Multi-terminal simultaneous connection heat control system based on NB-IoT on-off control valve and application thereof in intelligent analysis of vacant house

Abstract

The application provides a multi-end joint heat control system based on an NB-IoT on-off control valve and application thereof in intelligent analysis of vacant houses, and relates to the technical field of heat control. The system comprises a dual-mode sensing terminal, an edge computing gateway, a cloud decision platform and a hierarchical access interface. The dual-mode sensing terminal collects data through open and hidden dual modes and transmits the data in a classified and encrypted manner. The edge computing gateway processes data through double threads and extracts regular / abnormal features. The multi-dimensional collaborative detection engine of the cloud decision platform performs five-dimensional parallel analysis, calculates the contribution weight of the standardized abnormal score and the dynamic balance, weights the vacant state confidence, and generates a valve closing or verification instruction accordingly. The hierarchical access interface is used for realizing multi-role permission isolation. The system can accurately identify vacant houses and heat disguising behaviors, reduces heat loss and improves multi-end collaborative efficiency.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of centralized heating heat pipe control, and particularly relates to a multi-end joint heating control system based on an NB-IoT on-off control valve and application thereof in intelligent analysis of vacant houses. BACKGROUND

[0002] In the field of centralized heating in northern China, the energy utilization efficiency of the heating system, the heat loss control of vacant houses, and the standardization of heating behavior have always been the core problems of the industry development. With the acceleration of urbanization, the vacancy rate of urban residential buildings shows a stage-wise upward trend. A large number of vacant houses are in a long-term state of valve opening, resulting in increased heat loss of the heating pipe network. This not only causes serious waste of energy, but also significantly increases the operating cost and carbon emission pressure of the heat supply enterprise.

[0003] In order to solve the problem of heat loss of vacant houses, the existing technology mostly adopts the mode of manual inspection and single sensor monitoring. On the one hand, it relies on the personnel of the property or heat supply company to check regularly, which is not only inefficient, but also easily affected by factors such as privacy restrictions and human intervention, resulting in low accuracy. On the other hand, some systems introduce temperature sensors or ordinary on-off valves to monitor the water supply temperature and valve opening and closing state to determine vacancy. However, this kind of solution has significant defects: first, the sampling frequency of the sensor is low and the monitoring dimension is single, which cannot capture the short-term opening of the valve to create a normal heating illusion, the pseudo heating behavior of maintaining surface temperature without substantial heat consumption, etc. Second, the data transmission and security are insufficient. The existing systems mostly use GPRS or ordinary NB-IoT transmission, and do not classify and encrypt the normal heating data and sensitive operation data, so there is a risk of data leakage or tampering. Third, there is a lack of multi-end collaboration capability. The data between the heat supply company, the property, and the residents is fragmented, and the heat supply company cannot obtain cross-system data such as the occupancy status registered by the property, the water and electricity meter readings of the water and electricity supply enterprises, etc., resulting in low confidence in vacancy determination relying on local sensor data. The property lacks an efficient channel for receiving and feedback of verification instructions, and the manual verification process is lagging. The residents cannot query their own heating data and vacancy determination basis, which easily leads to disputes over false valve closure.

[0004] In addition, the response mechanism of the existing vacant house determination system is generally rigid. Most systems only trigger the valve closure instruction according to a single condition such as temperature threshold or valve opening time, and do not establish a multi-dimensional confidence evaluation model, resulting in frequent false positives such as short-term business trips being determined as vacant or false negatives such as pseudo heating residents not being identified. At the same time, the permission management is chaotic, and some systems do not divide the access rights of the heat supply company, the property, and the residents, resulting in disordered flow of sensitive data, which not only violates the data security regulations, but also damages the rights and interests of the residents.

[0005] In summary, the current central heating field urgently needs a heat control system with high-precision heat use camouflage behavior detection, multi-dimensional vacancy judgment, safe data transmission and hierarchical multi-end collaborative ability to solve the problems of inaccurate vacancy house identification, high heat loss, inefficient multi-end management and data security risks, and promote the transformation of the heating system to intelligence, energy saving and safety. SUMMARY

[0006] The purpose of the present application is to provide a multi-end joint heat control system based on NB-IoT on-off control valve and its application to overcome the defects of existing central heating systems in vacancy house monitoring and control.

