Non-intrusive power load identification method and system based on intelligent fusion terminal

Abstract

The invention belongs to the technical field of data security, and particularly relates to a non-intrusive power load identification method and system based on an intelligent fusion terminal, and the method comprises the steps: collecting power load data, and dividing an analysis window based on the power load data; calculating a volatility index corresponding to the analysis window based on the data in the analysis window so as to represent the data complexity in the analysis window; determining the number of iterations for encryption based on the volatility index; and executing a compression algorithm on the data in the analysis window to generate compressed data, and executing multiple rounds of encryption perturbations determined by the number of iterations on the compressed data to generate a dynamically encrypted compressed data stream. According to the method, the privacy protection intensity can be adaptively adjusted according to the dynamic complexity of the data, and the privacy leakage risk caused by the fact that the static encryption intensity cannot adapt to the data dynamics in the prior art is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data security. More particularly, the present application relates to a non-intrusive power load identification method and system based on an intelligent fusion terminal. BACKGROUND

[0002] Non-intrusive load monitoring (NILM) technology collects aggregated data such as voltage and current of the total circuit through an intelligent terminal arranged at the user's power inlet, and then uses artificial intelligence algorithms to deeply analyze these aggregated data, thereby identifying the specific running state and energy consumption of each independent electrical appliance. NILM technology is a core and key technology for building future smart grids and realizing fine energy management of smart homes.

[0003] With the popularization of NILM technology, the privacy and security of user's power consumption data have become increasingly prominent. The existing technology proposes a protection method combining data compression and data encryption. For example, a classic compression algorithm such as LZ77 can be used to process power data streams with high repetition. The core idea of the LZ77 algorithm is to maintain a sliding window as a dynamic dictionary to find and replace repeated sequences in the data stream, and encode them into a triple (offset, match length, next data point), thereby achieving the purpose of compressing data. To achieve encryption, a chaotic sequence generated by a shared key between the two parties can be further used to disturb the numerical values of the offset and match length in the triple output after compression, thereby achieving the effect of encryption. This idea of combining compression and encryption processes takes into account the efficiency and security of data transmission to some extent.

[0004] However, user's daily power consumption behavior is not constant, and the characteristics of the generated load data show significant dynamics. When most electrical appliances are in standby or off state, the total load curve is very smooth and highly periodic, and the information entropy of the data is very low. If a fixed moderate intensity encryption disturbance is used at this time, the attacker may be able to infer the user's work and rest time, whether he is at home, and other highly sensitive personal privacy information by conducting long-term statistical analysis on the encrypted data stream. Conversely, during the power consumption peak period, the total load curve is complex and changes rapidly, and the data itself is close to a random state. If a fixed high-intensity encryption disturbance is still used at this time, it will bring an excessive computational burden to the relatively limited intelligent terminal, affecting the real-time response capability and energy efficiency of the system. Therefore, the existing technology urgently needs a mechanism that can dynamically adjust the privacy protection strength according to the sensitivity and complexity of the data itself. SUMMARY

[0005] To solve the technical problems of the prior art that there is a risk of privacy leakage in the flat power consumption period, and there is a waste of computing resources in the power consumption peak period, the present application provides solutions in the following aspects.

[0006] In a first aspect, the present application provides a non-intrusive power load identification method based on an intelligent fusion terminal, comprising: collecting power load data and dividing an analysis window based on the power load data; calculating a volatility index corresponding to the analysis window based on the data in the analysis window to represent the data complexity in the analysis window; determining the number of iterations for encryption based on the volatility index, wherein the higher the degree of deviation of the volatility index from the normal volatility level, the greater the number of iterations; executing a compression algorithm on the data in the analysis window to generate compressed data, and performing multiple rounds of encryption perturbation determined by the number of iterations on the compressed data to generate a dynamic encrypted compressed data stream, achieving adaptive adjustment of privacy protection strength.

[0007] The present application can adaptively adjust the encryption strength according to the dynamic complexity of the power data itself, automatically enhance the protection strength in periods with strong data regularity and high risk of privacy leakage, and reduce unnecessary computing overhead when the data itself is complex enough, thereby achieving a better balance between privacy security, system efficiency and load identification accuracy.

[0008] Preferably, the volatility index is the statistical variance of all power load data points in the analysis window.

[0009] Variance is an effective indicator of data dispersion, which can intuitively and efficiently quantify the dynamic complexity of the current data segment, and provide an accurate basis for subsequent sensitivity judgment.

[0010] Preferably, the number of iterations for encryption is determined based on the volatility index, comprising: calculating a privacy sensitivity index according to the difference between the volatility index and a preset normal volatility level reference value; determining the number of iterations according to the privacy sensitivity index and a basic iteration coefficient.

