Air conditioner load monitoring and anomaly detection method and system

Abstract

An air conditioner load monitoring and anomaly detection method and system. The method comprises: acquiring user power consumption data and meteorological data, the user power consumption data and the meteorological data comprising the total power of an electricity meter, the power of an air conditioner, and the outdoor temperature, preprocessing the data, and computing electrical characteristics of the user power consumption data; on the basis of the electrical characteristics, learning by using a two-tower neural network so as to obtain operating rules of the air conditioner, and extracting an air conditioner operation curve; on the basis of the air conditioner operation curve, computing electrical characteristics and energy consumption characteristics of the air conditioner, and performing cluster analysis by using a K-means clustering algorithm; and, on the basis of clustering results, comparing the clustering results with clustering results of collected electrical data of the air conditioner under normal and abnormal operating conditions to determine whether the air conditioner is in an abnormal state, so as to implement detection of anomalies in the operation of the air conditioner, thereby improving the precision of load decomposition of the air conditioner, reducing application costs, providing a decision-making reference for equipment maintenance personnel, and prolonging the service life of the equipment.

Description

Air conditioning load monitoring and abnormality detection method and system Technical Field

[0001] The present invention relates to the technical field of air conditioning load monitoring and anomaly detection, and in particular to an air conditioning load monitoring and anomaly detection method and system. Background Art

[0002] Air conditioners are widely used in my country and consume a high proportion of energy. Load monitoring and anomaly detection for air conditioners play an important role in demand response and identifying energy waste, appliance repair or replacement needs. Non-invasive load monitoring can capture patterns from users' historical electricity usage data, accurately identify air conditioner loads, and evaluate the air conditioner's response capabilities as a basis for demand response decisions. Anomaly detection can improve equipment maintenance and extend its service life by promptly identifying faults in electrical components of user equipment, thereby achieving the goal of energy conservation and emission reduction.

[0003] Currently, air conditioning load monitoring primarily relies on non-invasive load identification methods. These methods pre-extract operational features or employ neural network models to capture patterns in timing signals to identify air conditioner starts and stops, or decompose operating power curves. However, these methods lack the ability to integrate the electrical characteristics of air conditioner operation with the patterns in timing signals, leading to increased accuracy in load decomposition. Research on air conditioner anomaly detection currently primarily relies on measuring the operating status of individual devices to detect faults. Studies combining NILM with anomaly detection are limited. While some literature has examined the power consumption patterns of individual devices to detect faults, this approach suffers from the high installation cost associated with a large number of devices. Summary of the Invention

[0004] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides an air conditioning load monitoring and anomaly detection method and system, which can solve the problems mentioned in the background technology.

[0007] To solve the above technical problems, the present invention provides the following technical solution: an air conditioning load monitoring and anomaly detection method, comprising:

[0008] Acquiring user electricity usage data and meteorological data, including the total power of the electricity meter, air conditioning power, and outdoor temperature, and preprocessing the data to calculate electrical characteristics of the user electricity usage data;

[0009] Based on the electrical characteristics, a dual-tower neural network is used to learn the operating rules of the air conditioner and extract the air conditioner operating curve;

[0010] According to the air conditioner operation curve, the electrical characteristics and energy consumption characteristics of the air conditioner are calculated, and the K-means clustering algorithm is used for cluster analysis;

[0011] The clustering result is compared with the collected clustering results of the air conditioner electrical data under normal and abnormal operating conditions to determine whether the air conditioner is in an abnormal state, thereby realizing abnormal detection of the air conditioner operation.

[0012] As a preferred embodiment of the air conditioning load monitoring and anomaly detection method of the present invention, the calculation of the electrical characteristics and energy consumption characteristics of the air conditioner based on the air conditioner operation curve and the cluster analysis using the K-means clustering algorithm include:

[0013] In the first stage, the characteristics are calculated based on the collected electrical data of the air conditioner under normal and abnormal operation;

[0014] The collected data is segmented by time step T2, and the energy consumption characteristics of each time period are calculated. The energy consumption characteristics are calculated by counting the duration of the air conditioner in the i-th opening state. Combined with the power during this duration Get its power consumption

[0015] Similarly, for its standby state, the duration and power The tuple consisting of energy consumption and its duration is used as the energy consumption feature Then calculate the electrical characteristics for each duration Including active power mean, maximum, minimum, variance, peak-to-valley ratio, load rate, minimum load rate, power distribution, and peak value;

[0016] The K-means clustering method is used to cluster the energy consumption characteristics and electrical characteristics respectively, and the cluster centers under normal and abnormal operation of the air conditioner are obtained as the energy consumption and electrical comparison values ​​of the normal and abnormal operation of the air conditioner;

[0017] In the second stage, based on the extracted air conditioner power curve, the electrical characteristics and energy consumption characteristics are calculated, and K-means clustering is performed to obtain the corresponding cluster centers as the basis for air conditioner anomaly detection.