[0007] To solve the above technical problems, the technical scheme of the present application is:

[0008] A multi-end joint heat control system based on NB-IoT on-off control valve, comprising a dual-mode perception terminal, an edge computing gateway, a cloud decision platform and a hierarchical access interface connected in sequence and forming a closed loop control;

[0009] The dual-mode perception terminal comprises an NB-IoT on-off control valve, a dual-mode sensor group and an encrypted communication unit. The NB-IoT on-off control valve is the core execution component, which is configured with an electric actuator and a state monitoring module. The electric actuator is used to receive control instructions and execute heating pipeline on-off operation, and the state monitoring module is used to collect valve state data. The dual-mode sensor group includes a public detection module and a hidden detection module, wherein the public detection module is used to collect regular sensing data, and the hidden detection module is used to collect hidden sensing data. The encrypted communication unit classifies and encrypts the valve state data, regular sensing data and hidden sensing data, and then sends them to the edge computing gateway through the NB-IoT network, and parses the received control instructions to drive the NB-IoT on-off control valve to act. The edge computing gateway is used to realize the communication connection between the dual-mode perception terminal and the cloud decision platform, which is configured with a first data processing thread and a second data processing thread for parallel processing. The first data processing thread receives and parses the basic information in the regular sensing data and the valve state data, generates a visual heat curve and synchronizes it to the hierarchical access interface, while extracting the regular feature parameters of the corresponding target resident and sending them to the cloud decision platform. The second data processing thread receives and decrypts the operation behavior information in the hidden sensing data and the valve state data, extracts the abnormal feature parameters of the corresponding target resident and sends them to the cloud decision platform;

[0010] The cloud decision platform comprises a multi-dimensional cooperative detection engine and a dynamic response module; the multi-dimensional cooperative detection engine is used for parallel analysis of five dimensions of spatial correlation, time sequence, operation behavior, environmental influence and multi-source data cross verification of the regular characteristic parameters and abnormal characteristic parameters of the target household, for each dimension, extracts the single-dimensional abnormal characteristic under the dimension pointing to vacancy, then calculates the standardized abnormal score reflecting the abnormal degree based on the single-dimensional abnormal characteristic, and dynamically determines the contribution weight of each dimension according to the degree of deviation of the single-dimensional abnormal characteristic from the normal threshold, and then calculates the weighted score of each dimension by multiplying the dimension standardized abnormal score and the corresponding contribution weight; finally, the weighted scores of the five dimensions are summed to obtain the vacancy state confidence of the target household; and in the operation behavior dimension analysis, whether there is heat camouflage behavior is determined by comparing the consistency of the hidden sensing data and the regular sensing data, if there is heat camouflage behavior, the behavior characteristic is taken as the key abnormal characteristic parameter of the dimension, and the contribution weight is additionally increased by 20%-50% based on the basic value; the dynamic response module generates corresponding control instructions according to the vacancy state confidence of the target household, and the control instructions and the heat camouflage behavior detection result are delivered through the edge computing gateway, the control instructions are used to drive the NB-IoT on-off control valve of the target household to act; the edge computing gateway triggers the high-frequency acquisition instruction of the dual-mode perception terminal of the target household after receiving the heat camouflage behavior detection result if there is heat camouflage behavior; at the same time, the decision result including the vacancy state confidence of the target household, the control instruction type, the heat camouflage behavior detection result and the basis of triggering the control instruction is synchronized to the hierarchical access interface; the basis of triggering the control instruction is the first three abnormal characteristics and their corresponding data values sorted by contribution weight;

[0011] Further, the valve state data is a characteristic parameter reflecting the running state and operation behavior of the valve; the regular sensing data is a temperature parameter reflecting the basic running state of the heating system, and the hidden sensing data is a high-frequency characteristic parameter reflecting the subtle heat behavior; the valve state data comprises switch state identifier, cumulative opening time, single operation duration, operation trigger source, real-time opening degree value and fault diagnosis code; the regular sensing data comprises water supply temperature and return water temperature; the hidden sensing data comprises flow micro-pulse signal and pipe radial temperature gradient.

[0012] Further, the classified encryption mechanism of the encryption communication unit is specifically: for the basic running state information and the conventional sensing data in the valve state data, an AES-128 symmetric encryption algorithm is used, the last 16 bits of the device unique identifier of the dual-mode sensing terminal are used as a fixed key, and the conventional data frame is generated after encryption, and the frame header marks the data type identification bit 0x01; for the sensitive operation information and the concealed sensing data in the valve state data, an AES-256 symmetric encryption algorithm is used, the first 16 bits of the device unique identifier of the dual-mode sensing terminal and the real-time time stamp are spliced as a dynamic key, and the concealed data frame is generated after encryption, and the frame header marks the data type identification bit 0x02; the encrypted data frame adopts a composite transmission structure of the conventional data frame and the concealed data frame, the concealed data frame is embedded in the check bit field of the conventional data frame, the two types of data are distinguished and analyzed through the frame header identification bit, and all the encrypted data is attached with an 8-bit CRC check code for integrity verification.

