A method for monitoring the operating status of household appliances based on an online feature library
Through the online feature library method, the operating status of household appliances is monitored using Pearson similarity coefficient and sliding window function, and the problem of low monitoring accuracy of continuous variable state appliances in the prior art is solved, and higher monitoring accuracy and more accurate load identification are achieved.
Patent Information
- Application Number
- CN202210971530.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-12
- Publication Date
- 2025-06-17
- Estimated Expiration
- 2042-08-12
AI Technical Summary
The existing household appliance operating status monitoring methods have low monitoring accuracy on continuously variable-state electrical appliances (such as variable-frequency air conditioners), resulting in poor load identification effect of the residential bus.
Using the method based on the online feature library, the periodic current waveform similarity function and sliding window function are established through the Pearson similarity coefficient, the array in the online feature library is updated in real time, and the operating status of household appliances is monitored.
It improves the accuracy of operating status monitoring of household appliances, reduces the missed detection rate, and enhances the accuracy of load identification.
Smart Images
Figure CN115389838B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to non-intrusive load monitoring technology, and particularly to a method for monitoring the operating state of household appliances based on an online feature library. Background Art
[0002] With the development of social economy, residential electricity consumption is developing in the direction of a large variety of electrical appliances and complex electricity consumption behaviors, which makes it difficult for power grid companies to obtain residential electricity consumption information. Therefore, it is necessary to study a load monitoring system suitable for the residential side. The purpose of load monitoring is to online monitor the load type, operating state, and energy consumption by real-time collecting the electrical parameters of the power load.
[0003] The power load monitoring at the residential side is mainly divided into two types: intrusive and non-intrusive. Compared with the intrusive load monitoring method, which has high hardware costs and inconvenient maintenance, the non-intrusive load monitoring method only installs load monitoring equipment at the residential electricity inlet, with less impact on residents' daily lives. Currently, non-intrusive load monitoring technology has become one of the research hotspots among many scholars at home and abroad.
[0004] As an important link in the non-intrusive load monitoring process, the monitoring accuracy of the load state will directly affect the subsequent load decomposition and identification accuracy. The existing methods for monitoring the operating state of household appliances have low accuracy in monitoring the state of continuously variable state appliances (such as variable frequency air conditioners), resulting in poor load identification effects at the residential side bus, and there are significant differences between the identified results and the actual appliance types. Summary of the Invention
[0005] The purpose of the present invention is to provide a method for monitoring the operating state of household appliances based on an online feature library.
[0006] The technical solution for realizing the present invention is: A method for monitoring the operating state of household appliances based on an online feature library, comprising the following steps:
[0007] Step 1, using the Pearson similarity coefficient, establish a periodic current waveform similarity function and a sliding window function;
[0008] Step 2, based on the periodic current array in the process of monitoring the operating state of household appliances, establish an online feature library, and update the array in the online feature library in real time according to the needs of state changes;
[0009] Step 3, based on the periodic current waveform similarity function and the online feature library, monitor the operating state of household appliances.
[0010] Step 1, using the Pearson similarity coefficient, establish a periodic current waveform similarity function and a sliding window function for judging the operating state of household appliances. The specific method is as follows:
[0011] Step 1.1, define the Pearson similarity coefficient function between two periodic current arrays, denoted as the periodic current waveform similarity function, as shown in Equation (1):
[0012]
[0013] In the formula, x is the periodic current array of the device to be sampled at the current moment, and x i is the previous consecutive periodic current array adjacent to the periodic current array x in time;
[0014] Step 1.2, based on the periodic current waveform similarity function, establish a sliding window function, as shown in Equation (2);
[0015]
[0016] In the formula, {x 1, ...,x m} is the set of current arrays in the sliding window for judging the state of the periodic current array x at the current moment, m is the dynamic index of the number of current arrays in the sliding window, α is the sliding window attenuation coefficient, and m - i is the exponent of the attenuation coefficient between the periodic current array x at the current moment and the current array x i in the sliding window; F(x) is the cumulative sum of the products of the similarity coefficients and attenuation coefficients of the periodic current array x at the current moment and each consecutive periodic current array in the sliding window, and the effective value range is between 0 and 1. When F(x) is greater than the threshold, it is determined that the operating state of the current household appliance is continuous, otherwise it is determined that the operating state has changed, as shown in Equation (3):
[0017]
[0018] Step Two, based on the typical periodic current arrays in the process of monitoring the operating state of household appliances, establish an online feature library and update the arrays in the online feature library in real time according to the needs of state changes. The specific method is as follows:
[0019] Step 2.1, collect the high-frequency current data of household appliances in the resident's home, encapsulate the first periodic current array collected as (x1, y1), and add (x1, y1) to the online feature library for initialization, where y1 is the feature attenuation coefficient corresponding to x1, and the initial value is set to 1;
[0020] Step 2.2, perform update processing on the periodic current arrays in the online feature library that are similar to the periodic current array x at the current moment. The update function is as shown in Equation (4), and the remaining current arrays in the feature library need to be updated for the attenuation coefficient, as shown in Equation (5);
[0021]
[0022]
[0023] Among them, k is the index value of the current array in the online feature library, and n is the number of periodic current arrays in the online feature library at the current moment;
[0024] Step 2.3, if the sliding window function values of the periodic current array x at the current moment and the periodic current arrays in the feature library are both less than the similarity threshold, and the values of each attenuation coefficient are greater than the set redundancy threshold of the online feature library, then add (x, y) to the online feature library, and the number of periodic current arrays in the online feature library becomes (n + 1), where y is the feature attenuation coefficient corresponding to x, otherwise the online feature library is not updated.
