference, the more satisfied the user is with the given incentive measures, and the more likely the user is to execute the power-saving behavior according to the rational strategy, indicating that the power-saving incentive measure is better; conversely, the larger the difference, the more likely it is that the theoretical power-saving measure has caused dissatisfaction among users or made it difficult for users to execute, indicating that the power-saving incentive measure is worse. Based on this, in this embodiment, a quantitative analysis is performed on the power-saving effects before and after implementing the corresponding target power-saving incentive measures for each user in the target area to determine the individual evaluation results. Similarly, in this embodiment, by comparing and analyzing the total actual power consumption in the target area after implementing the corresponding target power-saving incentive measures with the pre-calculated theoretical power consumption, the area evaluation results are determined. The theoretical power consumption is the power consumption generated when residential users can rationally respond to electricity price signals or policy incentives and take power-saving behaviors during the theoretically most reasonable time periods under the condition of complete rationality.

[0092] Dynamically adjust the target power-saving incentive measures according to the feedback information, the area evaluation results, and the individual evaluation results. Exemplarily, inefficient measures with an actual power-saving effect < 60% of the theoretical value and a satisfaction level < 3 points are eliminated. For measures with an effectiveness index > 0.9, expand the coverage of users or increase the reward intensity. Then, input the latest data after dynamic adjustment into the clustering algorithm and the CNN model to update the cluster classification and further perform strategy matching.

[0093] It should be noted that in some embodiments of the present invention, the process of quantitatively analyzing the power-saving effect is specifically as follows:

[0094] Collect the subjective and objective dimension power-saving behavior data of each user in the target area, including subjective power-saving behavior data and objective environment data. In this embodiment, qualitative indicators such as the user's power-saving awareness information and behavior habits (such as whether they are accustomed to using high-power electrical appliances during peak hours) obtained through questionnaires are collected; and quantitative indicators such as the user's electrical appliance configuration (such as the energy efficiency level of electrical appliances) and family structure are obtained. Perform preprocessing such as standardization on the obtained power-saving behavior data to eliminate the dimension difference.

[0095] Furthermore, in this embodiment, the entropy weight method is used to perform weight analysis on the preprocessed subjective power-saving behavior data and objective environment data, and the user's power-saving potential Q is determined according to the weight analysis result, which is expressed as follows:

[0096]

[0097] where w j is the weight of the jth index calculated by the entropy weight method, and X j is the standardized index value.

[0098] Exemplarily, according to the calculation results of the above formula, the power-saving potential is classified. Those with a power-saving potential greater than 0.8 are classified as high potential, those between [0.5, 0.8] are classified as medium potential, and those below 0.5 are classified as low potential.

[0099] Load data before and after the implementation of the target power-saving incentive measure are collected in real time, and the power consumption difference is calculated based on the load data. The actual power-saving benefit S of the user is calculated with this power consumption difference value, expressed as:

[0100] S = ΔE·P - C

[0101] In the formula, ΔE is the power consumption difference; P is the electricity price; C is the implementation cost of the incentive measure (such as equipment subsidy).

[0102] According to the above process of quantifying the power consumption difference value, the actual power-saving cost benefit, and the user's power-saving potential, this embodiment realizes the quantification of the user's actual power-saving effect.

[0103] Dynamic baseline analysis is carried out by combining multi-dimensional electricity consumption characteristics and the user's power-saving potential to fit the theoretical maximum reasonable power consumption to determine a baseline. The actual power consumption is compared with this baseline, the corresponding theoretical power consumption difference is calculated, and the product of the two is calculated in combination with the electricity price to determine the theoretical power-saving benefit. In this embodiment, this baseline is used as the theoretical power consumption.

[0104] In some embodiments of the present invention, the theoretical-actual difference can be further quantified. Exemplarily, it can be calculated by the following formula:

[0105] Theoretical - actual difference = 1 - |theoretical value - actual value| / theoretical value

[0106] Furthermore, based on the theoretical-actual difference, a quantitative analysis result of the difference between the actual power-saving effect and the theoretical power-saving effect is determined.

[0107] In this embodiment, by converting the user's electricity consumption characteristics into quantifiable sensitivity indicators, it drives the dynamic construction and optimization of the incentive measure library. From data standardization to measure generation, and then to feedback-driven iterative update, a closed-loop management is formed to ensure that the power-saving measures always fit the actual needs of users, improving the energy-saving effect and user satisfaction.

