Electric bicycle illegal charging behavior monitoring method based on non-intrusive load monitoring

Abstract

An electric bicycle illegal charging behavior monitoring method based on non-intrusive load monitoring comprises the following steps: 1) using a data acquisition card or a sensor to synchronously acquire voltage and current data during operation of an electric appliance in a circuit; 2) performing time domain simulation on the current and voltage characteristic data to obtain a preliminary charging voltage and current oscillogram; analyzing the oscillogram, and searching and eliminating error data; preprocessing through data cleaning, integration and conversion; 3) signal smoothing and noise removal; 4) establishing an EBCL (electric bicycle charging load) template library; 5) matching the real-time circuit data with the EBCL template library according to the characteristics of the constant-voltage charging stage of the electric bicycle, and identifying the charging load of the electric bicycle; and 6) if the illegal charging behavior of the electric bicycle is found, result feedback is carried out through an alarm or a monitoring platform. According to the invention, illegal charging behavior monitoring of the electric bicycle is effectively realized, and an efficient and simple monitoring alarm mode is established.

Description

Technical Field

[0001] The invention belongs to the field of electric bicycle monitoring and relates to a method for monitoring illegal charging behavior of an electric bicycle. Background Art

[0002] Non-intrusive load monitoring technology (NILM) is a method of monitoring power loads. It does not require the installation of sensors on each user's electrical equipment, but instead installs a monitoring device at the user's bus entrance. For example, NILM can install smart meters or smart circuit breakers at the user's electricity meter, use advanced signal processing and data analysis technology to extract features of the electrical quantities of the total power load (such as active power, reactive power, etc.), and then decompose the total equipment power consumption into the power consumed by each electrical device. This technology can identify the working status of each electrical appliance and analyze the user's power consumption behavior, thereby providing users with a more refined power management solution.

[0003] The online identification of electric bicycle charging load mainly uses NILM technology, which measures and collects signals at the user's home wires and combines intelligent algorithms to identify the charging load of electric bicycles. However, the online identification technology of electric bicycle charging load implemented by NILM currently has the following two difficulties:

[0004] 1) Complexity of feature extraction: The feature extraction of electric bicycle charging load involves in-depth analysis of power data such as current and voltage. Since electric bicycle chargers may exhibit nonlinear and severely distorted current waveforms during charging, and have high instantaneous power peaks, these characteristics make it difficult to accurately extract the features of electric bicycle charging loads from complex power data. Different types of electric bicycle chargers may have different load characteristics, which increases the complexity of feature extraction.

[0005] 2) Accuracy of pattern recognition: Online identification of electric bicycle charging loads relies on pattern recognition algorithms, which need to be able to accurately identify the charging load of electric bicycles from a large number of power-consuming devices. However, due to the complexity and uncertainty of power data and the diversity of power-consuming devices, the accuracy of pattern recognition is challenged. Especially when multiple devices are running at the same time, how to accurately distinguish the loads of different devices and avoid misjudgment and missed judgment is a key problem that pattern recognition algorithms need to solve. Summary of the invention

[0006] In order to overcome the shortcomings of the existing technology, in order to achieve the deep integration of the non-invasive load monitoring system and the electric bicycle illegal charging monitoring system, and establish an efficient and concise monitoring and alarm mode, the present invention provides an electric bicycle illegal charging behavior monitoring method based on non-invasive load monitoring, and utilizes the characteristics of the electric bicycle in the constant voltage charging stage to establish an electric bicycle charging load (EBCL) template. By identifying the load characteristics of the electric bicycle in the constant voltage charging stage, the active power slow slope decrease and the reactive power slow slope increase, it is determined whether there is electric vehicle charging behavior in the circuit, and an early warning is issued according to the actual situation and reported to the management personnel, so as to reduce fires and casualties caused by illegal charging of electric bicycles.

[0007] The technical solution adopted by the present invention to solve its technical problem is:

[0008] A method for monitoring illegal charging behavior of electric bicycles based on non-intrusive load monitoring comprises the following steps:

[0009] 1) Data Acquisition

[0010] Use data acquisition cards or sensors (such as voltage sensors and current sensors) to synchronously collect the voltage and current data of electrical appliances in the circuit when they are running; at the same time, ensure that the sampling frequency meets certain requirements (such as 6000Hz) to ensure that subtle changes in the charging process of the electric bicycle can be captured.