[0018] As a preferred embodiment of the air conditioning load monitoring and anomaly detection method of the present invention, the clustering result is compared with the clustering result of the air conditioning electrical data collected under normal and abnormal operating conditions to determine whether the air conditioning is in an abnormal state, thereby realizing abnormality detection of the air conditioning operation, including:

[0019] Calculate the distance between the cluster center and the air conditioning electrical data cluster center under normal and abnormal operation conditions, which is defined as: ij =||C i -μ j ||2,i=1,2,...,k1,j=1,2,...,k2 k ij =||C i -ξ j ||2,i=1,2,...,k1,j=1,2,...,k2

[0020] Among them, C i is the i-th cluster center of the air conditioning curve to be detected, μ j is the jth cluster center under normal operation of the collected air conditioning curves, ξ j is the jth cluster center of the collected air conditioning curve under abnormal operation, l ij 、k ij C i With μ j ,ξ j The L2 norm between ;

[0021] Set the threshold ε1, if l ij ≤ε1, it means the air conditioner is in normal operation, k ij ≤ε1, the air conditioner is judged to be in an abnormal operating state, and corresponds to the specific abnormal operating state of the cluster center.

[0022] As a preferred solution of the air conditioning load monitoring and anomaly detection method described in the present invention, the method of using a dual-tower neural network to learn the operating rules of the air conditioner includes: one branch uses a long short-term memory recursive neural network to accept the input of a timing signal, including total power data and outdoor temperature, and the other branch uses a fully connected network to accept the input of electrical characteristics. The two branches of the model are spliced ​​and output, and the air conditioning load signal with the same length as the input aliasing signal can be decomposed.

[0023] As a preferred solution of the air conditioning load monitoring and anomaly detection method of the present invention, the electrical characteristics of the user's electricity consumption data are calculated as follows:

[0024] The acquired data is segmented by time step T1, and the electrical characteristics of the user bus power data in each period are calculated, including the active power mean, maximum, minimum, variance, duty cycle, peak-to-valley difference rate, load rate, minimum load rate, power distribution, and peak value;

[0025] The power distribution calculation formula is as follows: P unb =|P med -P mean |

[0026] Where N represents the total number of samples in the power sequence, Indicates power is less than 0.75P min +0.25P max The number of samples, Indicates that the power sequence is greater than 0.75P max +0.25P min The number of samples, P unb Describe the imbalance of the sample, P med represents the median of the power series, P mean represents the average value of the power series;

[0027] The spike quantity is obtained by performing wavelet decomposition on the power sequence, averaging the high-frequency detail coefficients of the first layer of the decomposition, comparing the detail coefficients with a multiple of the average value a, and calculating the number of detail coefficients greater than the value as a quantitative description value of the spike.

[0028] As a preferred solution of the air conditioning load monitoring and anomaly detection method described in the present invention, the acquisition of user electricity consumption data and meteorological data includes: data used for the training phase and the testing phase, the data in the training phase are the user's total meter power, air conditioning power and outdoor temperature during normal and abnormal operation of the air conditioner, and the data in the testing phase are the user's total power and outdoor temperature in actual scenarios.

[0029] As a preferred solution of the air conditioning load monitoring and anomaly detection method described in the present invention, the preprocessing of the data includes: deleting abnormal values ​​and missing values ​​in the data, aligning the electrical data with the outdoor temperature data along the time axis, and using forward filling for missing data in the alignment process.

