Security analysis method for elderly people living alone based on electricity consumption behaviors

Abstract

The invention relates to the technical field of data processing, and discloses a safety analysis method for elderly people living alone based on power consumption behaviors. Comprising the steps of generating a user power utilization request, judging whether a power utilization behavior safety analysis requirement is met or not, obtaining comprehensive behavior data of a user side in a power utilization interval, predicting a power utilization state value, judging whether a power utilization abnormity early warning prompt is sent or not, identifying an abnormal behavior from the comprehensive behavior data, and formulating safety early warning information. Compared with the prior art, the method can avoid the phenomenon of inaccurate safety analysis caused by abnormal conditions such as power failure in a power supply plan, can accurately predict the electricity utilization state value of the residence of the elderly living alone according to the comprehensive behavior data in combination with the machine learning model, and can accurately analyze the electricity utilization state value of the residence of the elderly living alone when the electricity utilization potential safety hazard occurs. And safety early warning information is formulated in time and in a targeted manner, so that related departments and personnel can secondarily confirm whether the elderly living alone is safe or not, and more serious electricity utilization safety accidents are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and more specifically, to a safety analysis method for elderly people living alone based on electricity usage behavior. Background Art

[0003] The patent application with reference publication number CN117609794A discloses a method and system for identifying the safety behavior of elderly people living alone through electrical parameter features and algorithms. The method collects living habit information through user APP and equipment related to elderly care living alone, establishes a basic daily safety behavior model, and if the collected daily behavior characteristics of the elderly living alone have a high degree of match with the basic daily safety behavior model, if the learning time limit threshold set by the child-side APP user is met, the basic daily safety behavior is output. If the matching degree with the model is not satisfied, the current daily behavior characteristics are extracted and abnormal behavior alarms are output. After confirmation by the child-side or parent-side APP, the abnormal behavior characteristics are supplemented as new feature information to the basic daily behavior model, and the electrical parameter characteristics are used as the core, assisted by the number of entry and exit characteristics and the number of night-time wake-up characteristics, so as to establish a more accurate safety behavior model, which can accurately identify the daily safety behavior of elderly people living alone and achieve a more accurate and intelligent effect.

[0004] The existing safety analysis method for elderly people living alone collects various electricity usage behaviors of elderly people living alone in real time and continuously, and compares the electricity usage behaviors with the pre-stored standard electricity usage behaviors one by one, and analyzes the safety of elderly people living alone through the different behaviors. Although the safety analysis of the electricity usage behaviors of elderly people living alone can be achieved, the real-time and uninterrupted collection method will lead to a large amount of electricity usage behaviors participating in the safety analysis, thereby lowering the threshold for safety analysis and increasing the burden of safety analysis. It is also impossible to formulate targeted early warning measures for situations with safety hazards, and thus it is impossible to confirm the personal safety of elderly people living alone in the first time, reducing the safety of electricity use.

[0005] In view of this, the present invention proposes a safety analysis method for elderly people living alone based on electricity usage behavior to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solutions: a safety analysis method for elderly people living alone based on electricity usage behavior, applied to a security cloud platform, comprising:

[0007] S1: Receive the power consumption area and power consumption interval of the user, and combine the power consumption area and power consumption interval to generate the user's power consumption request;

[0008] S2: Filter out the user power supply plan corresponding to the user's power consumption request from the power database, identify the power supply status of the user power supply plan, and determine whether it meets the power consumption behavior safety analysis requirements; if it meets the power consumption behavior safety analysis requirements, execute S3; if it does not meet the power consumption behavior safety analysis requirements, repeat S1-S2;

[0009] S3: Obtain comprehensive behavior data of the user end within the power consumption interval, the comprehensive behavior data including time period type, unit power consumption excess value, abnormal power consumption frequency and abnormal power consumption duration;

[0010] S4: Input the collected comprehensive behavior data into the machine learning model trained in advance, predict the power consumption status value, and determine whether to issue an abnormal power consumption warning prompt. The power consumption status value includes a normal power consumption status and an abnormal power consumption status. If an abnormal power consumption warning prompt is issued, execute S5; if no abnormal power consumption warning prompt is issued, repeat S3-S4.

[0011] S5: Identify abnormal behaviors from the comprehensive behavior data, formulate safety warning information, and send corresponding safety warning information to the warning terminal.

[0012] Furthermore, the method for obtaining the electricity consumption area and the electricity consumption interval includes:

[0013] Mark the location of the residence of the elderly living alone on the electronic map and record it as the user point;

[0014] Taking the user point as the base point, four extension lines are drawn in the horizontal and vertical directions.

[0015] The road that first intersects the four extension lines on the electronic map is recorded as a regional road, and the area enclosed by the four regional roads is recorded as a power consumption area;

[0016] The electronic calendar of the electricity consumption area is queried through the calendar system, and the electronic date and electronic time in the electronic calendar are identified through natural language processing technology;

[0017] A basic interval having a first interval position and a second interval position is constructed, and the electronic date and the electronic time are respectively introduced into the first interval position and the second interval position of the basic interval to obtain a power consumption interval.

[0018] Furthermore, the method for generating the user's electricity consumption request includes:

[0019] Constructing a basic request with two request bits, and marking the two request bits as a first request bit and a second request bit in sequence;

[0020] Query the area number corresponding to the electricity consumption area through the administrative area system, and import the area number into the first request bit of the basic request to form a number request bit;

[0021] Mark the first digital group and the second digital group of the power consumption interval one by one, connect the first digital group and the second digital group end to end to form an interval group, and import the interval group into the second request bit of the basic request to form an interval request bit;

[0022] The basic request with the number request bit and the interval request bit is recorded as the user power request.