[0011] 2) Data preprocessing

[0012] Perform time domain simulation on current and voltage characteristic data to obtain preliminary charging voltage and current waveforms; analyze the waveforms to determine whether there are errors in the data, find and eliminate erroneous data; make the data neater and more orderly through pre-processing steps such as data cleaning, integration and conversion, thereby accelerating the subsequent data processing and analysis process;

[0013] 3) Data analysis, the process is as follows:

[0014] Signal smoothing: The Savitzky-Golay filtering method is used to process the power data. The data is fitted with a local polynomial of a sliding window and the fitted polynomial is used to replace the original signal for smoothing. The first-order and second-order derivative information of the power signal is retained, thereby better retaining the characteristics of the signal.

[0015] Noise removal: Savitzky-Golay filter is used to retain high-frequency signal components and remove high-frequency noise and fluctuations in power monitoring signals. For example, in the load active power sequence data monitored by smart meters, Savitzky-Golay filtering can effectively remove noise and provide a more accurate data basis for subsequent load decomposition;

[0016] 4) Establish an EBCL template library for electric bicycle charging load

[0017] Through data collection, data analysis, template design, parameter setting, verification and optimization, documentation and maintenance, an accurate and practical EBCL template library is established to provide support for the monitoring and management of electric bicycle charging loads;

[0018] 5) Identification and judgment

[0019] By matching the real-time circuit data with the EBCL template library based on the characteristics of the constant voltage charging stage of electric bicycles, the charging load of electric bicycles can be accurately identified and distinguished from other household appliance loads;

[0020] 6) Result feedback

[0021] If an electric bicycle is found to be charging in violation of regulations, feedback will be provided through an alarm or monitoring platform.

[0022] Further, in step 4), the process of establishing the EBCL template library is as follows:

[0023] NILM collects charging data of electric bicycles under different conditions, including charging data of different models, different battery capacities, different charging equipment and different charging environments; the data includes key information such as charging power, charging time, and charging start and end time; the charging data is analyzed in detail to understand the basic characteristics of the charging load of electric bicycles, including the changing pattern of charging power and the distribution of charging time;

[0024] According to the data analysis results, the basic structure of the EBCL template is designed. The template can reflect the main characteristics of the electric bicycle charging load, including the charging power curve and the charging time distribution. Considering that the operation time of the electric bicycle charging load is long and it may overlap with other electrical loads, the template design should have certain flexibility and adaptability. Parameters are set in the template, including the charging power range and the charging time range. These parameters are adjusted according to actual data and needs to ensure the accuracy and practicality of the template.

[0025] Use actual data to verify the established EBCL template to check whether it can accurately reflect the characteristics of the electric bicycle charging load. According to the verification results, optimize and adjust the template to improve its accuracy and practicality. Document the established EBCL template, including the template's design principle, structure, parameter setting and usage method, establish a complete EBCL template library, and regularly maintain and update the EBCL template library to adapt to changes in electric bicycle technology and the market.

[0026] Furthermore, in the step 5), first, the difference of the power signal is calculated, and then the continuous negative subsequences in the active power differential signal are focused on and matched with the continuous positive subsequences in the reactive power differential signal; once the matching continuous subsequences are found, the distances between these subsequences and the EBCL template are calculated; if the calculated distance is less than the preset threshold, this means that the electric bicycle battery is in the constant voltage charging stage, and there may be illegal charging at this time; in this case, the system recommends that the user immediately stop the illegal charging behavior or send a message to the management personnel; if the calculated distance is not less than the threshold, it is considered that the electric bicycle battery is not currently in the constant voltage charging stage, or there is no charging load; then the sliding window is moved backward for a set time period, and the data of the total active power P and the total reactive power Q in the next set time area are continued to be analyzed.