[0013] Further, the concealed detection module includes a micro turbine flow sensor, a three-way distributed temperature sensor and an event-triggered wake-up unit; the micro turbine flow sensor is used to obtain the circulating water flow in the heating pipeline of the target household, the collected data is processed by the encryption communication unit, and is embedded in the check bit field of the conventional data frame as the concealed data frame, and the frame header identification bit 0x02 is used for camouflage transmission; the three-way distributed temperature sensor is used to obtain the pipeline radial temperature difference to identify the false temperature change caused by the short opening of the valve; the event-triggered wake-up unit is used to automatically activate the micro turbine flow sensor and the three-way distributed temperature sensor when detecting that the valve is opened for less than 5 minutes at 2-4 a.m., and continuously collect data for 1 hour, and the data transmission interval during the collection period is shortened to 1 / 5 of the regular transmission interval.

[0014] Further, the corresponding single-dimension abnormal feature in the parallel analysis of the five dimensions of the multi-dimension cooperative detection engine is:

[0015] The spatial correlation dimension: taking the heat usage mode data of the neighborhood households of the target household as the processing object, the spatial anomaly is identified by calculating the heat usage mode similarity between the target household and more than 90% of the neighborhood households, and the single-dimension abnormal feature output is the neighborhood heat usage mode deviation; the neighborhood households are households of the same type, the same unit or the same building;

[0016] The time sequence dimension: taking the valve opening and closing time stamp and the single opening duration data of the target household for more than 10 consecutive days as the processing object, the time anomaly is identified by comparing three time thresholds of working days 8:00-18:00, holidays and day and night 22:00-6:00, and the single-dimension abnormal feature output is the operation frequency in the irregular period;

[0017] Operation behavior dimension: The valve opening frequency of the target household, the single operation time, the trigger source, and the historical operation model based on the LSTM neural network training are taken as the processing objects. The operation abnormality is identified by calculating the deviation rate of the real-time operation parameters and the historical operation model. The output single-dimension abnormal feature is the operation behavior deviation rate;

[0018] Environmental impact dimension: The outdoor temperature, wind speed, and supply and return water temperature difference of the target household are taken as the processing objects. The heat loss theoretical model is constructed based on the outdoor temperature-wind speed-heat loss correlation formula. The environmental abnormality is identified by comparing the deviation proportion of the actual heat loss and the theoretical heat loss value. The output single-dimension abnormal feature is the heat loss deviation rate;

[0019] Multi-source data cross-validation dimension: The analysis object also includes cross-system associated data associated with the target household. The cross-system associated data is derived from the property information management system, the water and power supply enterprise data platform, the heat company charging system, and the household identity authentication system, including 5 basic data such as the occupancy status registered by the property, the water meter reading in the past 30 days, the electricity meter reading in the past 30 days, the heat payment record in the past heating season, and the matching degree of the communication address and the house address of the real-name authentication of the household. The abnormality credibility is verified by counting the number of items pointing to the vacant state in the above cross-system associated data. The output single-dimension abnormal feature is the vacancy associated data matching degree.

[0020] Further, the standardized abnormal score of each dimension and the corresponding contribution weight are as follows:

[0021] Spatial correlation dimension: Its standardized abnormal score = 1- the heat use mode similarity of the target household and 90% of the neighborhood households; the basic weight is 0.2; the normal similarity threshold is ≥0.15. When the similarity is < the threshold, the weight increases by 10% on the basis value for every 0.05 decrease in similarity, and the highest increase is 0.3. When the similarity is ≥ the threshold, the weight decreases to 0.15;

[0022] Time series dimension: Its standardized abnormal score = the number of abnormal days of the target household ÷ the total monitoring days; the basic weight is 0.2. When the continuous abnormal days are ≥7 days, the weight increases by 20% on the basis value, and the highest is 0.24. When the continuous abnormal days are <3 days, the weight decreases to 0.15;

[0023] Operation behavior dimension: Its standardized abnormal score = the deviation rate of the valve operation parameters of the target household and the historical model ÷300%; when the deviation rate is ≥300%, the score =1; the basic weight is 0.2;

[0024] The environmental impact dimension: its standardized abnormal score = 1-actual ratio, wherein the actual ratio = actual heat loss of target household ÷ theoretical heat loss value; the basic weight is 0.15; when the actual ratio < 0.6, the weight is linearly improved according to the rule of basic weight x [1+ (0.6-actual ratio) ÷ 0.6], and the maximum improvement is 0.3;

[0025] The multi-source data cross verification dimension: its standardized abnormal score = the number of associated data items pointing to vacancy of target household ÷ the total number of associated data items; the basic weight is 0.25; when the number of associated data items pointing to vacancy is greater than or equal to 3, the weight is improved to the maximum value 0.3; when the number of items is less than 2, the weight is reduced to 0.15;

[0026] First, the contribution degree weight initial value of each dimension is calculated according to the above rules, then the sum of all contribution degree weight initial values is calculated, if the initial value sum ≠ 1, then the final contribution degree weight of each dimension = the contribution degree weight initial value of the dimension x correction coefficient, correction coefficient = 1 ÷ initial value sum.