[0023] Furthermore, the power supply state includes a continuous power supply state and a suspended power supply state;

[0024] The identification method of the continuous power supply state and the suspended power supply state includes:

[0025] Taking the area number in the number request bit as the screening condition, i power supply plans to be identified that are consistent with the area number in the number request bit are screened out from the power database;

[0026] Identify the power supply dates and power supply times of i power supply plans to be identified one by one through natural language processing technology, and obtain i power supply dates and i power supply times;

[0027] The power supply plan to be identified, in which the power supply date and power supply time are consistent with the first digit group and the second digit group in the interval request bit, is recorded as the user power supply plan;

[0028] The power supply status remarks column of the user's power supply plan is queried through the power supply management system, and the digital combination of remarks in the power supply status remarks column is identified;

[0029] When the digital combination is 1-1, the power supply state is identified as a continuous power supply state;

[0030] When the digital combination is 1-0, the power supply state is identified as a power supply suspension state.

[0031] Furthermore, the determination method of whether the power consumption behavior safety analysis requirements are met includes:

[0032] When the power supply state is a continuous power supply state, it is determined that the power consumption behavior safety analysis requirements are met;

[0033] When the power supply state is a suspended power supply state, it is determined that the power consumption behavior safety analysis requirement is not met.

[0034] Furthermore, the time period types include peak power consumption period, low-peak power consumption period and flat-peak power consumption period;

[0035] Methods for obtaining the unit electricity consumption excess value include:

[0036] In the time period corresponding to the power consumption interval, w collection time periods that are adjacent to each other are marked based on the preset power consumption duration, and the w collection time periods are numbered in ascending order;

[0037] Taking the collection period corresponding to the first number as the starting point, S collection periods are selected as valid periods in a manner of one number interval;

[0038] Mark the starting time and the ending time of the S valid time periods one by one, and use the electric energy meter to count the power consumption corresponding to the S starting time and the S ending time one by one, so as to obtain the power consumption of s starting points and s ending points;

[0039] Subtract the power consumption of the s end points from the power consumption of the s starting points one by one, and compare them with the duration of the effective period to obtain s sub-power consumption values;

[0040] The expression of the sub-power consumption value is:

[0041]

[0042] In the formula, YD zs is the electricity consumption value of the sth child, YD zds is the power consumption of the sth terminal, YD qds is the electricity consumption at the sth starting point, SC yx is the duration of the effective period;

[0043] The sub-power consumption value greater than the standard power consumption value is recorded as a sub-excess value, y sub-excess values ​​are obtained, and the y sub-excess values ​​are accumulated and averaged to obtain a unit power consumption excess value;

[0044] The expression of unit power consumption excess value is:

[0045]

[0046] In the formula, YD ce is the excess value of unit electricity consumption, YD za is the power consumption value of the ath child.

[0047] Furthermore, the method for obtaining abnormal power consumption frequency includes:

[0048] Randomly mark q non-adjacent moments in S valid time periods to obtain q detection moments;

[0049] The current sensor detects the real-time current at q detection moments respectively, obtains q detection current values, and records the detection current value greater than the standard current value as an abnormal current value;

[0050] Counting the number of abnormal current values ​​in the S effective time periods one by one, and comparing the number of abnormal current values ​​with the duration of the effective time period to obtain S sub-frequencies;

[0051] The expression of sub-frequency is:

[0052]

[0053] Where PL zs is the sth sub-frequency, SL ycs is the number of abnormal current values ​​of s, SC yx is the duration of the effective period;

[0054] Remove the maximum and minimum values ​​of the sub-frequencies, accumulate the remaining s-2 sub-frequencies and calculate the average to obtain the abnormal power consumption frequency;

[0055] The expression of abnormal power consumption frequency is:

[0056]

[0057] Where PL yc is the abnormal power consumption frequency, PL zb is the bth sub-frequency.

[0058] Furthermore, the method for obtaining the abnormal power consumption duration includes:

[0059] During the period corresponding to the power consumption interval, the real-time current at the user end is detected by the current sensor, and the moment when the real-time current reaches the preset current lower limit for the first time is marked as the zero time;

[0060] After the zero point, mark the moment when the real-time current is less than the preset current lower limit for the first time, record it as the termination moment, and record the duration from the zero point to the termination moment as the zero termination duration, and obtain v zero termination durations;

[0061] The zero-end duration greater than the safe power consumption duration is recorded as the abnormal duration, X abnormal durations are obtained, and the abnormal power consumption duration is obtained by accumulating the X abnormal durations;

[0062] The expression of abnormal power consumption duration is:

[0063]

[0064] In the formula, YD yc SC is the abnormal power consumption duration. lzc is the cth abnormal duration.

[0065] Furthermore, the training method of the machine learning model includes:

[0066] Collect multiple groups of comprehensive behavior data in normal and abnormal power consumption states in advance;

[0067] Assign the peak power consumption period to A, the low-peak power consumption period to B, and the flat-peak power consumption period to C;

[0068] The comprehensive behavior data is converted into a corresponding set of feature vectors, and the feature vectors are used as the input of the machine learning model. The power consumption state value is converted into a label corresponding to the comprehensive behavior data, the normal power consumption state is converted into 0, and the abnormal power consumption state is converted into 1. The power consumption state value corresponding to each set of comprehensive behavior data is used as the output of the machine learning model. The power consumption state value is used as the prediction target, and the sum of the prediction errors of all training data is minimized as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence and the training is stopped;

[0069] The determination method of whether to issue an abnormal power consumption warning prompt includes:

[0070] When the output of the machine learning model is 0, it is determined that no abnormal power consumption warning prompt will be issued;

[0071] When the output of the machine learning model is 1, it is determined that an abnormal power consumption warning prompt is issued.

[0072] Furthermore, the method for identifying abnormal behavior includes:

[0073] The unit power consumption excess value is compared with the standard power consumption excess value. When the unit power consumption excess value is greater than the standard power consumption excess value, the unit power consumption excess value is recorded as an abnormal behavior.

[0074] The abnormal power consumption frequency is compared with the standard power consumption frequency. When the abnormal power consumption frequency is greater than the standard power consumption frequency, the abnormal power consumption frequency is recorded as an abnormal behavior.