[0027] In step 5), the matching process between the circuit data and the EBCL template library is as follows:

[0028] 5.1) Read the real-time circuit data from the NILM system, remove abnormal values, missing values ​​or invalid data, and after Savitzky-Golay filtering and differentiation, obtain the active power differential signal divided into a continuous negative subsequence, and the reactive power differential signal divided into a continuous positive subsequence;

[0029] 5.2) Feature extraction

[0030] During the constant voltage charging stage of an electric bicycle, the decreasing slope of its active power and the increasing slope of its reactive power show precise synchronization in time; therefore, in order to ensure accurate analysis of the charging process, special attention should be paid to and matching the continuously decreasing active power differential signal subsequences and the continuously increasing reactive power differential signal subsequences; these subsequences should overlap in time to accurately reflect the synchronization characteristics of power changes;

[0031] 5.3) Template loading

[0032] Load the EBCL template and load the pre-defined EBCL template from the template library, including the template's charging power range and charging time range parameters.

[0033] 5.4) Template Matching:

[0034] Calculate the difference of the active power differential signal subsequence obtained by feature extraction in 5.2), and calculate the distance with the template in the EBCL template library. The calculation formula is:

[0035]

[0036] Where m i It represents the differential power template at time i, ni It represents the difference of the active power differential signal subsequence obtained at the same time, k is the number of data in the signal subsequence, and ΔL is the calculated distance value;

[0037] 5.5) Matching result output

[0038] Matching success: If the ΔL of the actual circuit data is less than the set threshold ε, it is considered a successful match, and there is an electric bicycle battery in the circuit that is in the constant voltage charging stage, which may be a case of illegal charging. At the same time, the result is broadcasted or telephoned to stop the illegal charging behavior, or fed back to the management personnel through the monitoring platform;

[0039] Matching failure: If the ΔL of the actual circuit data is not less than the set threshold ε, it is judged as a matching failure. There is no electric bicycle battery in the circuit that is in the constant voltage charging stage. The system will continue to monitor the circuit's active power P and total reactive power Q data.

[0040] The technical concept of the present invention is as follows: the batteries on the electric bicycle market are mainly divided into two categories: lithium-ion batteries and lead-acid batteries. During the charging process, lithium-ion batteries usually go through two stages: constant current charging and constant voltage charging, while lead-acid batteries additionally include a floating charging stage. Regardless of the type of battery, electric bicycles will go through a significant constant voltage charging stage when charging, which is significantly different from other electrical loads.

[0041] During the constant voltage charging stage, as the battery is gradually charged, the active power of the charger (that is, the power actually transmitted to the battery) will show a gradual downward trend. This is because as the battery voltage rises, the charging current will gradually decrease, resulting in a decrease in active power. And the reduction in active power is not sudden, but forms a relatively gentle slope. The reactive power characteristics are opposite to the active power. During the constant voltage charging stage, the reactive power of the charger (that is, the reactive power caused by the grounding capacitor current on the line) will show a gradual upward trend. By identifying the load characteristics of the slow slope of active power decrease and the slow slope of reactive power increase of electric bicycles during the constant voltage charging stage, it can be determined whether there is any illegal charging of electric bicycles in the circuit.

[0042] In the present invention, circuit data of densely populated buildings such as residential buildings and student dormitories are analyzed through non-intrusive load monitoring (NILM), the characteristics of electric vehicle charging are mined from historical data, and an electric bicycle charging load template library is established. By real-time monitoring of the load characteristics of the active power decline slope and reactive power rise slope of electric bicycles in the constant voltage charging stage, the charging behavior of electric vehicles in the circuit is identified and an alarm message is generated.

[0043] The beneficial effects of the present invention are mainly manifested in: effectively realizing the monitoring of illegal charging behavior of electric bicycles and establishing an efficient and concise monitoring and alarm method. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a principle block diagram of a method for monitoring illegal charging behavior of electric bicycles based on non-intrusive load monitoring.

[0045] Figure 2 The graphs are the slow slope of active power decrease and the slow slope of reactive power increase of the electric bicycle during the constant voltage charging stage, where (1) and (3) respectively show the changes of active power and reactive power over time, and (2) and (4) respectively show the changes of the differential signals of active power and reactive power over time. DETAILED DESCRIPTION

[0046] The present invention will be further described below in conjunction with the accompanying drawings.

[0047] Reference Figure 1 and Figure 2 , a method for monitoring illegal charging behavior of electric bicycles based on non-intrusive load monitoring, comprising the following steps:

[0048] 1) Data Acquisition

[0049] Use data acquisition cards or sensors (such as voltage sensors and current sensors) to synchronously collect the voltage and current data of electrical appliances in the circuit when they are running; at the same time, ensure that the sampling frequency meets certain requirements (such as 6000Hz) to ensure that subtle changes in the charging process of the electric bicycle can be captured.