[0030] An air conditioning load monitoring and anomaly detection system, characterized by comprising: a data acquisition and calculation module, a curve acquisition module, a cluster analysis module and an anomaly detection module,

[0031] a data acquisition and calculation module, which is used to acquire user electricity usage data and meteorological data, including the total power of the electricity meter, the air conditioning power, and the outdoor temperature, and pre-process the data to calculate the electrical characteristics of the user electricity usage data;

[0032] A curve acquisition module, the curve acquisition module is used to obtain the operating rules of the air conditioner based on the electrical characteristics using a dual-tower neural network learning method and extract the air conditioner operating curve;

[0033] A cluster analysis module is used to calculate the electrical characteristics and energy consumption characteristics of the air conditioner according to the air conditioner operation curve, and perform cluster analysis using the K-means clustering algorithm;

[0034] The abnormality detection module is used to compare the clustering results with the collected air-conditioning electrical data clustering results under normal and abnormal operating conditions according to the clustering results, determine whether the air-conditioning is in an abnormal state, and realize abnormality detection of the air-conditioning operation.

[0035] A computer device includes a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the above method when executing the computer program.

[0036] A computer-readable storage medium stores a computer program thereon, wherein the computer program implements the steps of the method described above when executed by a processor.

[0037] Beneficial Effects of the Invention: The present invention proposes a method and system for monitoring and detecting air conditioner load anomalies. The method obtains user electricity consumption data and meteorological data, including the total power of the meter, air conditioner power, and outdoor temperature, preprocesses the data, and calculates the electrical characteristics of the user electricity consumption data. Based on the electrical characteristics, a dual-tower neural network is used to learn the operating patterns of the air conditioner and extract the air conditioner operating curve. Based on the air conditioner operating curve, the air conditioner's electrical characteristics and energy consumption characteristics are calculated and clustered using a K-means clustering algorithm. The clustering results are compared with the clustering results of collected air conditioner electrical data under normal and abnormal operating conditions to determine whether the air conditioner is in an abnormal state, thereby detecting anomalies in air conditioner operation. To address the problem of insufficient air conditioner load decomposition accuracy, a dual-tower neural network model is proposed. This model combines the time series signal patterns of the air conditioner with its electrical characteristics. A long short-term memory recurrent neural network and a fully connected network are used to mine the patterns of the time series signals and electrical characteristics, respectively. Finally, the outputs of the two branches are concatenated. This method provides an air conditioner monitoring method with improved load decomposition accuracy, providing a foundation for air conditioners to participate in demand response, explore their response potential, or further perform anomaly detection. To address the high cost and limited scale of single-device anomaly detection, a method combining non-invasive load identification with anomaly detection was proposed. The results of load decomposition were applied to anomaly detection, enabling anomaly detection for air conditioners, reducing application costs, providing decision-making support for equipment maintenance personnel, and extending equipment life. Furthermore, to address the low accuracy of anomaly detection, a method was proposed to calculate the corresponding features of air conditioners, perform cluster analysis, and compare the clustering results. This provides a method for anomaly detection, improves its accuracy, and enhances its practicality. 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 describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0039] FIG1 is a flowchart of an overall method of an air conditioning load monitoring and anomaly detection method and system provided by one embodiment of the present invention;

[0040] FIG2 is a flow chart of an air conditioning load monitoring and anomaly detection method and system using a dual-tower neural network according to an embodiment of the present invention;

[0041] FIG3 is a schematic diagram of the specific structure of a dual-tower neural network of an air conditioning load monitoring and anomaly detection method and system provided by one embodiment of the present invention;

[0042] FIG4 is an internal structural diagram of a computer device of an air conditioning load monitoring and anomaly detection method and system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0044] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0045] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0046] The present invention is described in detail with reference to schematic diagrams. For ease of illustration, cross-sectional views of device structures may be partially enlarged and not to scale when describing embodiments of the present invention. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0047] In the description of the present invention, it should be noted that the terms "upper, lower, inner, and outer" and other references to orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0048] In this disclosure, unless otherwise specified or limited, the terms "mounted," "connected," and "connected" should be interpreted broadly. For example, they may refer to fixed, removable, or integral connections. They may also refer to mechanical, electrical, or direct connections, indirect connections through an intermediary, or internal communication between two components. Those skilled in the art will understand the specific meanings of these terms in this disclosure.