[0027] Further, the output vacancy state confidence value ranges from 0 to 100%; the corresponding relationship between the control instruction generated by the dynamic response module and the vacancy state confidence of the target household is:

[0028] When the vacancy state confidence of the target household is greater than or equal to 85%, a "valve closing instruction" for the target household is generated, including an execution time window of valve closing (1-5 minutes after the instruction is issued) and a state monitoring period after closing (1 hour / time);

[0029] When the vacancy state confidence of the target household is greater than or equal to 85%, a "valve closing instruction" for the target household is generated, including an execution time window of valve closing (1-5 minutes after the instruction is issued) and a state monitoring period after closing (1 hour / time);

[0030] When the vacancy state confidence of the target household is less than 60%, a "maintenance monitoring instruction" for the target household is generated, including the normal feature label of the current heat mode and the time node of the next comprehensive evaluation (default 7 days later);

[0031] If the basis for triggering the control instruction includes the heat disguise behavior feature, the decision result is separately marked with the prompt information "there is suspected heat disguise behavior", and the "manual verification instruction" or "valve closing instruction" is preferentially generated under the same confidence.

[0032] Further, the first three abnormal features in the basis of the trigger control instruction all include: dimension name, specific abnormal parameter, measured data value of target household and normal threshold range; when there are contribution weights, the weights of the parallel abnormal features are adjusted according to the dimension attribution: if the parallel features belong to the operation behavior dimension or the multi-source data cross verification dimension, the weight is additionally increased by 5% for assignment; if the parallel features belong to other dimensions, the weight remains unchanged; based on the adjusted weight, secondary sorting is performed, and the features with high ranking are preferentially selected to be included in the basis of the trigger control instruction.

[0033] Further, the determination standard of the heat camouflage behavior includes the following abnormal modes: mode one: the concealed sensing data shows that the pipeline has no substantial heat consumption, while the normal sensing data presents normal heat use characteristics; mode two: the valve single operation time is less than 3 minutes, and the temperature change rate is greater than 2℃ / min; mode three: the flow micro-pulse signal in the concealed sensing data lasts for less than 0.05L / h, while the normal sensing data shows that the daily average water supply temperature is greater than 45℃; mode four: the valve opening times within 1 hour are greater than or equal to 5 times, and the single opening time is less than 2 minutes, and the radial temperature gradient change rate of the pipeline is greater than 2℃ / min; when any of the above abnormal modes or contradictory states is detected, it is determined that there is a heat camouflage behavior.

[0034] Further, the specific configuration of the hierarchical access interface is:

[0035] The hierarchical access interface, as a human-computer interaction node, is in communication connection with the visualization module of the edge computing gateway and the response module of the cloud decision platform, and provides an access channel with permission isolation for different role access subjects;

[0036] The hierarchical access interface receives the visual heat curve transmitted by the edge computing gateway and the decision result output by the cloud decision platform, and divides the data access permission according to the access subject type:

[0037] For the heat company management end: open the read permission of the full amount of target household's normal characteristic parameters, abnormal characteristic parameters, vacancy state confidence and decision result, and simultaneously open the remote control instruction issuing permission of the target household's NB-IoT on-off control valve;

[0038] For the property execution end: open the read permission of the vacancy state confidence, artificial checking instruction and heat camouflage behavior detection result of the target household within the jurisdiction, and simultaneously open the feedback uploading permission of the checking result;

[0039] For the user query end: only open the desensitized normal sensing data of the corresponding target household and the vacancy state confidence query permission of the user's own house, and shield other household data and abnormal characteristic parameters.

[0040] The application of a multi-end co-connection heat control system based on an NB-IoT on-off control valve in intelligent analysis of vacant houses, which realizes intelligent identification of the vacant state of the target house in the heating area, abnormal heat monitoring and dynamic management and control by using the multi-end co-connection heat control system, specifically comprising the following steps:

[0041] S1: data acquisition and encrypted transmission: through the dual-mode perception terminal corresponding to the target house in the system, the conventional sensing data, hidden sensing data and valve state data of the heating pipeline of the house are collected, and after being classified and encrypted by the encrypted communication unit, they are transmitted to the edge computing gateway through the NB-IoT network;

[0042] S2: data preprocessing and feature extraction: the edge computing gateway analyzes and processes the received data through double threads, extracts the conventional feature parameters and abnormal feature parameters of the target house, and synchronizes the two types of parameters to the cloud decision platform, and when receiving the heat camouflage behavior detection result fed back by the cloud decision platform, triggers the dual-mode perception terminal to perform high-frequency data acquisition;

[0043] S3: multi-dimensional vacant analysis and confidence calculation: the multi-dimensional collaborative detection engine of the cloud decision platform calculates the vacant state confidence of the target house based on the conventional feature parameters, abnormal feature parameters and cross-system associated data of the target house through parallel analysis of five dimensions of spatial correlation, time series, operation behavior, environmental influence and multi-source data cross-validation;

[0044] S4: dynamic response and result output: the cloud decision platform generates corresponding control instructions according to the vacant state confidence, and sends them to the NB-IoT on-off control valve of the target house through the edge computing gateway; at the same time, the vacant state confidence, control instruction type, heat camouflage behavior detection result and basis for triggering the control instruction are synchronized to the hierarchical access interface, and the heating company and the property can obtain and execute subsequent management and control actions according to the permission, and the user can obtain the desensitization data and vacant state confidence of his own house through the user query terminal.