[0050] 2) Data preprocessing

[0051] Perform time domain simulation on current and voltage characteristic data to obtain preliminary charging voltage and current waveforms; analyze the waveforms to determine whether there are errors in the data, find and eliminate erroneous data; make the data neater and more orderly through pre-processing steps such as data cleaning, integration and conversion, thereby accelerating the subsequent data processing and analysis process;

[0052] 3) Data analysis, the process is as follows:

[0053] Signal smoothing: The Savitzky-Golay filtering method is used to process the power data. The data is fitted with a local polynomial of a sliding window and the fitted polynomial is used to replace the original signal for smoothing. The first-order and second-order derivative information of the power signal is retained, thereby better retaining the characteristics of the signal.

[0054] Noise removal: Savitzky-Golay filter is used to retain high-frequency signal components and remove high-frequency noise and fluctuations in power monitoring signals. For example, in the load active power sequence data monitored by smart meters, Savitzky-Golay filtering can effectively remove noise and provide a more accurate data basis for subsequent load decomposition;

[0055] 4) Establish an EBCL template library for electric bicycle charging load

[0056] Through data collection, data analysis, template design, parameter setting, verification and optimization, documentation and maintenance, an accurate and practical EBCL template library is established to provide support for the monitoring and management of electric bicycle charging loads;

[0057] In the step 4), the process of establishing the EBCL template library is as follows:

[0058] NILM collects charging data of electric bicycles under different conditions, including charging data of different models, different battery capacities, different charging equipment and different charging environments; the data includes key information such as charging power, charging time, and charging start and end time; the charging data is analyzed in detail to understand the basic characteristics of the charging load of electric bicycles, including the changing pattern of charging power and the distribution of charging time;

[0059] According to the data analysis results, the basic structure of the EBCL template is designed. The template can reflect the main characteristics of the electric bicycle charging load, including the charging power curve and the charging time distribution. Considering that the operation time of the electric bicycle charging load is long and it may overlap with other electrical loads, the template design should have certain flexibility and adaptability. Parameters are set in the template, including the charging power range and the charging time range. These parameters are adjusted according to actual data and needs to ensure the accuracy and practicality of the template.

[0060] Use actual data to verify the established EBCL template to check whether it can accurately reflect the characteristics of the electric bicycle charging load. According to the verification results, optimize and adjust the template to improve its accuracy and practicality. Document the established EBCL template, including the template's design principle, structure, parameter setting and usage method, establish a complete EBCL template library, and regularly maintain and update the EBCL template library to adapt to changes in electric bicycle technology and the market.

[0061] 5) Identification and judgment

[0062] By matching the real-time circuit data with the EBCL template library based on the characteristics of the constant voltage charging stage of electric bicycles, the charging load of electric bicycles can be accurately identified and distinguished from other household appliance loads;

[0063] In the step 5), first, the difference of the power signal is calculated, and then the continuous negative subsequences in the active power differential signal are focused on and matched with the continuous positive subsequences in the reactive power differential signal; once the matching continuous subsequences are found, the distances between these subsequences and the EBCL template are calculated; if the calculated distance is less than a preset threshold, this means that the electric bicycle battery is in a constant voltage charging stage, and there may be illegal charging at this time; in this case, the system recommends that the user immediately stop the illegal charging behavior or send a message to the management personnel; if the calculated distance is not less than the threshold, it is considered that the electric bicycle battery is not currently in a constant voltage charging stage, or there is no charging load; then the sliding window is moved backward for a set time period, and the data of the total active power P and the total reactive power Q in the next set time area are continuously analyzed.