[0075] Compare the abnormal power usage duration with the standard power usage duration. When the abnormal power usage duration is greater than the standard power usage duration, record the abnormal power usage duration as abnormal behavior.

[0076] Safety warning information includes mild warning information and severe warning information;

[0077] The method for formulating mild warning information and severe warning information includes:

[0078] Count the number of abnormal behaviors, and when the number of abnormal behaviors reaches 1, generate a mild warning message;

[0079] When the number of abnormal behaviors is 2 or 3, a severe warning message is formulated.

[0080] Technical effects and advantages of the present invention's safety analysis method for elderly people living alone based on electricity usage behavior:

[0081] The present invention generates a user electricity request by receiving the electricity consumption area and electricity consumption interval of the user end, and combining the electricity consumption area and the electricity consumption interval, screens out the user power supply plan corresponding to the user electricity consumption request from the power database, identifies the power supply status of the user power supply plan, and determines whether it meets the power consumption behavior safety analysis requirements, obtains the comprehensive behavior data of the user end in the power consumption interval, inputs the collected comprehensive behavior data into a machine learning model trained in advance, predicts the power consumption status value, and determines whether to issue an abnormal electricity consumption warning prompt, identifies abnormal behavior from the comprehensive behavior data, formulates safety warning information, and sends the corresponding safety warning information to the warning end; compared with the prior art, by forming the user electricity request with the electricity consumption area and the electricity consumption interval, it can A preliminary verification is conducted on whether the residence of an elderly person living alone meets the safety analysis requirements for electricity usage, thereby avoiding inaccurate safety analysis caused by abnormal situations in the power supply plan such as power outages. At the same time, it also raises the threshold for safety analysis, reduces the collection and analysis of a large number of electricity usage behaviors that do not meet the safety analysis threshold, and reduces the workload of safety analysis. At the same time, combined with a machine learning model, it can accurately predict the electricity usage status value of the residence of an elderly person living alone based on comprehensive behavioral data, and formulate safety warning information in a timely and targeted manner when there are safety hazards in electricity use, thereby providing relevant departments and personnel with secondary confirmation of the safety of the elderly living alone, thereby avoiding more serious electricity safety accidents and effectively improving the electricity safety of the elderly living alone. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 A schematic flow chart of a method for analyzing the safety of elderly people living alone based on electricity usage behavior provided in the first embodiment of the present invention;

[0083] Figure 2 A schematic diagram of a safety analysis system for elderly people living alone based on electricity usage behavior provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0084] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0085] Example 1: Please refer to Figure 1 As shown, the safety analysis method for elderly people living alone based on electricity usage behavior described in this embodiment is applied to a security cloud platform, including:

[0086] S1: Receive the power consumption area and power consumption interval of the user, and combine the power consumption area and power consumption interval to generate the user's power consumption request;

[0087] The user end refers to the data end that needs to collect electricity usage behavior to determine the safety of elderly people living alone. It is used to accurately collect and store the electricity usage behavior of elderly people living alone, and can send the collected and stored data corresponding to the electricity usage behavior to the security cloud platform;

[0088] The electricity consumption area refers to the location of the residence where the elderly living alone use electricity, which can indicate the area where the elderly living alone use electricity and provide data support for subsequent user electricity requests. The electricity consumption interval refers to the date and time when the elderly living alone use electricity, which can indicate the time of the elderly living alone's electricity use behavior;

[0089] Methods for obtaining electricity consumption areas and electricity consumption intervals include:

[0090] Mark the location of the residence of the elderly living alone on the electronic map and record it as the user point;

[0091] Taking the user point as the base point, four extension lines are drawn in the horizontal and vertical directions.

[0092] The road that intersects the four extension lines for the first time on the electronic map is recorded as a regional road, and the area enclosed by the four regional roads is recorded as the power consumption area; by intersecting the extension lines with the roads, the accuracy of the corresponding position of the power consumption area can be ensured, and the deviation of the geographical location can be minimized as much as possible;

[0093] The electronic calendar of the electricity consumption area is queried through the calendar system, and the electronic date and electronic time in the electronic calendar are identified through natural language processing technology;

[0094] A basic interval with a first interval position and a second interval position is constructed, and the electronic date and electronic time are respectively imported into the first interval position and the second interval position of the basic interval to obtain a power consumption interval. The basic interval is a blank structure that constitutes the power consumption interval, and the interval position is the smallest unit that constitutes the basic interval and provides an accurate position for the import of the electronic date and electronic time.

[0095] It should be noted that after the power consumption area and the power consumption interval are obtained, the power consumption area and the power consumption interval can be effectively combined, so that the power consumption area and the power consumption interval can be bundled together and form a request that can be used for subsequent judgment.

[0096] A user electricity request is a request generated by combining electricity consumption areas and electricity consumption intervals, which can be used for subsequent electricity consumption behavior safety analysis to ensure that the user's electricity consumption areas and electricity consumption intervals can be accurately and effectively analyzed, and provide an execution basis for subsequent steps;

[0097] The method for generating a user electricity request includes:

[0098] A basic request with two request bits is constructed, and the two request bits are marked as the first request bit and the second request bit in sequence; a basic request refers to a meaningless request template without actual data, and a request bit is the smallest unit constituting a basic request, and serves as the data import location for subsequent user power requests, ensuring the rationality and accuracy of user power requests;

[0099] Query the area number corresponding to the electricity consumption area through the administrative area system, and import the area number into the first request bit of the basic request to form a number request bit;

[0100] Mark the first digit group and the second digit group of the power consumption interval one by one, connect the first digit group and the second digit group end to end to form an interval group, and import the interval group into the second request bit of the basic request to form an interval request bit; the first digit group and the second digit group refer to the digits of the electronic date and the digits of the electronic time in the power consumption interval, so as to accurately represent the power consumption interval;

[0101] The basic request with the number request bit and the interval request bit is recorded as the user power request.