[0108] In summary, in the embodiments of the present invention, by clustering the load data of a large number of residential users, diverse electricity consumption characteristics are determined. Secondly, combined with the electricity consumption characteristics of typical users, multi-faceted scoring is carried out in aspects such as environmental perception and social factors to obtain the best incentive measures suitable for typical benchmark users; using artificial intelligence algorithms such as convolutional neural networks, for each independent user, the best incentive measures are selected from the power-saving strategy library; finally, according to the satisfaction and evaluation results feedback by users, dynamic optimization is carried out to obtain more user-demand-adapted power-saving incentive measures.

[0109] An embodiment of the present invention provides a power-saving incentive system considering user feedback. Specifically, please refer to Figure 4 , Figure 4 which shows a schematic structural diagram of the power-saving incentive system considering user feedback in one of the embodiments of the present invention, including:

[0110] A data acquisition module M1, configured to collect user electricity load data within a target area to generate an electricity load sequence;

[0111] A clustering analysis module M2, configured to input the electricity load sequence into a preset clustering algorithm to determine the clustering cluster centers, and use each clustering cluster center as the corresponding benchmark user;

[0112] A measure library construction module M3, configured to extract multi-dimensional electricity consumption characteristics of each benchmark user, and construct a power-saving incentive measure library according to the multi-dimensional electricity consumption characteristics;

[0113] A measure matching module M4, configured to input the electricity consumption characteristics of a target user into a neural network model trained with the multi-dimensional electricity consumption characteristics and the power-saving incentive measure library, and output a target power-saving incentive measure matching the target user;

[0114] A measure optimization module M5, configured to dynamically adjust and implement the target power-saving incentive measure through the feedback information of the user and the quantified evaluation result.

[0115] As Figure 5 shown, an embodiment of the present invention also provides a computer device, Figure 5 which is a structural block diagram of a preferred embodiment of the computer device provided by the present invention. The computer device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the power-saving incentive method considering user feedback as described above is implemented.

[0116] Preferably, the computer program may be divided into one or more modules / units (such as computer program 1, computer program 2, ……), and the one or more modules / units are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program in the computer device.

[0117] The processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may also be any conventional processor. The processor is the control center of the terminal device and connects various parts of the terminal device through various interfaces and circuits.

[0118] The memory mainly includes a program storage area and a data storage area. Among them, the program storage area may store an operating system, application programs required for at least one function, etc., and the data storage area may store relevant data, etc. In addition, the memory may be a high-speed random access memory, or may also be a non-volatile memory, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., or the memory may also be other volatile solid-state storage devices.

[0119] It should be noted that the above terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that Figure 5The structural block diagram is only an example of the terminal device and does not constitute a limitation on the terminal device. It may include more or fewer components than shown, or combine certain components, or different components. Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it may include the processes of the embodiments of the above methods. Among them, the storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0120] Correspondingly, an embodiment of the present invention provides a computer-readable storage medium. The computer-readable storage medium includes a stored computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute the steps in the power-saving incentive method considering user feedback in the above embodiment, for example Figure 1 the steps S1 to S5 described therein.

[0121] The technical features and technical effects of the power-saving incentive system considering user feedback proposed in the embodiments of the present invention are the same as those of the power-saving incentive method considering user feedback proposed in the embodiments of the present invention, and will not be elaborated herein.

[0122] The above embodiments only represent several implementation manners of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention should be subject to the appended claims.

Claims

1. A power-saving incentive method considering user feedback, characterized in that Including: Collecting the user electricity load data in the target area to generate an electricity load sequence; Inputting the electricity load sequence into a preset clustering algorithm to determine the cluster centers, and taking each cluster center as the corresponding benchmark user; Extracting the multi-dimensional electricity consumption characteristics of each benchmark user, and constructing an energy-saving incentive measure library according to the multi-dimensional electricity consumption characteristics; Inputting the electricity consumption characteristics of the target user into a neural network model trained with the multi-dimensional electricity consumption characteristics and the energy-saving incentive measure library, and outputting a target energy-saving incentive measure matching the target user; Dynamically adjusting and implementing the target energy-saving incentive measure through the feedback information of the user and the quantified evaluation result.