[0064] In step 5), the matching process between the circuit data and the EBCL template library is as follows:

[0065] 5.1) Read the real-time circuit data from the NILM system, remove abnormal values, missing values ​​or invalid data, and after Savitzky-Golay filtering and differentiation, obtain the active power differential signal divided into a continuous negative subsequence, and the reactive power differential signal divided into a continuous positive subsequence;

[0066] 5.2) Feature extraction

[0067] During the constant voltage charging stage of an electric bicycle, the decreasing slope of its active power and the increasing slope of its reactive power show precise synchronization in time; therefore, in order to ensure accurate analysis of the charging process, special attention should be paid to and matching the continuously decreasing active power differential signal subsequences and the continuously increasing reactive power differential signal subsequences; these subsequences should overlap in time to accurately reflect the synchronization characteristics of power changes;

[0068] 5.3) Template loading

[0069] Load the EBCL template and load the pre-defined EBCL template from the template library, including the template's charging power range and charging time range parameters.

[0070] 5.4) Template Matching:

[0071] Calculate the difference of the active power differential signal subsequence obtained by feature extraction in 5.2), and calculate the distance with the template in the EBCL template library. The calculation formula is:

[0072]

[0073] Where m i It represents the differential power template at time i, ni It represents the difference of the active power differential signal subsequence obtained at the same time, k is the number of data in the signal subsequence, and ΔL is the calculated distance value;

[0074] 5.5) Matching result output

[0075] Matching success: If the ΔL of the actual circuit data is less than the set threshold ε, it is considered a successful match, and there is an electric bicycle battery in the circuit that is in the constant voltage charging stage, which may be a case of illegal charging. At the same time, the result is broadcasted or telephoned to stop the illegal charging behavior, or fed back to the management personnel through the monitoring platform;

[0076] Matching failure: If the ΔL of the actual circuit data is not less than the set threshold ε, it is judged as a matching failure. There is no electric bicycle battery in the circuit that is in the constant voltage charging stage. The system will continue to monitor the circuit's data on active power P and total reactive power Q;

[0077] 6) Result feedback

[0078] If an electric bicycle is found to be charging in violation of regulations, feedback will be provided through an alarm or monitoring platform.

[0079] In this embodiment, through a non-invasive load monitoring system and utilizing the load characteristics of the electric bicycle's active power decreasing slope and reactive power increasing slope during the constant voltage charging stage, it is possible to determine whether there is electric vehicle charging behavior in the circuit, issue an early warning based on the actual situation and report to the management personnel.

[0080] The contents described in the embodiments of this specification are merely an enumeration of the implementation forms of the inventive concept and are for illustrative purposes only. The protection scope of the present invention should not be considered to be limited to the specific forms described in this embodiment, and the protection scope of the present invention also extends to equivalent technical means that can be thought of by ordinary technicians in this field based on the inventive concept.

Claims

1. A method for monitoring illegal charging behavior of electric bicycles based on non-intrusive load monitoring, characterized in that: The method comprises the following steps: 1) Data Acquisition Use data acquisition cards or sensors to synchronously collect voltage and current data of electrical appliances in the circuit when they are running; at the same time, ensure that the sampling frequency meets the set requirements to ensure that subtle changes in the charging process of the electric bicycle can be captured; 2) Data preprocessing Perform time domain simulation on current and voltage characteristic data to obtain preliminary charging voltage and current waveforms; analyze the waveforms to determine whether there are errors in the data, find and eliminate erroneous data; make the data neater and more orderly through data cleaning, integration and conversion preprocessing, thereby accelerating the subsequent data processing and analysis process; 3) Data analysis, the process is as follows: Signal smoothing: The Savitzky-Golay filtering method is used to process the power data. The data is fitted with a local polynomial of a sliding window and the fitted polynomial is used to replace the original signal for smoothing. The first-order and second-order derivative information of the power signal is retained, thereby better retaining the characteristics of the signal. Noise removal: Use Savitzky-Golay filter to retain high-frequency signal components and remove high-frequency noise and fluctuations in power monitoring signals; 4) Establish an EBCL template library for electric bicycle charging load Through data collection, data analysis, template design, parameter setting, verification and optimization, documentation and maintenance, an accurate and practical EBCL template library is established to provide support for the monitoring and management of electric bicycle charging loads; 5) Identification and judgment By matching the real-time circuit data with the EBCL template library based on the characteristics of the constant voltage charging stage of electric bicycles, the charging load of electric bicycles can be accurately identified and distinguished from other household appliance loads; 6) Result feedback If an electric bicycle is found to be charging in violation of regulations, feedback will be provided through an alarm or monitoring platform.