[0102] It should be noted that when a user's electricity request is constructed, it can provide a precondition for judging the safety of electricity use in the residence of the elderly living alone, and then conduct subsequent collection and analysis of electricity use behavior.

[0103] S2: Filter out the user power supply plan corresponding to the user's power consumption request from the power database, identify the power supply status of the user power supply plan, and determine whether it meets the power consumption behavior safety analysis requirements; if it meets the power consumption behavior safety analysis requirements, execute S3; if it does not meet the power consumption behavior safety analysis requirements, repeat S1-S2;

[0104] The user power supply plan refers to the specific power supply plan that the power supply management department pre-stores in the power database based on the actual power supply demand to indicate whether different areas are powered on. It can be used as the basis for comparative analysis with the user's power demand request and as the basis for subsequent safety analysis of whether the power consumption behavior is met;

[0105] After obtaining the user's power supply plan, the power supply status can be identified from the user's power supply plan, and according to the different power supply status, it is determined whether the power consumption behavior safety analysis requirements are met, and then whether to execute the subsequent steps;

[0106] The power supply status is used to indicate whether the power supply is provided in the user's power supply plan. Different power supply statuses represent different power supply situations. The power supply status includes continuous power supply status and suspended power supply status. The continuous power supply status indicates that the corresponding power supply situation in the user's power supply plan is continuous power supply, and there will be no power outage. The suspended power supply status indicates that the corresponding power supply situation in the user's power supply plan is suspended power supply, and there will be power outage.

[0107] The identification method of the continuous power supply state and the suspended power supply state includes:

[0108] Taking the area number in the number request bit as the screening condition, i power supply plans to be identified that are consistent with the area number in the number request bit are screened out from the power database;

[0109] Identify the power supply dates and power supply times of i power supply plans to be identified one by one through natural language processing technology, and obtain i power supply dates and i power supply times;

[0110] The power supply plan to be identified, in which the power supply date and power supply time are consistent with the first digit group and the second digit group in the interval request bit, is recorded as the user power supply plan;

[0111] The power supply status remark column of the user's power supply plan is queried through the power supply management system, and the digital combination of the remark in the power supply status remark column is identified; the power supply status remark column is used to truly indicate whether the power supply indicated by the user's power supply plan is available;

[0112] When the digital combination is 1-1, it means that the corresponding power supply plan in the user's power supply plan is continuous power supply, and there will be no power outage, so the power supply state is identified as a continuous power supply state;

[0113] When the digital combination is 1-0, it means that the corresponding power supply plan in the user's power supply plan is to suspend power supply, and a power outage will occur, and the power supply state is identified as a suspended power supply state.

[0114] When a continuous power supply state is identified, it means that when the residence of the elderly living alone conducts a safety analysis of electricity usage behavior, there will be no power outage in the area of ​​the residence, which can meet the needs of the safety analysis of electricity usage behavior, and it can be ensured that the electricity usage behavior can be effectively and accurately collected. On the contrary, when a suspended power supply state is identified, it means that when the residence of the elderly living alone conducts a safety analysis of electricity usage behavior, there will be a power outage in the area of ​​the residence, which cannot meet the needs of the safety analysis of electricity usage behavior, and it cannot be ensured that the electricity usage behavior can be effectively and accurately collected. Therefore, it can be determined whether to perform subsequent steps according to different power supply states;

[0115] The methods for determining whether the safety analysis requirements of electricity consumption behavior are met include:

[0116] When the power supply status is continuous power supply, it means that the area where the elderly person living alone lives will have continuous power supply, which can meet the relevant needs of power consumption behavior collection and safety analysis, and it is determined that the power consumption behavior safety analysis needs are met;

[0117] When the power supply status is suspended, it means that the power supply in the area where the elderly person living alone lives will be suspended, and the relevant needs of power consumption behavior collection and safety analysis cannot be met, so it is determined that the power consumption behavior safety analysis needs are not met.

[0118] When the power usage behavior safety analysis requirements are not met, S1-S2 needs to be repeated until the power usage behavior safety analysis requirements are met before entering the next step, thereby ensuring the effectiveness and accuracy of subsequent power usage behavior collection and safety analysis.

[0119] S3: Obtain comprehensive behavior data of the user end within the power consumption interval, the comprehensive behavior data including time period type, unit power consumption excess value, abnormal power consumption frequency and abnormal power consumption duration;

[0120] Comprehensive behavior data refers to data that can be used by the user end to represent the power consumption of various electrical equipment in the residence of elderly people living alone within the power consumption range, so as to serve as the data basis for subsequent analysis of power safety;

[0121] Comprehensive behavior data includes time period type, unit power consumption excess value, abnormal power consumption frequency and abnormal power consumption duration;

[0122] The time period type refers to the specific type of time period when the user collects relevant power consumption behaviors within the power consumption interval, so as to effectively distinguish different time periods of a day. Since the power consumption behaviors of users in different time periods are uneven, different power consumption time periods can be effectively identified by distinguishing the time period types. The time period types include peak power consumption period, low peak power consumption period and flat peak power consumption period.

[0123] Exemplarily, when the time period when electricity consumption is between 6:00 and 8:00 or between 17:00 and 19:00, the time period type is a peak electricity consumption period; when the time period when electricity consumption is between 21:00 and 6:00, the time period type is a low-peak electricity consumption period; when the time period when electricity consumption is between 8:00 and 17:00 or between 19:00 and 21:00, the time period type is an off-peak electricity consumption period.