[0045] Compared with the prior art, the beneficial effects of the present application are:

[0046] By integrating a dual-mode sensor group of public + hidden, combining 10Hz high-frequency sampling flow micro-pulse monitoring and three-way distributed temperature gradient detection, subtle heat behavior can be captured; at the same time, a five-dimensional collaborative detection engine of spatial correlation, time series, operation behavior, environmental influence and multi-source data cross-validation is constructed, which can accurately identify heat camouflage behaviors such as short-time valve opening and no substantial heat consumption, and improve the accuracy of vacant state determination.

[0047] Through the closed-loop control of terminal acquisition, gateway processing, cloud decision and terminal execution, the dynamic management and control of the heat loss of the vacant house is realized, the energy waste caused by invalid heat supply is reduced, the operation cost of the heat supply enterprise is reduced, at the same time, transparent heat data query service is provided for the residents, and misjudgment disputes are avoided.

[0048] The classified encryption and composite transmission mechanism are adopted, the conventional data is encrypted by using AES-128, the hidden data is encrypted by using AES-256 dynamic key, and the hidden data frame is embedded in the check bit of the conventional data frame to realize the camouflage transmission, so that the safety of data transmission is guaranteed, at the same time, the low-power consumption characteristic of the NB-IoT is utilized, combined with the event triggered high-frequency collection, the terminal power consumption and transmission cost are reduced while the data integrity and safety are guaranteed.

[0049] By constructing the hierarchical access interface, the differences of the differentiated data orderly flow of the full amount data and control authority of the heat supply company, the checking and right-to-reply authority of the property and the desensitization data query authority of the residents are determined, the data orderly flow is realized, at the same time, the correlation mechanism of the vacancy confidence-dynamic response is established, the valve is automatically closed when the confidence is greater than or equal to 85%, the manual checking is carried out when the confidence is 60%-85%, and the monitoring is maintained when the confidence is less than 60%, so that the multi-end collaborative transformation from passive inspection to active response is promoted.

[0050] In summary, the present application not only solves the core technical problems in the current central heating field, but also can produce significant energy saving, cost reduction and management optimization benefits, has technical innovation and industrial practicality, and can be widely applied to the intelligent upgrading of the central heating system in northern cities and towns. BRIEF DESCRIPTION OF DRAWINGS

[0051] Figure 1 is the system structure schematic diagram of the present application;

[0052] Figure 2 is the structure composition and working mechanism schematic diagram of the dual-mode perception terminal;

[0053] Figure 3 is the structure schematic diagram of the cloud decision platform;

[0054] Figure 4 is the structure schematic diagram of the hierarchical access interface. DETAILED DESCRIPTION

[0055] In order to facilitate the understanding of the present application, the present application will be further described in detail in combination with the drawings and specific embodiments. Those skilled in the art should understand that the embodiments are only to help understand the present application, and should not be regarded as the specific limitation of the present application.

[0056] In the following description of embodiments, specific details are set forth, such as particular system configurations, techniques, etc., in order to provide a thorough understanding of embodiments of the present application. However, persons having ordinary skill in the art will appreciate that embodiments of the present application can be practiced without these specific details. In other instances, well-known systems, structures, circuits, and techniques have not been shown in detail in order not to obscure the understanding of this description.

[0057] It is to be understood that the terminology "includes", "has", "holds", "contains" and / or "comprising", "including", "having" and their conjugates, as used herein, are used in the sense of "including but not limited to", and not in the sense of "consisting only of the mater-ials recited one or more times in the claim or introduction" or "consisting of only of the materials recited one or more times in the claim or introduction", unless the context clearly indicates otherwise. It is also to be understood that the terminology "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items, as well as the lack of any associated item.

[0058] As used in the description of embodiments and the appended claims herein, the term "if' can be interpreted to mean "when" or "upon" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [a described condition or event] is detected" can be interpreted to mean "upon determining" or "in response to determining" or "upon detecting [the described condition or event]" or "in response to detecting [the described condition or event]", depending on the context.