[0011] Preferably, the privacy sensitivity index satisfies the expression: ; wherein, is the privacy sensitivity index, is the volatility index, is the normal volatility level reference value, is a variance normalization factor of the volatility index itself, is a hyperbolic tangent function.

[0012] In this way, the leakage risk of data in different fluctuation states can be accurately reflected, and the privacy sensitivity is the highest when the data fluctuation is in an extremely stable or extremely volatile state, and the real scene is more consistent.

[0013] Preferably, the performing the multiple rounds of encryption perturbation on the compressed data determined by the iteration number comprises: generating a chaotic sequence according to a shared key; initializing the compressed data to be encrypted; and for each round from the first round to the iteration number, performing an exclusive or operation on the compressed data encrypted in the last round by using a value in the chaotic sequence to obtain an encryption result of the current round.

[0014] The iteration encryption perturbation mechanism can materialize the abstract sensitivity index into a specific executable encryption operation intensity, thereby providing a scalable and robust privacy protection means.

[0015] Preferably, the compression algorithm is an LZ77 compression algorithm, and the compressed data is a triple containing an offset and a matching length.

[0016] Preferably, the method further comprises: receiving the dynamically encrypted compressed data stream in the cloud; synchronously obtaining the iteration number according to the dynamically encrypted compressed data stream; and performing reverse de-perturbation operation of the iteration number to restore the compressed data and decompress the compressed data.

[0017] Since the encryption process adopts a symmetric exclusive or operation, the decryption process is also an exclusive or operation, which ensures the reversibility and efficiency of the algorithm, and the cloud can accurately restore the original compressed data by synchronously receiving the iteration number, thereby ensuring the integrity and consistency of the entire data processing link.

[0018] In a second aspect, the present application provides a non-intrusive power load identification system based on an intelligent fusion terminal, comprising a processor and a memory, the memory storing computer program instructions, when the computer program instructions are executed by the processor, the above-mentioned non-intrusive power load identification method based on an intelligent fusion terminal is realized.

[0019] By using the above technical solution, the above-mentioned non-intrusive power load identification method based on an intelligent fusion terminal is generated into a computer program and stored in a memory to be loaded and executed by a processor, so as to manufacture a terminal device according to the memory and the processor, and facilitate use.

[0020] The present application realizes the self-adaptation and intelligence of the privacy protection strength, directly links the fluctuation characteristics of the power data itself with the encryption perturbation strength, so that the encryption protection can be dynamically adjusted, in the period when the user's power consumption behavior is simple, regular, and the privacy leakage risk is high, the system will automatically increase the iteration number of the encryption perturbation, effectively resisting statistical analysis attacks on the data regularity.

[0021] Further, the application optimizes terminal computing resource allocation and improves system energy efficiency. When the power consumption behavior is complex and variable or in a regular state, the system will correspondingly adopt a lower encryption disturbance intensity, avoiding over-encryption in the case where the data itself has high randomness, significantly reducing the computing power consumption and processing delay of the intelligent fusion terminal, so that the system maintains high efficiency while providing effective privacy protection. BRIEF DESCRIPTION OF DRAWINGS

[0022] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description read in conjunction with the accompanying drawings. In the drawings, several embodiments of the application are shown by way of example, and like or corresponding elements are identified by like or corresponding reference numbers, in which: Figure 1 is a flowchart schematically showing a non-intrusive power load identification method based on an intelligent fusion terminal in the application; Figure 2 is a schematic diagram schematically showing the adaptive change of privacy protection intensity with data complexity; Figure 3 is a schematic diagram schematically showing data lossless recovery verification. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only some of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the application.

[0024] The specific embodiments of the application will be described in detail below with reference to the drawings.

[0025] The embodiments of the application disclose a non-intrusive power load identification method based on an intelligent fusion terminal, referring to Figure 1 , comprising steps S1-S4: S1, collecting power load data and dividing an analysis window based on the power load data.

[0026] Specifically, an intelligent fusion terminal deployed at a user power inlet can be used to continuously collect the total active power data of the user at a preset sampling frequency, for example, 1 Hz, to form an original time series data stream P={ , ,..., ,...} where is a floating point data. It is worth mentioning that in order to facilitate subsequent calculation, the floating point power values can be quantized, and in the embodiment, the power values are quantized to the nearest integer watt value by rounding off, to obtain the quantized data stream ={ , ,..., ,...}.

[0027] Next, in order to segmentally analyze the data stream, the continuous quantized data stream is cut into multiple analysis windows of fixed length and no overlap, and in the embodiment, the size W of each analysis window is set to 100 data points. Therefore, the first analysis window contains data points to , and so on, to obtain multiple analysis windows.