2. The method for monitoring illegal charging behavior of electric bicycles based on non-invasive load monitoring as claimed in claim 1, characterized in that: In the step 4), the process of establishing the EBCL template library is as follows: NILM collects charging data of electric bicycles under different conditions, including charging data of different models, different battery capacities, different charging equipment and different charging environments; the data includes key information such as charging power, charging time, and charging start and end time; Conduct detailed analysis of charging data to understand the basic characteristics of e-bike charging load, including the changing pattern of charging power and charging time distribution; According to the data analysis results, the basic structure of the EBCL template is designed. The template can reflect the main characteristics of the electric bicycle charging load, including the charging power curve and the charging time distribution. Considering that the operation time of the electric bicycle charging load is long and it may overlap with other electrical loads, the template design should have certain flexibility and adaptability. Parameters are set in the template, including the charging power range and the charging time range. These parameters are adjusted according to actual data and needs to ensure the accuracy and practicality of the template. Use actual data to verify the established EBCL template to check whether it can accurately reflect the characteristics of the electric bicycle charging load. According to the verification results, optimize and adjust the template to improve its accuracy and practicality. Document the established EBCL template, including the template's design principle, structure, parameter setting and usage method, establish a complete EBCL template library, and regularly maintain and update the EBCL template library to adapt to changes in electric bicycle technology and the market.

3. The method for monitoring illegal charging behavior of electric bicycles based on non-intrusive load monitoring as claimed in claim 1 or 2, characterized in that: In the step 5), first, the difference of the power signal is calculated, and then the continuous negative subsequences in the active power differential signal are focused on and matched with the continuous positive subsequences in the reactive power differential signal; once the matching continuous subsequences are found, the distances between these subsequences and the EBCL template are calculated; if the calculated distance is less than a preset threshold, this means that the electric bicycle battery is in a constant voltage charging stage, and there may be illegal charging at this time; in this case, the system recommends that the user immediately stop the illegal charging behavior or send a message to the management personnel; if the calculated distance is not less than the threshold, it is considered that the electric bicycle battery is not currently in a constant voltage charging stage, or there is no charging load; then the sliding window is moved backward for a set time period, and the data of the total active power P and the total reactive power Q in the next set time area are continuously analyzed.

4. The method for monitoring illegal charging behavior of electric bicycles based on non-intrusive load monitoring as claimed in claim 3 is characterized in that: In step 5), the matching process between the circuit data and the EBCL template library is as follows: 5.1) Read the real-time circuit data from the NILM system, remove abnormal values, missing values ​​or invalid data, and after Savitzky-Golay filtering and differentiation, obtain the active power differential signal divided into a continuous negative subsequence, and the reactive power differential signal divided into a continuous positive subsequence; 5.2) Feature extraction: During the constant voltage charging stage of an electric bicycle, the decreasing slope of its active power and the increasing slope of its reactive power show precise synchronization in time; therefore, in order to ensure accurate analysis of the charging process, special attention should be paid to and matching the continuously decreasing active power differential signal subsequences and the continuously increasing reactive power differential signal subsequences; these subsequences should overlap in time to accurately reflect the synchronization characteristics of power changes; 5.3) Template loading: Load the EBCL template, and load the pre-defined EBCL template from the template library, including the charging power range and charging time range parameters of the template; 5.4) Template Matching: Calculate the difference of the active power differential signal subsequence obtained by feature extraction in 5.2), and calculate the distance with the template in the EBCL template library. The calculation formula is: Where m i It represents the differential power template at time i, n i It represents the difference of the active power differential signal subsequence obtained at the same time, k is the number of data in the signal subsequence, and ΔL is the calculated distance value; 5.5) Matching result output: Matching success: If the ΔL of the actual circuit data is less than the set threshold ε, it is considered a successful match, and there is an electric bicycle battery in the circuit that is in the constant voltage charging stage, which may be a case of illegal charging. At the same time, the result is broadcasted or telephoned to stop the illegal charging behavior, or fed back to the management personnel through the monitoring platform; Matching failure: If the ΔL of the actual circuit data is not less than the set threshold ε, it is judged as a matching failure. There is no electric bicycle battery in the circuit that is in the constant voltage charging stage. The system will continue to monitor the circuit's active power P and total reactive power Q data.

Images