[0059] In addition, the terms "first", "second", "third", etc. as used in the description of embodiments and the appended claims herein are used only to differentiate one element from another, and do not imply a relative importance or a chronological sequence. The terms "one embodiment", "some embodiments", "an embodiment", "another embodiment", "at least one embodiment", "one example", "some examples", "an example", "at least one example", etc. as used in the description of embodiments herein are to be understood to refer to one or more embodiments of the present application, unless otherwise stated. As used in the description of embodiments herein, the term "embodiment" means an implementation or implementation example. As used in the description of embodiments, the terms "include", "including", "comprise", "comprising", "have", "having" and variants thereof are meant to be open-ended terms that specifically allow for the inclusion of more than one component in a given medium, composition, process, method, step, etc. As used in the description of embodiments herein, the term "about" means that a value is within a range of values that one of ordinary skill in the art would consider to be equivalent to the value being about. As used in the description of embodiments herein, the term "coupled" means directly or indirectly connected, linked, or associated.

[0060] Embodiment 1 :

[0061] As Figure 1As shown, the application provides a multi-end joint heating control system based on NB-IoT on-off control valve, which is characterized by comprising a dual-mode sensing terminal, an edge computing gateway, a cloud decision platform and a hierarchical access interface.

[0062] The dual-mode sensing terminal comprises an NB-IoT on-off control valve, a dual-mode sensor group and an encrypted communication unit; the NB-IoT on-off control valve is configured with an electric actuator and a state monitoring module, the electric actuator is used to receive control instructions and execute heating pipeline on-off operation, and the state monitoring module is used to collect valve state data; the dual-mode sensor group comprises an open detection module and a hidden detection module, wherein the open detection module is used to collect regular sensing data, and the hidden detection module is used to collect hidden sensing data; the encrypted communication unit classifies and encrypts the valve state data, the regular sensing data and the hidden sensing data before sending them to the edge computing gateway, and parses the received control instructions to drive the NB-IoT on-off control valve to act; the edge computing gateway is used to realize the communication connection between the dual-mode sensing terminal and the cloud decision platform, and is configured with a first data processing thread and a second data processing thread for parallel processing; the first data processing thread receives and parses the basic information in the regular sensing data and the valve state data, generates a visual heating curve and synchronizes it to the hierarchical access interface, extracts the regular feature parameters of the corresponding target household and sends them to the cloud decision platform; the second data processing thread receives and decrypts the operation behavior information in the hidden sensing data and the valve state data, extracts the abnormal feature parameters of the corresponding target household and sends them to the cloud decision platform.

[0063] The cloud decision platform comprises a multi-dimensional collaborative detection engine and a dynamic response module; the multi-dimensional collaborative detection engine is used to perform parallel analysis on the regular feature parameters and the abnormal feature parameters of the target household in five dimensions of spatial correlation, time sequence, operation behavior, environmental influence and multi-source data cross verification: for each dimension, extract the single-dimensional abnormal feature pointing to vacancy under this dimension, then calculate the standardized abnormal score reflecting the abnormal degree based on the single-dimensional abnormal feature, and dynamically determine the contribution weight of each dimension according to the degree of deviation of the single-dimensional abnormal feature from the normal threshold, and then calculate the weighted score of each dimension by multiplying the standardized abnormal score of this dimension by the corresponding contribution weight; finally, the weighted scores of the five dimensions are summed to obtain the vacancy state confidence of the target household; and in the operation behavior dimension analysis, whether there is heat camouflage behavior is determined by comparing the consistency of the hidden sensing data and the regular sensing data, if there is heat camouflage behavior, the behavior feature is taken as the key abnormal feature parameter of this dimension, and the contribution weight is increased.

[0064] The dynamic response module generates a corresponding control instruction according to the vacancy state confidence of the target household, and the control instruction and the heat use camouflage behavior detection result are delivered together through the edge computing gateway, and the control instruction is used to drive the NB-IoT on-off control valve of the target household to act; after receiving the heat use camouflage behavior detection result, if there is a heat use camouflage behavior, the edge computing gateway triggers the dual-mode perception terminal of the target household to execute the high-frequency acquisition instruction; at the same time, the decision result including the vacancy state confidence of the target household, the control instruction type, the heat use camouflage behavior detection result and the basis for triggering the control instruction is synchronized to the hierarchical access interface; the basis for triggering the control instruction is the first three abnormal features and the corresponding data values in the contribution weight order.

[0065] Embodiment 2:

[0066] The application also provides an application of a multi-end co-connection heat control system based on an NB-IoT on-off control valve in intelligent analysis of vacant houses, which realizes intelligent identification of the vacancy state of a target household house in a heating area, abnormal heat use monitoring and dynamic control by using the above multi-end co-connection heat control system, and specifically includes the following steps:

[0067] S1: data acquisition and encrypted transmission: through the dual-mode perception terminal corresponding to the target household, the conventional sensing data, hidden sensing data and valve state data of the heating pipeline of the household are acquired, and after being classified and encrypted by the encrypted communication unit, they are transmitted to the edge computing gateway through the NB-IoT network;

[0068] S2: data preprocessing and feature extraction: the edge computing gateway analyzes and processes the received data through double threads respectively, extracts the conventional feature parameters and abnormal feature parameters of the target household, and synchronizes the two types of parameters to the cloud decision platform, and when receiving the camouflage behavior detection result fed back by the cloud decision platform, triggers the dual-mode perception terminal to execute high-frequency data acquisition;