[0028] In this way, by collecting and windowing the continuous physical signal, it is converted into standardized discrete data units suitable for subsequent algorithm analysis, laying a foundation for dynamic analysis.

[0029] S2, based on the data in the analysis window, the volatility index corresponding to the analysis window is calculated to represent the data complexity in the analysis window.

[0030] In an optional embodiment, for the data in each analysis window, its volatility index can be calculated to represent the data complexity or regularity degree in the analysis window, wherein the volatility index is the statistical variance of all power load data points in the analysis window, and the calculation formula of the volatility index is as follows: wherein W is the size of the analysis window, is the i-th quantized power value in the analysis window, t is the last index in the current analysis window, is the first index in the current analysis window, is the arithmetic mean of all power values in the current analysis window, the greater the variance, the more dispersed the data points, and the more violent the power consumption behavior; the smaller the variance, the more concentrated the data points, and the more stable the power consumption behavior.

[0031] For example, assuming that the current analysis window W is 10, and the quantized power data in the window is {20, 22, 21, 20, 23, 22, 20, 21, 21, 20}, the corresponding mean is 21, and the variance is 1, that is, the volatility index of the current analysis window is 1.

[0032] Thus, by using variance as the volatility indicator, the dynamic complexity of the current data segment can be quantified intuitively and efficiently, providing an accurate basis for subsequent adaptive assessment of the privacy risk level.

[0033] S3, determining the number of iterations for encryption based on the volatility indicator, wherein the higher the degree of deviation of the volatility indicator from the normal volatility level, the greater the number of iterations.

[0034] In an optional embodiment, a privacy sensitivity index can be calculated according to the difference between the volatility indicator and the preset normal volatility level reference value, which is used to reflect that when the data volatility is extremely low (high regularity, easy to be analyzed) or extremely high (may contain important events), the risk of privacy leakage is higher, thus more in line with the real scene demand. The privacy sensitivity index satisfies the relationship: wherein, is the privacy sensitivity index, is the volatility indicator, is the normal volatility level reference value, is the variance normalization factor of the volatility indicator itself, is the hyperbolic tangent function. In this embodiment, by long-term statistical learning of a large amount of historical normal power consumption data, the reference value of the normal volatility level is set to 50, and the value of the variance normalization factor is set to 1000.

[0035] For example, when the power consumption is very stable, is 1, the result is close to 2, indicating that the privacy sensitivity is high; when the power consumption is close to the normal volatility, set to 55, then the corresponding is 1.025, close to 1, indicating that the privacy sensitivity is low at this time.

[0036] Further, the number of iterations required for the final encryption perturbation can be determined according to the privacy sensitivity index , and the corresponding calculation formula is: wherein k is a basic iteration coefficient for adjusting the overall encryption strength baseline, and in this embodiment, the value is 2, is a ceiling function to ensure that the number of iterations is an integer.

[0037] For example, in a high sensitivity scenario, rounds, and in a low sensitivity scenario, rounds. Therefore, the higher the degree of deviation of the volatility indicator from the normal volatility level, the greater the number of iterations. ​​

[0038] like Figure 2 The diagram illustrates how privacy protection strength adapts to data complexity. The low-fluctuation zone (green area) shows very stable data with a variance close to 0, thus identified as high privacy sensitivity, corresponding to a high encryption strength. The medium-fluctuation zone (orange background) shows data volatility close to the normal fluctuation baseline, thus judged as low privacy sensitivity, with the encryption strength significantly decreasing and stabilizing at the lowest baseline level. The high-fluctuation zone (red background) shows drastic data fluctuation, also identified as high privacy sensitivity, with a high encryption strength.

[0039] In this way, the physical volatility of data can be successfully mapped to the strength of encryption operations, providing a clear and reasonable logical basis for adjusting the encryption strength.

[0040] S4. Perform a compression algorithm on the data in the analysis window to generate compressed data, and perform multiple rounds of encryption perturbation on the compressed data, determined by the number of iterations, to generate a dynamically encrypted compressed data stream, thereby achieving adaptive adjustment of the privacy protection strength.

[0041] In an optional embodiment, the standard LZ77 compression algorithm is first applied to the data within the current analysis window to generate compressed data. This algorithm encodes duplicate data sequences into triples containing an offset o and a matching length l. , where s is the next unmatched data point.

[0042] Next, the generated compressed data, i.e., the offset o and matching length l in the triplet, are subjected to multiple rounds of encryption perturbation determined by the number of iterations, thereby generating a dynamically encrypted compressed data stream. The encryption process relies on a chaotic sequence generated by a key shared by both communicating parties, and each round of encryption is an XOR operation.