[0124] The unit power consumption excess value refers to the extent to which the power consumption of the user end exceeds the standard power consumption per unit time in the power consumption interval, which can be used to represent the power consumption of the user end per unit time. The larger the unit power consumption excess value, the greater the extent to which the user end exceeds the standard power consumption per unit time, the more serious the abnormal power consumption in the residence of the elderly living alone, the greater the probability of power accidents, and the lower the safety of the elderly living alone;

[0125] Methods for obtaining the unit electricity consumption excess value include:

[0126] In the time period corresponding to the power consumption interval, w adjacent collection time periods are marked based on the preset power consumption duration, and the w collection time periods are numbered in ascending order; the power consumption duration is used as the basis for dividing the corresponding duration of the power consumption interval, that is, the duration of the power consumption interval can be equally divided to ensure the consistency of the duration of each collection time period;

[0127] Taking the collection period corresponding to the first number as the starting point, S collection periods are selected as valid periods in a manner of one number interval; this can ensure that there will be no continuous adhesion between two adjacent valid periods, and at the same time reduce the number of valid periods, thereby reducing the subsequent data collection and calculation burden;

[0128] The starting time and the ending time of the S valid time periods are marked one by one, and the power consumption corresponding to the S starting time and the S ending time is counted one by one by the electric energy meter to obtain the power consumption of the S starting points and the S ending points;

[0129] Subtract the power consumption of the S endpoints from the power consumption of the S starting points one by one, and compare them with the duration of the effective period to obtain S sub-power consumption values;

[0130] The expression of the sub-power consumption value is:

[0131]

[0132] In the formula, YD zs is the electricity consumption value of the sth child, TD zds is the power consumption of the sth terminal, YD qds is the electricity consumption at the sth starting point, SC yx is the duration of the effective period;

[0133] The sub-power consumption value greater than the standard power consumption value is recorded as the sub-excess value, and y sub-excess values ​​are obtained. The y sub-excess values ​​are accumulated and averaged to obtain the unit power consumption excess value. The standard power consumption value is a numerical basis for distinguishing whether the sub-power consumption value is a sub-excess value, which can accurately distinguish the size of the sub-power consumption value, thereby accurately identifying the sub-excess value.

[0134] The expression of unit power consumption excess value is:

[0135]

[0136] In the formula, YD ce is the excess value of unit electricity consumption, YD za is the power consumption value of the ath child.

[0137] The abnormal power consumption frequency refers to the number of times abnormal current occurs per unit time at the user end in the power consumption interval, which can be used to indicate the abnormal current phenomenon at the user end per unit time. The greater the abnormal power consumption frequency, the more times abnormal current occurs per unit time at the user end, the more serious the abnormal power consumption phenomenon in the residence of the elderly living alone, the greater the probability of power accidents, and the lower the safety of the elderly living alone;

[0138] Methods for obtaining abnormal power consumption frequency include:

[0139] Randomly mark q non-adjacent moments in S valid time periods to obtain q detection moments;

[0140] The real-time current at q detection moments is detected by the current sensor respectively, and q detection current values ​​are obtained, and the detection current value greater than the standard current value is recorded as an abnormal current value; the standard current value is a numerical basis for distinguishing whether the detection current value is an abnormal current value, and the magnitude of the detection current value can be accurately distinguished, so as to accurately identify the abnormal current value;

[0141] Counting the number of abnormal current values ​​in the S effective time periods one by one, and comparing the number of abnormal current values ​​with the duration of the effective time period to obtain S sub-frequencies;

[0142] The expression of sub-frequency is:

[0143]

[0144] Where PL zs is the sth sub-frequency, SL ycs is the number of abnormal current values ​​of s, SC yx is the duration of the effective period;

[0145] Remove the maximum and minimum values ​​of the sub-frequencies, accumulate the remaining s-2 sub-frequencies and calculate the average to obtain the abnormal power consumption frequency;

[0146] The expression of abnormal power consumption frequency is:

[0147]

[0148] Where PL yc is the abnormal power consumption frequency, PL zb is the bth sub-frequency.

[0149] The abnormal power consumption duration refers to the length of time corresponding to abnormal power consumption at the user end in the power consumption interval, which can be used to indicate the duration of abnormal power consumption at the user end. The longer the abnormal power consumption duration, the longer the time corresponding to abnormal power consumption at the user end, the more serious the abnormal power consumption phenomenon in the residence of the elderly living alone, the greater the probability of power accidents, and the lower the safety of the elderly living alone;

[0150] Methods for obtaining abnormal power consumption duration include:

[0151] During the period corresponding to the power consumption interval, the real-time current at the user end is detected by the current sensor, and the moment when the real-time current reaches the preset current lower limit for the first time is marked as the zero time; the preset current lower limit is the minimum value of the current when it is judged as abnormal power consumption, so as to ensure the accuracy of the zero time;

[0152] After the zero point, mark the moment when the real-time current is less than the preset current lower limit for the first time, record it as the termination moment, and record the duration from the zero point to the termination moment as the zero termination duration, and obtain v zero termination durations;

[0153] The zero-end duration that is greater than the safe power usage duration is recorded as an abnormal duration, and x abnormal durations are obtained. After the x abnormal durations are accumulated, the abnormal power usage duration is obtained. The safe power usage duration is a numerical basis for distinguishing whether the zero-end duration is an abnormal duration, and the size of the zero-end duration can be accurately distinguished, thereby accurately identifying the abnormal duration.

[0154] The expression of abnormal power consumption duration is:

[0155]

[0156] In the formula, YD yc SC is the abnormal power consumption duration. lzc is the cth abnormal duration.

[0157] S4: Input the collected comprehensive behavior data into the machine learning model trained in advance, predict the power consumption status value, and determine whether to issue an abnormal power consumption warning prompt; if an abnormal power consumption warning prompt is issued, execute S5; if no abnormal power consumption warning prompt is issued, repeat S3-S4;

[0158] After obtaining the time period type, unit power consumption excess value, abnormal power consumption frequency and abnormal power consumption duration, the time period type, unit power consumption excess value, abnormal power consumption frequency and abnormal power consumption duration can be input into a pre-trained machine learning model, so that the machine learning model can predict the corresponding power consumption status value according to the input time period type, unit power consumption excess value, abnormal power consumption frequency and abnormal power consumption duration, thereby providing a basis for whether to issue an early warning prompt in the future;

[0159] The machine learning model is used to predict the corresponding power consumption status value based on the collected time period type, unit power consumption excess value, abnormal power consumption frequency and abnormal power consumption duration, so as to provide predictive data for the power consumption safety of elderly people living alone;

[0160] The power usage status value is used to indicate the safety of the power usage behavior of elderly people living alone. The power usage status value includes normal power usage status and abnormal power usage status. The power usage status value is obtained by collecting a large number of time period types, unit power consumption excess values, abnormal power usage frequencies and abnormal power usage durations in normal power usage status and abnormal power usage status.