[0025] Step three, based on the periodic current waveform similarity function and the online feature library, monitor the operating state of household appliances. The specific method is as follows:
[0026] (1) When (x, y) is similar to a certain periodic current array (x i , y i ) in the online feature library, and the attenuation coefficient values of all periodic current arrays in the online feature library are greater than the set redundancy threshold of the online feature library, it is determined that the operating state of the current household appliance is continuous;
[0027] (2) When (x, y) is not similar to all periodic current arrays (x i , y i ) in the online feature library, add (x, y) to the online feature library. If the attenuation coefficient value of a periodic current array in the online feature library is less than the set redundancy threshold of the online feature library, it is determined that the operating state of the current household appliance is interrupted, otherwise it is determined to be continuous;
[0028] (3) When the attenuation coefficient y i , y i of a certain periodic current array (x i ) in the online feature library is less than the set redundancy threshold of the online feature library, it indicates that the set of feature arrays in the online feature library {(x1, y1), …, (x n , y n )} is no longer suitable for monitoring the operating state of the current household appliance, that is, the continuous current waveform or pattern during the operation of the household appliance has changed. At this time, without considering the similarity between (x, y) and the periodic current arrays in the online feature library, it is determined that the operating state of the household appliance changes at the current moment, and the online feature library is re-initialized at the same time.
[0029] A monitoring system for the operating state of household appliances based on an online feature library realizes the monitoring of the operating state of household appliances based on the online feature library through the described monitoring method for the operating state of household appliances based on the online feature library.
[0030] A computer device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the method for monitoring the operating state of a household appliance based on an online feature library is implemented.
[0031] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the method for monitoring the operating state of a household appliance based on an online feature library is implemented.
[0032] Compared with the prior art, the significant advantage of the present invention is that by calculating the Pearson similarity coefficient between the periodically sampled current array and the periodically sampled current array in the online feature library, the problem of monitoring the operating state of household appliances with slow-changing operating characteristics is effectively solved, the monitoring accuracy is improved, and the missed detection rate is reduced. Description of the Drawings
[0033] Figure 1 It is a schematic flowchart of a method for monitoring the operating state of a household appliance based on an online feature library according to the present invention;
[0034] Figure 2 It is a measured current waveform diagram of three types of household appliance devices;
[0035] Figure 3 It is a schematic diagram of the sliding window function calculation of an electric kettle;
[0036] Figure 4 It is a monitoring result diagram of the operating state of an electric kettle;
[0037] Figure 5 It is a monitoring result diagram of the operating state of a hair dryer;
[0038] Figure 6 It is a monitoring result diagram of the operating state of a variable-frequency air conditioner. Detailed Embodiments
[0039] In order to make the objectives, technical solutions, and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0040] As Figure 1 shown, a method for monitoring the operating state of a household appliance based on an online feature library. According to the different operating states of household appliances, household appliances can be divided into start-stop two-state devices, finite multi-state devices, and continuously variable state devices. Aiming at the problems of missed detection and false detection in the monitoring of the operating state of continuously variable state household appliances, the monitoring of the operating state of household appliances includes the following steps:
[0041] Step 1: Using the Pearson similarity coefficient, design a periodic current waveform similarity function and a sliding window function for judging the continuity of the operating state of household appliances;
[0042] First, define the Pearson similarity coefficient function between two periodic current arrays, denoted as the periodic current waveform similarity function, as shown in Equation (1):
[0043]
[0044] In the formula, x is the periodic current array of the device being sampled at the current moment, and x i is the previous continuous periodic current array adjacent to the x current array in time. The periodic current waveform similarity function can reflect the similarity of waveforms and amplitudes between two periodic current arrays. The discrimination criteria for similarity values are shown in Table 1.