[0043] For example, suppose the triples generated from the current data point are , The required chaotic sequence value is: , , , Furthermore, if the offset 'o' is represented by 16 bits and the matching length 'l' is represented by 8 bits, then in highly sensitive scenarios, the execution... The encryption process consists of four rounds: initialization: First iteration: , Second iteration , Similarly, after the third iteration , After the fourth iteration , .

[0044] After 4 iterations, the final dynamic encrypted compressed data is the encryption triple, and the encrypted data stream is then transmitted to the cloud, and the cloud server performs the opposite process to recover the data: first synchronize the number of iterations, then perform the reverse de-randomization operation (the same XOR operation) for the number of iterations to recover the original triple, and finally perform the LZ77 decompression algorithm to recover the original power load data stream for subsequent load identification.

[0045] As Figure 3 shown is a schematic diagram of data lossless recovery verification, wherein the blue solid line is the original power load data input, and the red dashed line is the recovered data after the whole process of encryption, transmission, decryption and decompression.

[0046] In this way, by performing the number of iteration encryption matching the sensitivity level, stronger protection can be applied to high-risk data segments, while avoiding unnecessary over-encryption of low-risk data, and finally achieving a dynamic balance between data security and optimizing terminal computing resources.

[0047] The embodiment of the application also discloses a non-intrusive power load identification system based on an intelligent fusion terminal, comprising a processor and a memory, and the memory stores computer program instructions, which realize the non-intrusive power load identification method based on the intelligent fusion terminal when executed by the processor.

[0048] The above system also comprises a communication bus and a communication interface and other components well known to those skilled in the art, and the settings and functions thereof are known in the art, so they will not be described here.

[0049] In the description of the present specification, the meaning of "a plurality of", "several" is at least two, such as two, three or more, etc., unless otherwise explicitly specifically limited.

[0050] Although the present specification has shown and described several embodiments of the present application, it will be apparent to those skilled in the art that such embodiments are provided only in an exemplary manner. Those skilled in the art will think of many changes, changes and alternatives without departing from the idea and spirit of the present application. It should be understood that various alternatives to the embodiments of the application described herein can be employed in practicing the present application.

Claims

1. A non-intrusive power load identification method based on an intelligent fusion terminal, characterized in that, include: Collect power load data and divide the analysis window based on the power load data; Based on the data within the analysis window, a volatility index corresponding to the analysis window is calculated to characterize the data complexity within the analysis window; The number of iterations for encryption is determined based on the volatility index, wherein the greater the deviation of the volatility index from the normal volatility level, the greater the number of iterations. A compression algorithm is executed on the data within the analysis window to generate compressed data, and multiple rounds of encryption perturbation determined by the number of iterations are performed on the compressed data to generate a dynamically encrypted compressed data stream, thereby achieving adaptive adjustment of the privacy protection strength.

2. The non-intrusive power load identification method based on an intelligent fusion terminal according to claim 1, characterized in that, The volatility index is the statistical variance of all power load data points within the analysis window.

3. The non-intrusive power load identification method based on an intelligent fusion terminal according to claim 2, characterized in that, The step of determining the number of iterations for encryption based on the volatility index includes: The privacy sensitivity index is calculated based on the difference between the volatility index and the preset normal volatility level benchmark. The number of iterations is determined based on the privacy sensitivity index and the basic iteration coefficient.

4. The non-intrusive power load identification method based on an intelligent fusion terminal according to claim 3, characterized in that, The privacy sensitivity index satisfies the following relationship: in, For privacy sensitivity index, The volatility indicator, This is the baseline value for the normal fluctuation level. This is the variance normalization factor for the volatility index itself. It is the hyperbolic tangent function.

5. A non-intrusive power load identification method based on an intelligent fusion terminal according to claim 1 or 3, characterized in that, The process of performing multiple rounds of encryption perturbation on the compressed data, determined by the number of iterations, includes: Generate a chaotic sequence based on the shared key; Initialize the compressed data to be encrypted; For each round from the first iteration to the specified number of iterations, the values ​​in the chaotic sequence are used to perform an XOR operation on the compressed data encrypted in the previous round to obtain the encryption result for the current round.

6. The non-intrusive power load identification method based on an intelligent fusion terminal according to claim 1, characterized in that, The compression algorithm is the LZ77 compression algorithm, and the compressed data is a triple containing the offset and the matching length.

7. The non-intrusive power load identification method based on an intelligent fusion terminal according to claim 1, characterized in that, The method further includes: Receive the dynamically encrypted compressed data stream in the cloud; The iteration count is synchronously obtained based on the dynamically encrypted compressed data stream; Perform the inverse de-perturbation operation for the specified number of iterations to recover the compressed data and decompress the compressed data.

8. A non-intrusive power load identification system based on an intelligent fusion terminal, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a non-intrusive power load identification method based on a smart fusion terminal according to any one of claims 1-7.

Images