[0161] Methods for training machine learning models include:

[0162] Collect multiple groups of comprehensive behavior data in normal and abnormal power consumption states in advance;

[0163] Assign values ​​to the time period types in sequence; for example, assign the peak power consumption period to A, the low-peak power consumption period to B, and the flat-peak power consumption period to C;

[0164] The comprehensive behavior data is converted into a corresponding set of feature vectors, and the feature vectors are used as the input of the machine learning model. The power consumption state value is converted into a label corresponding to the comprehensive behavior data, the normal power consumption state is converted into 0, and the abnormal power consumption state is converted into 1. The power consumption state value corresponding to each set of comprehensive behavior data is used as the output of the machine learning model. The power consumption state value is used as the prediction target, and the sum of the prediction errors of all training data is minimized as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence and the training is stopped;

[0165] Exemplarily, the machine learning model is any one of a CNN neural network model or AlexaNet.

[0166] After the collected comprehensive behavior data is input into the machine learning model, the corresponding power consumption status value can be predicted, thereby determining whether to issue an abnormal power consumption warning prompt;

[0167] The determination method of whether to issue an abnormal power consumption warning prompt includes:

[0168] When the output of the machine learning model is 0, the power consumption state value at this time is normal power consumption state, and there is no abnormality in the power consumption behavior of the elderly living alone, so it is determined that no abnormal power consumption warning prompt will be issued;

[0169] When the output of the machine learning model is 1, the power usage status value at this time is an abnormal power usage status, and the power usage behavior of the elderly living alone is abnormal, and it is determined to issue an abnormal power usage warning prompt.

[0170] S5: Identify abnormal behaviors from the comprehensive behavior data, formulate security warning information, and send the corresponding security warning information to the warning terminal;

[0171] Abnormal behavior refers to the specific parameters in the comprehensive behavior data that can cause the power consumption status value to be an abnormal power consumption status, and serves as the basis for the subsequent formulation of the safety warning level;

[0172] Methods for identifying abnormal behavior include:

[0173] The unit power consumption excess value is compared with the standard power consumption excess value; the standard power consumption excess value refers to the minimum value of the unit power consumption excess value that can cause the power consumption state value to be an abnormal power consumption state, which can limit the lower limit of the unit power consumption excess value;

[0174] When the unit electricity consumption excess value is greater than the standard electricity consumption excess value, the unit electricity consumption of the residence of the elderly living alone is higher than the safe value, and the unit electricity consumption excess value is recorded as abnormal behavior;

[0175] Compare the abnormal power consumption frequency with the standard power consumption frequency; the standard power consumption frequency refers to the minimum value of the abnormal power consumption frequency that can cause the power consumption state value to be an abnormal power consumption state, which can limit the lower limit of the abnormal power consumption frequency;

[0176] When the abnormal power consumption frequency is greater than the standard power consumption frequency, the abnormal frequency value of the residence of the elderly living alone is higher than the safety value, and the abnormal power consumption frequency is recorded as abnormal behavior;

[0177] The abnormal power consumption duration is compared with the standard power consumption duration; the standard power consumption duration refers to the minimum value of the abnormal power consumption duration that can cause the power consumption state value to be an abnormal power consumption state, which can limit the lower limit of the abnormal power consumption duration;

[0178] When the abnormal electricity usage duration is greater than the standard electricity usage duration, the abnormal electricity usage duration of the elderly living alone is higher than the safety value, and the abnormal electricity usage duration is recorded as abnormal behavior.

[0179] Safety warning information is used to specifically indicate the safety level represented by abnormal behavior parameters. Different safety warning information represents different safety levels for elderly people living alone.

[0180] Safety warning information includes mild warning information and severe warning information; mild warning information indicates that the number of abnormal behaviors is small and the safety hazard is low, while severe warning information indicates that the number of abnormal behaviors is large and the safety hazard is high;

[0181] The method for formulating mild warning information and severe warning information includes:

[0182] The number of abnormal behaviors is counted. When the number of abnormal behaviors is 1, the number of abnormal electricity consumption behaviors in the comprehensive behavior data is small, and a mild warning information is formulated;

[0183] When the number of abnormal behaviors is 2 or 3, the number of abnormal electricity usage behaviors in the comprehensive behavior data is large, and a severe warning information is formulated.

[0184] When mild warning information and severe warning information are formulated, the security cloud platform will promptly send the security warning information to the warning terminal corresponding to the user terminal, so that the warning terminal can rush to the residence of the elderly living alone for safety confirmation in time after receiving the security warning information, thereby preventing more serious electricity safety accidents. The warning terminal can be the warning platform of the management department corresponding to the residence of the elderly living alone, or it can be the smart device of the guardian of the elderly living alone, as long as it can receive safety warning information.