[0045] Table 1 Similarity discrimination criteria
[0046] <![CDATA[f(x,x i )]]> Similarity 1.0～0.8 Very strong similarity 0.8～0.6 Strong similarity 0.6～0.4 Moderate similarity 0.4～0.2 Weak similarity 0.2～0.0 Very weak similarity or dissimilarity
[0047] Then, based on the periodic current waveform similarity function, establish a sliding window, and monitor the operating state of the current household appliance by calculating the similarity of the current arrays within the sliding window. The sliding window function is shown in Equation (2).
[0048]
[0049] In the formula, {x 1, ...,x m} can be regarded as the set of current arrays in the sliding window for judging the state of the current current array x. m is the dynamic index value of the number of current arrays in the sliding window (a natural number starting from 1, which is also the current sliding window size). If the current arrays within a relatively long time belong to the same operating state, the value of m will continue to increase. Considering the online monitoring speed of computer performance and load status, take m = 10. α is the sliding window attenuation coefficient, and m - i is the exponent of the attenuation coefficient between the current array x and the current array x i in the sliding window (0 < α < 1, let's assume α = 0.9).
[0050] Finally, F(x) is the cumulative sum of the products of the similarity coefficients and attenuation coefficients of the current array x at the current moment and each continuous periodic current array in the sliding window. By calculating the value of F(x), the operating state of the household appliance is monitored. The effective value range of F(x) is between 0 and 1. As shown in Equation (3), 0.9 * α is the threshold for judging whether F(x) is continuous. 0.9 is the extremely strong similarity value in Table 1, and α is the sliding window attenuation coefficient. When F(x) is greater than the threshold, it means that the current periodic current array x is similar to the periodic current arrays in the sliding window. At this time, it is determined that the household appliance is in a continuous operating state. Otherwise, it indicates that the operating state of the household appliance has changed.
[0051]
[0052] Step 2: Establish an online feature library, which contains typical periodic current arrays during the operation state monitoring of household appliances, and update the arrays in the online feature library in real time according to the needs of state changes;
[0053] Due to different equipment types, the continuous periodic current waveforms of various household appliances show different regularities during operation. A data structure (x, y) is proposed, where x is the periodic current array and y is the characteristic attenuation coefficient corresponding to x. For the regular fluctuations of the continuous periodic current waveforms of continuously variable state devices, the sliding window in Step 1 is improved to an online feature library, so as to realize the monitoring of the operation state of this type of household appliances.
[0054] First step: Collect the high-frequency current data of the appliances in the residents' homes. Package the first collected periodic current array as (x1, y1), where the initial value of y1 is set to 1, and add (x1, y1) to the online feature library for initialization. If the current current array x is not similar to the online feature library, then the online feature library has nothing to do with the state of the current current array x, and y1 is attenuated. If this situation persists for a period of time and y1 decays to a very small value, it means that the online feature library not only has nothing to do with the state of the current current array x, but also has nothing to do with the operation state of the household appliances for a period of time. At this time, an update operation needs to be performed on the online feature library. The update operation is shown in the second step of Step 2.
[0055] Second step: Design constant coefficients 0.01 and 0.99. When reading the real-time high-frequency current data, update the periodic current arrays in the online feature library that are similar to x. As can be seen from Equation (3), when F(x) k (k is the index value of the current array in the online feature library) is greater than the similarity threshold, the array in the corresponding online feature library needs to be updated, and the update function is shown in Equation (4).
[0056]
[0057] And the remaining current arrays in the feature library need to be updated with the attenuation coefficient, as shown in Equation (5).
[0058]
[0059] Third step: When the values of the n F(x) sliding window functions are all less than the similarity threshold, and each attenuation coefficient value is greater than the set redundancy threshold of the online feature library, the current periodic current array (x, y) can be added to the online feature library, that is, the number of periodic current arrays in the online feature library becomes (n + 1).