[0069] S3: multi-dimensional vacancy analysis and confidence calculation: the multi-dimensional collaborative detection engine of the cloud decision platform, based on the conventional feature parameters, abnormal feature parameters and cross-system associated data of the target household, analyzes in parallel through five dimensions of spatial correlation, time series, operation behavior, environmental influence and multi-source data cross-validation, determines whether there is a camouflage behavior, and calculates the contribution weight of each dimension abnormal feature, and obtains the vacancy state confidence of the target household by weighted summation;

[0070] S4: Dynamic response and result output: the cloud decision platform generates corresponding control instructions according to the vacancy state confidence, and sends them to the NB-IoT on-off control valve of the target household through the edge computing gateway; at the same time, the vacancy state confidence, control instruction type, camouflage behavior detection result and basis for triggering control instructions are synchronized to the hierarchical access interface, and the heat company and the property can obtain and execute subsequent control actions according to their permissions, and the user can obtain the desensitization data and vacancy state confidence of his own house through the user query terminal.

[0071] Embodiment 3:

[0072] I. Application scenario

[0073] This embodiment is applied to a certain provincial capital city XX garden community in the north (12 buildings in total, 30 floors for each building, 2 households for each floor, and a total of 720 households), which adopts centralized heating. The heating season in winter is from November 15 to March 15 of the next year, and there are about 20% of vacant houses (mostly new houses that have not been occupied, houses of owners on short-term business trips or houses of owners who want to avoid heating fees by pretending to use heat). Through the multi-terminal joint heat control system, vacancy house identification, heat loss control and multi-terminal collaboration are realized, and scenes such as single household accurate detection, whole building batch analysis, hardware detail analysis and multi-role permission management are covered, verifying the all-round practicality of the system.

[0074] II. Specific configuration and collaboration logic of system modules

[0075] (1) Dual-mode sensing terminal

[0076] The internal components and working mechanism details of the dual-mode sensing terminal are shown in Figure 2

[0077] Core execution component: NB-IoT on-off control valve, using a DN20 size electric ball valve, the electric actuator is a direct current electric actuator with model ZJS-20, the rated power is 5W, the response time is less than or equal to 1 second, and the ball valve opening is driven to 30% within 1 second after receiving the "valve opening 30%" instruction, which performs the heat limiting operation of the heating pipeline.

[0078] The state monitoring module is built-in Hall sensor with model A3144, which collects valve state data in real time, including:

[0079] ​Switch status identifier: binary, 0 = off, 1 = on; cumulative on duration: unit minutes / day, e.g. a certain household's cumulative on duration on January 12th is 180 minutes; single operation duration: unit seconds; operation trigger source: enumerated value, 0 = remote instruction, 1 = local manual, 2 = automatic adjustment, e.g. the operation trigger source of the opening instruction at 14:00 is 0 (remote from the heat company); real-time opening degree value: 0-100% continuous value, accuracy 1%; fault diagnosis code: hexadecimal, 0x00 = normal, 0x01 = blockage, 0x02 = under-voltage, 0x03 = communication abnormality, data transmitted to the encrypted communication unit through the internal I2C bus.

[0080] Dual-mode sensor group: the disclosed detection module uses PT1000 temperature sensors, installed at the water inlet end (supply water temperature) and the water outlet end (return water temperature) of the heating pipeline, respectively, with a sampling frequency of 30 minutes / second, a measurement range of 5-90°C, and an accuracy of ±0.5°C. For example, on January 10th, the average supply water temperature of a certain household is 29°C, and the average return water temperature is 22°C. On January 12th, the supply water temperature is 29°C at 9:00 and 30°C at 9:30, while the return water temperature is 23°C and 24°C at the same time, constituting the conventional sensor data reflecting the basic operation state of heating.

[0081] Hidden detection module: a miniature turbine flow sensor of model FS3003 is installed in the pipeline to collect flow micro-pulse signals, with a 10Hz high-frequency sampling rate (i.e. 10 times per second), a measurement range of 0-0.5L / h, and a resolution of 0.01L / h. For example, when a certain household is pretending to use heat, the flow micro-pulse signal is sustained at 0.03-0.04L / h (much lower than the normal heat consumption threshold of 10-20L / h, with no substantial heat consumption), and the data is transmitted through the SPI bus.

[0082] Three-way distributed temperature sensor: a model DS18B20 is installed at 1cm, 2cm, and 3cm along the radial direction of the pipeline, with a measurement range of 0-5°C and an accuracy of ±0.1°C, capable of collecting the radial temperature gradient of the pipeline. For example, on January 12th at 10:00, the sampling values are 29°C, 28.5°C, and 28°C, respectively, with a calculated temperature gradient of 0.5°C. When a certain household is pretending to use heat, the temperature rises from 27°C at 10:00:00 to 30°C at 10:00:30, with an instantaneous temperature change rate of 6°C / min (exceeding the threshold of ">2°C / min" for mode two heat pretending behavior).