[0185] In this embodiment, by receiving the power consumption area and power consumption interval of the user end, and combining the power consumption area and the power consumption interval, a user power consumption request is generated, the user power supply plan corresponding to the user power consumption request is screened out from the power database, the power supply status of the user power supply plan is identified, and it is determined whether the power consumption behavior safety analysis requirements are met, the comprehensive behavior data of the user end in the power consumption interval is obtained, the collected comprehensive behavior data is input into the machine learning model trained in advance, the power consumption status value is predicted, and it is determined whether to issue an abnormal power consumption warning prompt, abnormal behavior is identified from the comprehensive behavior data, safety warning information is formulated, and the corresponding safety warning information is sent to the warning end; Compared with the prior art, by forming the user power consumption request with the power consumption area and the power consumption interval, it can A preliminary verification is conducted on whether the residence of an elderly person living alone meets the safety analysis requirements for electricity usage, thereby avoiding inaccurate safety analysis caused by abnormal situations in the power supply plan such as power outages. At the same time, it also raises the threshold for safety analysis, reduces the collection and analysis of a large number of electricity usage behaviors that do not meet the safety analysis threshold, and reduces the workload of safety analysis. At the same time, combined with a machine learning model, it can accurately predict the electricity usage status value of the residence of an elderly person living alone based on comprehensive behavioral data, and formulate safety warning information in a timely and targeted manner when there are safety hazards in electricity use, thereby providing relevant departments and personnel with secondary confirmation of the safety of the elderly living alone, thereby avoiding more serious electricity safety accidents and effectively improving the electricity safety of the elderly living alone.

[0186] Example 2: Please refer to Figure 2 As shown, the part not described in detail in this embodiment is described in the first embodiment, and a safety analysis system for elderly people living alone based on electricity usage behavior is provided, which is applied to a security cloud platform and is used to implement a safety analysis method for elderly people living alone based on electricity usage behavior, including a request generation module, a demand determination module, a data acquisition module, a model prediction module and an early warning information module, wherein each module is connected via a wired or wireless network;

[0187] A request generation module, used to receive the power consumption area and power consumption interval of the user end, and combine the power consumption area and the power consumption interval to generate a user power consumption request;

[0188] The demand determination module is used to filter out the user power supply plan corresponding to the user's power consumption request from the power database, identify the power supply status of the user's power supply plan, and determine whether the power consumption behavior safety analysis requirements are met;

[0189] The data collection module is used to obtain the comprehensive behavior data of the user end within the power consumption interval, and the comprehensive behavior data includes the time period type, unit power consumption excess value, abnormal power consumption frequency and abnormal power consumption duration;

[0190] The model prediction module is used to input the collected comprehensive behavior data into the machine learning model trained in advance, predict the power consumption status value, and determine whether to issue an abnormal power consumption warning prompt. The power consumption status value includes normal power consumption status and abnormal power consumption status;

[0191] The early warning information module is used to identify abnormal behaviors from comprehensive behavior data, formulate safety early warning information, and send corresponding safety early warning information to the early warning terminal.

[0192] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A safety analysis method for elderly people living alone based on electricity usage behavior, applied to a security cloud platform, characterized in that: include: S1: Receive the power consumption area and power consumption interval of the user, and combine the power consumption area and power consumption interval to generate the user's power consumption request; S2: Filter out the user power supply plan corresponding to the user's power consumption request from the power database, identify the power supply status of the user power supply plan, and determine whether it meets the power consumption behavior safety analysis requirements; if it meets the power consumption behavior safety analysis requirements, execute S3; If the power consumption behavior safety analysis requirements are not met, repeat S1-S2; S3: Obtain comprehensive behavior data of the user end within the power consumption interval, the comprehensive behavior data including time period type, unit power consumption excess value, abnormal power consumption frequency and abnormal power consumption duration; S4: Input the collected comprehensive behavior data into the machine learning model trained in advance, predict the power consumption status value, and determine whether to issue an abnormal power consumption warning prompt. The power consumption status value includes a normal power consumption status and an abnormal power consumption status. If an abnormal power consumption warning prompt is issued, execute S5; if no abnormal power consumption warning prompt is issued, repeat S3-S4. S5: Identify abnormal behaviors from the comprehensive behavior data, formulate safety warning information, and send corresponding safety warning information to the warning terminal.

2. A safety analysis method for elderly people living alone based on electricity usage behavior according to claim 1, characterized in that: The method for obtaining the electricity consumption area and the electricity consumption interval includes: Mark the location of the residence of the elderly living alone on the electronic map and record it as the user point; Taking the user point as the base point, four extension lines are drawn in the horizontal and vertical directions. The road that first intersects the four extension lines on the electronic map is recorded as a regional road, and the area enclosed by the four regional roads is recorded as a power consumption area; The electronic calendar of the electricity consumption area is queried through the calendar system, and the electronic date and electronic time in the electronic calendar are identified through natural language processing technology; A basic interval having a first interval position and a second interval position is constructed, and the electronic date and the electronic time are respectively introduced into the first interval position and the second interval position of the basic interval to obtain a power consumption interval.

3. A safety analysis method for elderly people living alone based on electricity usage behavior according to claim 2, characterized in that: The method for generating the user's electricity consumption request comprises: Constructing a basic request with two request bits, and marking the two request bits as a first request bit and a second request bit in sequence; Query the area number corresponding to the electricity consumption area through the administrative area system, and import the area number into the first request bit of the basic request to form a number request bit; Mark the first digital group and the second digital group of the power consumption interval one by one, connect the first digital group and the second digital group end to end to form an interval group, and import the interval group into the second request bit of the basic request to form an interval request bit; The basic request with the number request bit and the interval request bit is recorded as the user power request.

4. A safety analysis method for elderly people living alone based on electricity usage behavior according to claim 3, characterized in that: The power supply state includes a continuous power supply state and a suspended power supply state; The identification method of the continuous power supply state and the suspended power supply state includes: Taking the area number in the number request bit as the screening condition, i power supply plans to be identified that are consistent with the area number in the number request bit are screened out from the power database; Identify the power supply dates and power supply times of i power supply plans to be identified one by one through natural language processing technology, and obtain i power supply dates and i power supply times; The power supply plan to be identified, in which the power supply date and power supply time are consistent with the first digit group and the second digit group in the interval request bit, is recorded as the user power supply plan; The power supply status remarks column of the user's power supply plan is queried through the power supply management system, and the digital combination of remarks in the power supply status remarks column is identified; When the digital combination is 1-1, the power supply state is identified as a continuous power supply state; When the digital combination is 1-0, the power supply state is identified as a power supply suspension state.