[0060] Step 3: Monitor the operating state of household appliances based on the periodic current waveform similarity function and the online feature library.
[0061] (1) If (x, y) is similar to a certain periodic current array in the online feature library, and the decay coefficient values corresponding to the periodic current arrays in the online feature library are all greater than the set redundancy threshold of the online feature library (in the embodiment, the redundancy threshold of the online feature library is set to 0.01), then it is determined that the current electrical appliance state is continuous.
[0062] (2) If there is a decay coefficient value y of the current array in the online feature library i < 0.01, it indicates that the set of feature arrays {(x1, y1), …, (x n , y n )} (n is the number of arrays in the online feature library) in the current online feature library is no longer suitable for monitoring the operating state of the current household appliance, and the continuous current waveform or pattern during the operation of the household appliance has changed. Therefore, it is determined that the operating state of the household appliance changes at the current moment, and the online feature library is initialized. The initialization method is shown in the first step of Step 2.
[0063] (3) If (x, y) is not similar to all periodic current arrays in the online feature library, add (x, y) to the online feature library. At the same time, calculate the decay coefficient values corresponding to each periodic current array in the online feature library according to Equation (5). If y < 0.01, it is determined that the operating state of the current power load has changed; otherwise, it is determined to be continuous.
[0064] Embodiment
[0065] To verify the effectiveness of the proposed solution of the present invention, the following experiments are carried out.
[0066] Due to the different internal circuit structures and uses of various household appliances, there are respective typical characteristics in the steady-state periodic current waveforms under continuous working conditions. Figure 2 The current waveforms of three devices in normal working states are shown in. According to the different operating states of household appliances during operation, electrical appliances can be divided into start-stop two-state devices (denoted as ON / OFF) represented by an electric kettle, finite multi-state devices (finnite state machin, denoted as FSM) represented by a hair dryer, and continuously varing state devices (continuously varing state machin, denoted as CVSM) represented by an air conditioner. In this embodiment, the NILM residential load recording device of the State Grid Corporation is used to sample household appliances, and the sampling frequency is 1.6 kHz.
[0067] Using the sliding window function, monitor the state of the current data of the ON / OFF device (taking the electric kettle as an example) for a period of time. The sliding window function in the calculation process is asFigure 3 As shown, the visualization result is as Figure 4 shown. Figure 4 The dark black part in it is the continuous current data of the electric kettle over a period of time. By analyzing the current amplitude, it can be seen that the electric kettle has only two working states, start and stop. The same working state will reappear due to the electrical appliance usage method within a certain time domain, and the periodic current waveforms and amplitudes under the same working state are consistent. Table 2 shows the characteristic current waveforms and amplitude ranges of the electric kettle state extracted from Figure 4 it.
[0068] Table 2 Characteristic Current Waveforms of Electric Kettle State
[0069]
[0070] For FSM devices, their working states are limited, and the periodic current waveforms and amplitudes under the same working state are consistent. It can be regarded as a generalized ON / OFF device. Similarly, the sliding window function is used to monitor the current data within a specific time domain to obtain the switching time points of each operating state, which are marked with different gray-scale color blocks, and the current waveforms and amplitudes of the same gray-scale color blocks are similar. Taking a hair dryer with two different working states, cold and hot, as an example, after the sliding window function is used to monitor the device operating state, the current data of the cold air operating state, hot air operating state, and shutdown state are respectively marked with different gray-scale color blocks. The visualization result is as Figure 5 shown.
[0071] For CVSM devices, their operating states are variable, and the periodic current waveforms and amplitudes under the same working state are not consistent. Therefore, the above method cannot be used for load monitoring of such devices. According to the regular fluctuations of the periodic current when this type of device works, as shown by the continuous current of the air conditioner in Figure 2 , the sliding window is improved to an online feature library, and the online feature library is used as the reference data for judging the operating state of household appliances in the sliding window function. Taking the air conditioner as an example, its visualization identification result is as Figure 6 shown
[0072] From the above research materials, it can be seen that the method of the present invention can better achieve the task of monitoring the operating state of CVSM devices, and is also applicable to the operating state monitoring scenarios of ON / OFF devices and FSM devices.
[0073] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.
[0074] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.