[0049] Example 1

[0050] 1-4 , which illustrate a first embodiment of the present invention, provide an air conditioning load monitoring and anomaly detection method and system, including:

[0051] Currently, air conditioning load monitoring primarily relies on non-invasive load identification methods. These methods pre-extract operational characteristics or employ neural network models to capture patterns in timing signals to identify air conditioner starts and stops or decompose operating power curves. However, this approach lacks the combined utilization of operational characteristics and timing signal patterns. Air conditioning anomaly detection typically relies on single-device detection, lacking integration with non-invasive load identification. This paper proposes a method for air conditioning load monitoring and anomaly detection based on a dual-tower neural network. First, a method for calculating the electrical characteristics of the air conditioner is designed. The dual-tower neural network leverages the patterns of timing signals and their electrical characteristics. One branch receives input from the air conditioner's timing signals using a long-short-term memory recurrent network, while the other branch receives input from its electrical characteristics using a fully connected network. Finally, the outputs of the two branches are combined to decompose the air conditioning load. Based on the decomposed air conditioning load, energy consumption and electrical characteristics are calculated. K-means clustering analysis is then used to compare the clustering results with those of pre-collected data. Thresholds are set to determine if an anomaly is present and what the anomaly is. The present invention has a good effect on air conditioning monitoring and realizes its anomaly detection, thereby improving the adaptability and practicality of air conditioning monitoring and anomaly detection. The specific steps of the method are as follows:

[0052] Obtain user electricity consumption data and meteorological data, including the total power of the meter, air conditioning power, and outdoor temperature, and pre-process the data to calculate the electrical characteristics of the user electricity consumption data;

[0053] Among them, obtaining user electricity consumption data and meteorological data includes: data used for training phase and testing phase. The data in the training phase is the user's total meter power, air conditioning power and outdoor temperature when the air conditioner is operating normally and abnormally. The data in the testing phase is the user's total power and outdoor temperature in actual scenarios.

[0054] Furthermore, data preprocessing includes: deleting outliers and missing values ​​in the data, aligning the electrical data with the outdoor temperature data along the time axis, and using forward filling for missing data in the alignment process.

[0055] Furthermore, the padded data is standardized, and the data in the training phase needs to be divided into training set and test set;

[0056] In the embodiment of the present application, the ratio of the training set to the total data is 0.85.

[0057] Furthermore, the electrical characteristics of the user's electricity consumption data are calculated including:

[0058] The acquired data is segmented by time step T1, and the electrical characteristics of the user bus power data in each period are calculated, including the active power mean, maximum, minimum, variance, duty cycle, peak-to-valley difference rate, load rate, minimum load rate, power distribution, and peak value;

[0059] The mean active power is calculated as the mean of a power sequence with n samples:

[0060] Where, P i is the value of each point in the power series.

[0061] Duty cycle calculation: The ratio of power sequences greater than a certain threshold to all samples:

[0062] Where N on is the number of samples whose power sequence is greater than the corresponding threshold, and N is the total number of samples.

[0063] The peak-to-valley difference rate is the ratio of the difference between the maximum and minimum values ​​of the power sequence to the maximum value, that is,

[0064] Where P max is the maximum value in the power sequence, P min Refers to its minimum value.

[0065] The load rate is the ratio of the average value to the maximum value of the power sequence, that is,

[0066] Where P mean Refers to the average value of the power series.

[0067] The minimum load rate is the ratio of the minimum value to the maximum value of the power sequence, that is,

[0068] The power distribution calculation formula is as follows: P unb =|P med -Pmean |

[0069] Where N represents the total number of samples in the power sequence, Indicates power is less than 0.75P min +0.25P max The number of samples, Indicates that the power sequence is greater than 0.75P max +0.25P min The number of samples, P unb Describe the imbalance of the sample, P med represents the median of the power series, P mean represents the average value of the power series;

[0070] Furthermore, the spike quantity is obtained by performing wavelet decomposition on the power sequence, averaging the high-frequency detail coefficients of the first layer of the decomposition, comparing the detail coefficients with a multiple of their average value, and finding the number of detail coefficients greater than this value as a quantitative description value of the spike.

[0071] In the embodiment of the present application, a=20, and the number of values ​​greater than this value is calculated as the quantitative description value of the peak.

[0072] Based on the electrical characteristics, a dual-tower neural network is used to learn the operating rules of the air conditioner and extract the air conditioner operating curve;

[0073] Furthermore, a dual-branch neural network is used to learn the operating patterns of air conditioners. One branch uses a long-short-term memory recurrent neural network to accept time-series inputs, including total power data and outdoor temperature. The other branch uses a fully connected network to accept electrical characteristics. By concatenating these two branches and outputting them, the air conditioner load signal, which is equal in length to the input aliased signal, can be decomposed. The detailed structure is shown in Figure 3.