5. A safety analysis method for elderly people living alone based on electricity usage behavior according to claim 4, characterized in that: The method for determining whether the power consumption behavior safety analysis requirement is met includes: When the power supply state is a continuous power supply state, it is determined that the power consumption behavior safety analysis requirements are met; When the power supply state is a suspended power supply state, it is determined that the power consumption behavior safety analysis requirement is not met.

6. A safety analysis method for elderly people living alone based on electricity usage behavior according to claim 5, characterized in that: The time period types include peak power consumption period, low-peak power consumption period and flat-peak power consumption period; Methods for obtaining the unit electricity consumption excess value include: In the time period corresponding to the power consumption interval, w collection time periods that are adjacent to each other are marked based on the preset power consumption duration, and the w collection time periods are numbered in ascending order; Taking the collection period corresponding to the first number as the starting point, S collection periods are selected as valid periods in a manner of one number interval; Mark the starting time and the ending time of the S valid time periods one by one, and use the electric energy meter to count the power consumption corresponding to the S starting time and the S ending time one by one, so as to obtain the power consumption of s starting points and s ending points; Subtract the power consumption of the s end points from the power consumption of the s starting points one by one, and compare them with the duration of the effective period to obtain s sub-power consumption values; The expression of the sub-power consumption value is: In the formula, YD zs is the electricity consumption value of the sth child, TD zds is the power consumption of the sth terminal, YD qds is the electricity consumption at the sth starting point, SC yx is the duration of the effective period; The sub-power consumption value greater than the standard power consumption value is recorded as a sub-excess value, y sub-excess values ​​are obtained, and the y sub-excess values ​​are accumulated and averaged to obtain a unit power consumption excess value; The expression of unit power consumption excess value is: In the formula, YD ce is the excess value of unit electricity consumption, YD za is the power consumption value of the ath child.

7. A safety analysis method for elderly people living alone based on electricity usage behavior according to claim 6, characterized in that: The method for obtaining the abnormal power consumption frequency includes: Randomly mark q non-adjacent moments in s valid time periods to obtain q detection moments; The current sensor detects the real-time current at q detection moments respectively, obtains q detection current values, and records the detection current value greater than the standard current value as an abnormal current value; Counting the number of abnormal current values ​​in the S effective time periods one by one, and comparing the number of abnormal current values ​​with the duration of the effective time period to obtain S sub-frequencies; The expression of sub-frequency is: Where PL zs is the sth sub-frequency, SL ycs is the number of abnormal current values ​​of s, SC yx is the duration of the effective period; Remove the maximum and minimum values ​​of the sub-frequencies, accumulate the remaining s-2 sub-frequencies and calculate the average to obtain the abnormal power consumption frequency; The expression of abnormal power consumption frequency is: Where PL yc is the abnormal power consumption frequency, PL zb is the bth sub-frequency.

8. A safety analysis method for elderly people living alone based on electricity usage behavior according to claim 7, characterized in that: The method for obtaining the abnormal power consumption duration includes: During the period corresponding to the power consumption interval, the real-time current at the user end is detected by the current sensor, and the moment when the real-time current reaches the preset current lower limit for the first time is marked as the zero time; After the zero point, mark the moment when the real-time current is less than the preset current lower limit for the first time, record it as the termination moment, and record the duration from the zero point to the termination moment as the zero termination duration, and obtain v zero termination durations; The zero-end duration greater than the safe power consumption duration is recorded as the abnormal duration, x abnormal durations are obtained, and the abnormal power consumption duration is obtained by accumulating the x abnormal durations; The expression of abnormal power consumption duration is: In the formula, YD yc SC is the abnormal power consumption duration. lzc is the cth abnormal duration.

9. A safety analysis method for elderly people living alone based on electricity usage behavior according to claim 8, characterized in that: The training method of the machine learning model includes: Collect multiple groups of comprehensive behavior data in normal and abnormal power consumption states in advance; Assign the peak power consumption period to A, the low-peak power consumption period to B, and the flat-peak power consumption period to C; The comprehensive behavior data is converted into a corresponding set of feature vectors, and the feature vectors are used as the input of the machine learning model. The power consumption state value is converted into a label corresponding to the comprehensive behavior data, the normal power consumption state is converted into 0, and the abnormal power consumption state is converted into 1. The power consumption state value corresponding to each set of comprehensive behavior data is used as the output of the machine learning model. The power consumption state value is used as the prediction target, and the sum of the prediction errors of all training data is minimized as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence and the training is stopped; The determination method of whether to issue an abnormal power consumption warning prompt includes: When the output of the machine learning model is 0, it is determined that no abnormal power consumption warning prompt will be issued; When the output of the machine learning model is 1, it is determined that an abnormal power consumption warning prompt is issued.

10. A safety analysis method for elderly people living alone based on electricity usage behavior according to claim 9, characterized in that: The abnormal behavior identification method includes: The unit power consumption excess value is compared with the standard power consumption excess value. When the unit power consumption excess value is greater than the standard power consumption excess value, the unit power consumption excess value is recorded as an abnormal behavior. The abnormal power consumption frequency is compared with the standard power consumption frequency. When the abnormal power consumption frequency is greater than the standard power consumption frequency, the abnormal power consumption frequency is recorded as an abnormal behavior. Compare the abnormal power usage duration with the standard power usage duration. When the abnormal power usage duration is greater than the standard power usage duration, record the abnormal power usage duration as abnormal behavior. Safety warning information includes mild warning information and severe warning information; The method for formulating mild warning information and severe warning information includes: Count the number of abnormal behaviors, and when the number of abnormal behaviors reaches 1, generate a mild warning message; When the number of abnormal behaviors is 2 or 3, a severe warning message is formulated.

Images