Claims
1. A method for monitoring the operating status of household appliances based on an online feature library, characterized in that, It includes the following steps: Step 1: Using the Pearson similarity coefficient, establish a periodic current waveform similarity function and a sliding window function; Step 2: Based on the periodic current array during the operation state monitoring of household appliances, establish an online feature library, and update the array in the online feature library in real time according to the needs of state changes; Step 3: Based on the periodic current waveform similarity function and the online feature library, monitor the operation state of household appliances; Step 1: Using the Pearson similarity coefficient, establish a periodic current waveform similarity function and a sliding window function for judging the operation state of household appliances. The specific method is as follows: Step 1.1: Define a Pearson similarity coefficient function between two periodic-level current arrays, denoted as the periodic current waveform similarity function, as shown in Equation (1): Where x is the periodic current array of the device to be sampled at the current moment, and x i is the previous consecutive periodic current array that is temporally adjacent to the periodic current array x; Step 1.2: Based on the periodic current waveform similarity function, establish a sliding window function, as shown in Equation (2); Wherein, {x 1, ..., x m} is a set of current arrays in a sliding window for judging the state of the current array x of the current period, m is a dynamic index of the number of current arrays in the sliding window, α is a sliding window attenuation coefficient, and m - i is the exponent of the attenuation coefficient between the current period current array x and the current array x i in the sliding window; F(x) is the cumulative sum of the products of the similarity coefficient and the attenuation coefficient between the current period current array x and each continuous period current array in the sliding window, and the effective value range is between 0 and 1; Step 2: Based on the typical periodic current array during the operation state monitoring of household appliances, establish an online feature library, and update the array in the online feature library in real time according to the needs of state changes. The specific method is as follows: Step 2.1: Collect the high-frequency current data of household appliances in the residents' homes. Package the first collected periodic current array as (x1, y1), and add (x1, y1) to the online feature library for initialization, where y1 is the feature attenuation coefficient corresponding to x1, and the initial value is set to 1; Step 2.2: Perform an update process on the periodic current array in the online feature library that is similar to the periodic current array x at the current moment. The update function is as shown in Equation (4), and the remaining current arrays in the feature library need to be updated with the attenuation coefficient, as shown in Equation (5); Among them, k is the current array index value in the online feature library, and n is the number of periodic current arrays in the online feature library at the current moment; Step 2.3: If the sliding window function values of the periodic current array x at the current moment and the periodic current arrays in the feature library are both less than the similarity threshold, and each attenuation coefficient value meets the threshold requirements, then add (x, y) to the online feature library, and the number of periodic current arrays in the online feature library becomes (n + 1), where y is the feature attenuation coefficient corresponding to x, otherwise the online feature library is not updated; Step 3: Based on the periodic current waveform similarity function and the online feature library, monitor the operation state of household appliances. The specific method is as follows: (1) When (x, y) is similar to a certain periodic current array (x i , y i ) in the online feature library, and the attenuation coefficient values of all periodic current arrays in the online feature library are greater than the set redundancy threshold of the online feature library, it is determined that the operating state of the current household appliance is continuous; (2) When (x, y) is not similar to all the periodic current arrays (x i , y i ) in the online feature library, add (x, y) to the online feature library. If the attenuation coefficient value of the periodic current array in the online feature library is less than the set redundancy threshold of the online feature library, it is determined that the operation state of the current household appliance is interrupted; otherwise, it is determined to be continuous. (3) When the decay coefficient y of a certain cycle current array (x i , y i ) in the online feature library i is less than the set redundancy threshold of the online feature library, it indicates that the set of feature arrays in the online feature library {(x1, y1), …, (x n , y n )} is no longer suitable for monitoring the operating state of the current household appliance, that is, the continuous current waveform or pattern has changed during the operation of the household appliance. At this time, without considering the similarity between (x, y) and the cycle current arrays in the online feature library, it is determined that the operating state of the household appliance has changed at the current moment, and the online feature library is re-initialized simultaneously.
2. A system for monitoring the operating status of household appliances based on an online feature library, characterized in that, Implement the operation state monitoring of household appliances based on the online feature library through the operation state monitoring method of household appliances based on the online feature library described in Claim 1.
3. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, the method for the operating status of household appliances based on the online feature library as claimed in claim 1 is implemented.
4. A computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for monitoring the operating status of household appliances based on the online feature library as claimed in claim 1 is implemented.