[0074] It should be noted that the LSTM recursion consists of three parts: the forget gate, the memory gate, and the output gate. "σ" and "tanh" represent the sigmoid and tanh activation functions, respectively. The "×" and "+" signs represent the Hadamard product and addition operations of matrices, respectively.

[0075] Furthermore, the forget gate is composed of f t =σ(W f ·[h t-1 ,x t ]+b f ) to get the output f t , where x t is the current input, including user bus power, outdoor temperature data, h t-1 is the output of the previous stage, [h t-1 ,x t] means connecting the two matrices to retain specific long-term memory information.

[0076] Furthermore, the memory gate is composed of i t =σ(W i ·[h t-1 ,x t ]+b i ) is achieved through i t control Some of the information is remembered.

[0077] Furthermore, the outputs of the forget gate and the memory gate control the previous long-term memory C t-1 Which information to retain and add new information to get the updated long-term memory C t ,

[0078] Furthermore, the current output h t for o t =σ(W o [h t-1 ,x t ]+b o ), h t =o t *tanh(C t ) and serves as the input to the next stage. Finally, h t Output through the linear output layer.

[0079] Furthermore, the number of neuron nodes and network layers of the long short-term memory recurrent neural network affects the recognition effect of the air conditioner. In this paper, the dimension of the input data is used as the number of input layer nodes. The input data includes user total power data and outdoor temperature data, and the dimension is 2. The number of hidden layer neurons z1 is selected as 15 considering the training speed and recognition accuracy. The dimension of the output data z2 is not compressed for the time being because it needs to be spliced ​​with the output of the fully connected neural network. It is still selected as 15. At the same time, in order to improve the recognition accuracy, the number of network layers z3 is selected as 2.

[0080] Furthermore, the input layer of the fully connected neural network takes the calculated electrical characteristic data as input. The dimension of the electrical characteristics calculated in this paper is 12, the number of hidden layers z4 is selected as 2 layers, the activation function is the relu function, and the specific expression is f(x) = max(0,x). Finally, the output layer obtains the electrical characteristic data output with a dimension z5 of 60.

[0081] Furthermore, the merged output is finally performed. The long short-term memory recursive neural network processes the input sequence, and the output data is still a segment of sequence data with a data dimension z2=15. The fully connected neural network part obtains the corresponding electrical characteristics of the sequence based on the sequence data, so its output sequence length is compressed to 1. In order to match the output with the output of the long short-term memory recursive neural network, the output sequence length of the fully connected neural network is expanded. Since its output represents the electrical characteristics of the entire input sequence, that is, each sample point can be considered to contain this feature, it is copied so that its sequence length is consistent with the output of the long short-term memory recursive. Finally, the output data is spliced ​​in sequence, so that the dimension of the output data is further increased.

[0082] Furthermore, after the output is merged, the data is sent to the fully connected layer. Here, it passes through two fully connected layers, the activation function is the ReLU function, and finally it is output by the linear output layer to obtain sequence data with a dimension of 1, which is the identified air conditioning power value.

[0083] Furthermore, the loss function is used to measure the difference between the predicted value of the neural network model and the label value. It is used for backpropagation of the model to update the parameters and plays an important role in correcting the weights in the network layer, making the output value of the model closer to the label value and achieving the purpose of learning. The MSE is selected and its expression is:

[0084] Among them, n is the number of labels, x i is the model prediction value, y i is the tag value.

[0085] It should be noted that the learning rate is 0.001, and the Adam method is used to adjust the weights. This algorithm uses the first and second moment estimation calculations of the gradient as the basis to iteratively correct the model parameters. It has the advantages of high computational efficiency and applicability to large-scale data and parameters.