2. The power-saving incentive method considering user feedback according to claim 1, wherein The step of inputting the electricity load sequence into a preset clustering algorithm to determine the cluster centers, and taking each cluster center as the corresponding benchmark user includes: Determining the target cumulative distance between every two electricity load sequences through a dynamic programming algorithm; Taking the target cumulative distance as the input of the preset clustering algorithm, performing clustering analysis on the electricity load sequences through the defined action space and reward function, and performing iterative update to determine the cluster centers; Taking each cluster center as the corresponding benchmark user.

3. The power-saving incentive method considering user feedback according to claim 1, wherein The construction process of the energy-saving incentive measure library includes: Standardizing several multi-dimensional electricity consumption characteristics into score data within a preset range; Grouping the multi-dimensional electricity consumption characteristics based on user social factor category, electricity price factor category, environmental perception factor category, and user behavior factor category, and calculating the sensitivity of each benchmark user to each category of factors according to the score data; Generating corresponding energy-saving incentive measures according to the sensitivity, and constructing an energy-saving incentive measure library.

4. The power-saving incentive method considering user feedback according to claim 1, wherein The step of inputting the electricity consumption characteristics of the target user into a neural network model trained with the multi-dimensional electricity consumption characteristics and the energy-saving incentive measure library, and outputting a target energy-saving incentive measure matching the target user includes: Inputting the multi-dimensional electricity consumption characteristics and the energy-saving incentive measure library into a preset neural network model for training; Inputting the actual electricity consumption characteristics of the obtained target user into the trained neural network model, and calculating the similarity probability between the target user and each benchmark user; Based on a preset probability threshold, determining the similar benchmark user corresponding to the target user; Matching the multi-dimensional electricity consumption characteristics corresponding to the similar benchmark user in the energy-saving incentive measure library to obtain the target energy-saving incentive measure.

5. The power-saving incentive method considering user feedback according to claim 1, characterized in that, The step of dynamically adjusting and implementing the target energy-saving incentive measure through the feedback information of the user and the quantified evaluation result includes: Constructing a user feedback path through the user-side mobile terminal, and collecting the feedback information of the user from the user feedback path; Determining the regional evaluation result by comparing and analyzing the total actual electricity consumption in the region after implementing the corresponding target energy-saving incentive measure with the pre-calculated theoretical electricity consumption; Quantitatively analyzing the energy-saving effectiveness before and after implementing the corresponding target energy-saving incentive measure for each user in the target region to determine the individual evaluation result; Dynamically adjusting the target energy-saving incentive measure according to the feedback information, the regional evaluation result, and the individual evaluation result.

6. The power saving incentive method considering user feedback according to claim 5, characterized in that, Quantitatively analyze the power-saving effects before and after implementing the corresponding target power-saving incentive measures for each user in the target area, including: Collect the subjective power-saving behavior data and objective environment data of each user in the target area, perform weight analysis on the subjective power-saving behavior data and the objective environment data using the entropy weight method, and determine the user's power-saving potential according to the weight analysis results; Real-time collect the load data before and after the implementation of the target power-saving incentive measures, and calculate the difference in power consumption according to the load data; Determine the actual power-saving effect of the user according to the difference in power consumption and the user's power-saving potential; Conduct baseline analysis by combining the multi-dimensional power consumption characteristics and the user's power-saving potential to determine the theoretical power-saving effect of the user.

7. The power-saving incentive method considering user feedback according to claim 1, characterized in that, The process of implementing the target power-saving incentive measures includes: Push personalized power-saving suggestions composed of the target power-saving incentive measures to users in the target area through the user-side mobile terminal as the push medium.

8. A power-saving incentive system considering user feedback, characterized in that Including: A data collection module for collecting the user power consumption load data in the target area to generate a power consumption load sequence; A clustering analysis module for inputting the power consumption load sequence into a preset clustering algorithm to determine the clustering cluster centers, and using each clustering cluster center as the corresponding benchmark user; A measure library construction module for extracting the multi-dimensional power consumption characteristics of each benchmark user and constructing a power-saving incentive measure library according to the multi-dimensional power consumption characteristics; A measure matching module for inputting the power consumption characteristics of the target user into a neural network model trained with the multi-dimensional power consumption characteristics and the power-saving incentive measure library, and outputting the target power-saving incentive measures matching the target user; A measure optimization module for dynamically adjusting and implementing the target power-saving incentive measures through the user's feedback information and quantitative evaluation results.

9. A computer device, characterized in that, Including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the power-saving incentive method considering user feedback as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program. When the device where the computer-readable storage medium is located executes the computer program, it implements the power-saving incentive method considering user feedback as described in any one of claims 1 to 7.

Images