[0086] Furthermore, based on the air conditioner operation curve, the electrical characteristics and energy consumption characteristics of the air conditioner are calculated, and the K-means clustering algorithm is used for cluster analysis;

[0087] Based on the air conditioner operation curve, the electrical characteristics and energy consumption characteristics of the air conditioner are calculated, and the K-means clustering algorithm is used for cluster analysis, including:

[0088] Furthermore, in the first stage, features are calculated from the collected electrical data of the air conditioners in normal and abnormal operation;

[0089] The collected data is segmented by time step T2, and the energy consumption characteristics of each time period are calculated. The energy consumption characteristics are calculated by counting the duration of the air conditioner in the i-th opening state. Combined with the power during this duration Get its power consumption

[0090] Similarly, for its standby state, the duration and power The tuple consisting of energy consumption and its duration is used as the energy consumption feature Then calculate the electrical characteristics for each duration Including active power mean, maximum, minimum, variance, peak-to-valley ratio, load rate, minimum load rate, power distribution, and peak value;

[0091] Furthermore, the K-means clustering method is used to cluster the energy consumption characteristics and electrical characteristics respectively, and the cluster centers under normal and abnormal operation of the air conditioner are obtained as the energy consumption and electrical comparison values ​​of the normal and abnormal operation of the air conditioner;

[0092] In the embodiment of the present application, considering the diversity of air-conditioning operation conditions, k=6 may be taken.

[0093] In the second stage, based on the extracted air conditioner power curve, the electrical characteristics and energy consumption characteristics are calculated, and K-means clustering is performed to obtain the corresponding cluster centers as the basis for air conditioner anomaly detection.

[0094] Furthermore, based on the clustering results, the collected air conditioner electrical data clustering results under normal and abnormal operating conditions are compared to determine whether the air conditioner is in an abnormal state, thereby realizing abnormal detection of air conditioner operation.

[0095] The distance between the calculated cluster center and the air conditioning electrical data cluster center under normal and abnormal operation conditions is defined as: ij =||C i -μ j ||2,i=1,2,...,k1,j=1,2,...,k2 k ij =||C i -ξ j ||2,i=1,2,...,k1,j=1,2,...,k2

[0096] Among them, C i is the i-th cluster center of the air conditioning curve to be detected, μ j is the jth cluster center under normal operation of the collected air conditioning curves, ξ j is the jth cluster center of the collected air conditioning curve under abnormal operation, l ij 、k ij C i With μj ,ξ j The L2 norm between ;

[0097] Set the threshold ε1, if l ij ≤ε1, it means the air conditioner is in normal operation, k ij ≤ε1, the air conditioner is judged to be in an abnormal operating state, and corresponds to the specific abnormal operating state of the cluster center.

[0098] In a preferred embodiment, an air conditioning load monitoring and anomaly detection system includes: a data acquisition and calculation module, a curve acquisition module, a cluster analysis module, and an anomaly detection module.

[0099] The data acquisition and calculation module is used to obtain user electricity consumption data and meteorological data, including the total power of the meter, air conditioning power, and outdoor temperature, and pre-process the data to calculate the electrical characteristics of the user electricity consumption data;

[0100] The curve acquisition module is used to learn the operating rules of the air conditioner based on the electrical characteristics using a dual-tower neural network and extract the air conditioner operating curve;

[0101] Cluster analysis module, which is used to calculate the electrical characteristics and energy consumption characteristics of the air conditioner based on the air conditioner operation curve, and uses the K-means clustering algorithm to perform cluster analysis;

[0102] The anomaly detection module is used to compare the clustering results with the collected air-conditioning electrical data clustering results under normal and abnormal operating conditions, determine whether the air-conditioning is in an abnormal state, and realize the anomaly detection of the air-conditioning operation.

[0103] The above-mentioned unit modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of the above-mentioned modules.

[0104] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG4 . The computer device includes a processor, memory, a communication interface, a display screen, and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal via wired or wireless communication. The wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. When executed by the processor, the computer program implements a method for air conditioning load monitoring and anomaly detection. The display screen of the computer device may be a liquid crystal display or an electronic ink display. The input device of the computer device may be a touch screen covering the display screen, or may be buttons, a trackball, or a touchpad provided on the computer device housing, or may be an external keyboard, touchpad, or mouse.

[0105] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0106] Obtain user electricity consumption data and meteorological data, including the total power of the meter, air conditioning power, and outdoor temperature, and pre-process the data to calculate the electrical characteristics of the user electricity consumption data;

[0107] Based on the electrical characteristics, a dual-tower neural network is used to learn the operating rules of the air conditioner and extract the air conditioner operating curve;

[0108] According to the air conditioner operation curve, the electrical characteristics and energy consumption characteristics of the air conditioner are calculated, and the K-means clustering algorithm is used for cluster analysis;

[0109] Based on the clustering results, the data are compared with the clustering results of the air conditioner electrical data under normal and abnormal operating conditions to determine whether the air conditioner is in an abnormal state, thereby realizing abnormal detection of the air conditioner operation.

[0110] Example 2

[0111] 2-3 , which illustrate an embodiment of the present invention, provide an air conditioning load monitoring and anomaly detection method and system. To verify the beneficial effects of the present invention, a comparative experiment is conducted to provide scientific evidence.

[0112] Load decomposition scheme based on long short-term memory recursive neural network: The traditional load decomposition algorithm based on long short-term memory recursive algorithm only learns relevant features through user total power data, so the decomposition accuracy of air conditioning load is relatively low. In order to verify that this method has higher air conditioning load decomposition accuracy than the traditional method, this embodiment will use the load decomposition algorithm based on long short-term memory recursive algorithm and the method of the present invention to compare the air conditioning operation curve obtained by user bus power decomposition with its actual operation curve (Coefficient of determination, R 2 ), mean squared error (MSE), and root mean square error (RMSE) were compared.

[0113] Test environment: A case study was conducted using one year of active power waveform data for air conditioners and buses collected from user homes in a foreign community. Data from the first half of May and the first half of June were selected for identification using the air conditioner load decomposition model based on a long-short-term memory recurrent neural network and the proposed method. The training data accounted for 0.85% of the total data, and the learning rate was 0.001. The R², MSE, and RMSE of the air conditioner load curve identified by the model and the actual operating curve were calculated using the following formulas:

[0114] Where, is the i-th predicted value, y i is the i-th true value, Represents the mean of the true values.

[0115] The results are compared as shown in the following table:

[0116] Table 1: Comparison of experimental results:

[0117] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0118] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiment of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal translation scripting language JavaScript, etc.

[0119] The present application is described with reference to the flow chart and / or block diagram of the method, device (system), and computer program product according to the embodiment of the present application. It should be understood that each flow process and / or box in the flow chart and / or block diagram and the combination of the flow process and / or box in the flow chart and / or block diagram can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processing machine or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for realizing the function specified in one flow chart flow or multiple flows and / or one box or multiple boxes of the block diagram.

[0120] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0121] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0122] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0123] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A method for monitoring and detecting air conditioning load anomalies, characterized in that: include: Acquire user electricity consumption data and meteorological data, the user electricity consumption data and meteorological data including the total power of the electric meter, the air conditioning power, and the outdoor temperature, and pre-process the data to calculate the electrical characteristics of the user electricity consumption data; According to the electrical characteristics, a dual-tower neural network is used to learn the operation law of the air conditioner and extract the air conditioner operation curve; According to the air conditioner operation curve, the electrical characteristics and energy consumption characteristics of the air conditioner are calculated, and the K-means clustering algorithm is used for cluster analysis; According to the clustering result, the clustering result of the air conditioner electrical data under the normal and abnormal operation states is compared with the collected data to determine whether the air conditioner is in an abnormal state, thereby realizing abnormal detection of the air conditioner operation.

2. The air conditioning load monitoring and abnormality detection method according to claim 1, characterized in that: The step of calculating the electrical characteristics and energy consumption characteristics of the air conditioner according to the air conditioner operation curve and performing cluster analysis using the K-means clustering algorithm includes: In the first stage, the characteristics are calculated by using the collected electrical data of the air conditioner in normal and abnormal operation; The collected data is segmented according to the time step T2, and the energy consumption characteristics of each time period are calculated. The energy consumption characteristics are calculated by counting the duration of the air conditioner being turned on for the i-th time. Combined with the power during this duration Get its power consumption Similarly, for its standby state, the duration And power The tuple consisting of energy consumption and its duration is used as the energy consumption feature Then calculate the electrical characteristics for each duration Including active power mean, maximum, minimum, variance, peak-to-valley ratio, load rate, minimum load rate, power distribution, and peak value; The K-means clustering method is used to cluster the energy consumption characteristics and electrical characteristics respectively, and the cluster centers under normal and abnormal operation of the air conditioner are obtained as the energy consumption and electrical comparison values ​​of the normal and abnormal operation of the air conditioner; In the second stage, the electrical characteristics and energy consumption characteristics are calculated based on the extracted air conditioner power curve, and K-means clustering is also performed to obtain the corresponding cluster center as the basis for air conditioner anomaly detection.

3. The air conditioning load monitoring and abnormality detection method according to claim 2, characterized in that: The clustering result is compared with the collected air conditioner electrical data clustering results under normal and abnormal operation states to determine whether the air conditioner is in an abnormal state, and the abnormality detection of the air conditioner operation includes: Calculate the distance between the cluster center and the air conditioning electrical data cluster center under normal and abnormal operation conditions, which is defined as: l ij =||C i -μ j ||2,i=1,2,...,k1,j=1,2,...,k2 k ij =||C i -ξ j ||2,i=1,2,...,k1,j=1,2,...,k2 Among them, C i is the i-th cluster center of the air conditioning curve to be detected, μ j is the jth cluster center of the collected air conditioning curves under normal operation, ξ j is the jth cluster center of the collected air conditioning curve under abnormal operation, l ij , k ij C i With μ j , j The L2 norm between ; Set the threshold ε1, if l ij ≤ε1, it means the air conditioner is in normal operation, k ij ≤ε1, the air conditioner is judged to be in an abnormal operating state, and corresponds to the specific abnormal operating state of the cluster center.

4. The air conditioning load monitoring and abnormality detection method according to claim 3, characterized in that: The air-conditioning operation rules learned by using a dual-tower neural network include: one branch accepts the input of a timing signal, including total power data and outdoor temperature, with a long short-term memory recursive neural network, and the other branch accepts the input of electrical characteristics with a fully connected network. The two branches of the model are spliced ​​and output, and the air-conditioning load signal with the same length as the input aliasing signal can be decomposed.

5. The air conditioning load monitoring and abnormality detection method according to claim 4, characterized in that: The electrical characteristics of calculating the user's electricity consumption data include: The acquired data is segmented according to the time step T1, and the electrical characteristics of the user bus power data in each period of time are calculated, including the active power mean, maximum value, minimum value, variance, duty cycle, peak-to-valley difference rate, load rate, minimum load rate, power distribution, and peak value; The power distribution calculation formula is as follows: P unb =|P med -P mean | Where N is the total number of samples in the power sequence, Indicates power is less than 0.75P min +0.25P max The number of samples, Indicates that the power sequence is greater than 0.75P max +0.25P min The number of samples, P unb Describe the imbalance of the sample, P med represents the median of the power series, P mean Indicates work The average value of the rate series; The peak value is obtained by performing wavelet decomposition on the power sequence, averaging the high-frequency detail coefficients of the first layer of the decomposition, comparing the detail coefficients with a multiple of the average value a, and calculating the number of the detail coefficients greater than the value as the quantitative description value of the peak.

6. The air conditioning load monitoring and abnormality detection method according to claim 5, characterized in that: The acquisition of user electricity consumption data and meteorological data includes: data used for training phase and testing phase, the data for the training phase is the user's total meter power and air conditioning power and outdoor temperature when the air conditioner is operating normally and abnormally, and the data for the testing phase is the user's total power and outdoor temperature in actual scenarios.

7. The air conditioning load monitoring and abnormality detection method according to claim 6, characterized in that: The preprocessing of the data includes: deleting abnormal values ​​and missing values ​​in the data, aligning the electrical data with the outdoor temperature data along the time axis, and forward filling the missing data in the alignment process.

8. An air conditioning load monitoring and abnormality detection system, characterized in that: include: Data acquisition and calculation module, curve acquisition module, cluster analysis module and anomaly detection module, A data acquisition and calculation module, which is used to acquire user electricity consumption data and meteorological data, including the total power of the electric meter, the air conditioning power, and the outdoor temperature, and pre-process the data to calculate the electrical characteristics of the user electricity consumption data; A curve acquisition module, the curve acquisition module is used to obtain the operating law of the air conditioner by using a double-tower neural network learning according to the electrical characteristics, and extract the air conditioner operating curve; A cluster analysis module, which is used to calculate the electrical characteristics and energy consumption characteristics of the air conditioner according to the air conditioner operation curve, and perform cluster analysis using a K-means clustering algorithm; The abnormality detection module is used to compare the clustering results with the collected air-conditioning electrical data clustering results under normal and abnormal operating conditions according to the clustering results, determine whether the air-conditioning is in an abnormal state, and realize abnormality detection of the air-conditioning operation.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images