[0083] Event-triggered wake-up unit: a microcontroller of model STM8L051 is used, with a preset trigger condition of "valve opening <5 minutes between 2-4am", e.g. on January 12th at 3:10, the valve opening is detected for 3 minutes, immediately activating the flow and temperature sensors, and continuously collecting data for 1 hour (3:10-4:10) with a sampling interval reduced from 30 minutes / second to 6 minutes / second.

[0084] Encryption communication unit: taking STM32L475 chip as the core, the valve state data, regular sensing data and hidden sensing data are classified and encrypted, and after the classified and encrypted data are transmitted to the edge computing gateway through the China Mobile NB-IoT network, the control instructions issued by the edge computing gateway are received to drive the valve action.

[0085] The valve basic operation information (on-off state identification, cumulative opening time) in the regular sensing data (water supply / return water temperature) and valve state data is encrypted by AES-128, and the fixed key is the last 16 bits of the device unique identifier (UUID) of the dual-mode sensing terminal, such as "8A7B6C5D4E3F2G1H" after the last 16 bits of UUID, and the encrypted data frame with frame header 0x01 is generated.

[0086] The sensitive operation information (operation trigger source, real-time opening value) in the hidden sensing data (flow micro-pulse, temperature gradient, instantaneous temperature change rate) and valve state data is encrypted by AES-256, and the dynamic key is the first 16 bits of UUID + real-time timestamp "YYYYMMDDHH", which is updated once a day, such as "1A2B3C4D5E6F7G8H" + timestamp "2024011010" for the first 16 bits of UUID, and the encrypted hidden data frame with frame header 0x02 is generated.

[0087] Composite transmission: embedding the hidden data frame into the CRC check bit (expanding from 8 bits to 32 bits) of the regular data frame, and transmitting it through the NB-IoT module (model BC20); after receiving by the edge computing gateway, it is parsed according to the frame header 0x01 / 0x02, and the CRC check code (8-bit regular check + 24-bit hidden check) is verified to ensure data integrity.

[0088] (2) Edge computing gateway

[0089] Deployed in the community heat exchange station, using an industrial gateway (model EC200S), configured with dual data processing threads, single-household data processing and whole-building data transfer can be realized:

[0090] 1) Single-household data processing (taking 101 room of 1 unit of No. 1 building and 202 room of 2 unit of No. 3 building as examples):

[0091] The first data processing thread: parsing regular sensing data and valve basic state, generating resident daily heat curve, such as 202 room heat curve from 8:00 to 12:00 on January 10, horizontal axis time, vertical axis temperature 22-29℃, synchronizing to hierarchical access interface; extracting regular feature parameters, such as 101 room daily average temperature 25℃, valve opening rate 50%, 202 room daily average temperature 25.5℃, valve opening rate 8% (cumulative opening 9 minutes ÷ 240 minutes of regular heat period), uploading to cloud decision platform.

[0092] Second data processing thread: decrypt covert sensor data, extract abnormal feature parameters. For example, room 101 has daily flow micro-pulse 0.03 L / h, temperature gradient change rate 0.5 ℃ / min, valve opening 4 times (single <2 minutes) within 1 hour; room 202 has normal heating period "opening 2 times (single <3 minutes) within 1 hour, flow 0.04 L / h, instantaneous temperature change rate 6 ℃ / min", non-heating period "operation 1 time (single ≥3 minutes)", and abnormal features are marked and uploaded to the cloud decision platform.

[0093] High-frequency acquisition trigger: when receiving the cloud decision platform feedback "suspected heating camouflage behavior", for example, on January 10, 14:00, the cloud determines that room 202 has mode two heating camouflage, immediately issues an instruction to the dual-mode perception terminal to shorten the sampling interval from 30 minutes / time to 6 minutes / time, and continuously collects data for 1 hour (14:00-15:00), verifying that there is no substantial heat consumption.

[0094] 2) Whole building data transfer (take building 3 with 60 households as an example):

[0095] Receive regular / abnormal feature parameters of all households in building 3, such as room 3-1-102 with daily average temperature 26 ℃, valve opening rate 60%, flow micro-pulse 0.06 L / h; synchronize cross-system associated data (property "15 households not occupied", water supply "12 households with water consumption <5 m 3 " in the last 30 days, power supply "10 households with electricity consumption <10 kWh in the last 30 days"), and push them to the cloud decision platform in real time through the API interface; after receiving the cloud instruction, forward it to the corresponding dual-mode perception terminal, such as pushing the 12 households to the property APP for manual verification.

[0096] (3) Cloud decision platform

[0097] The whole process logic of the cloud decision platform's five-dimensional analysis, confidence calculation, and dynamic response is shown in the following figure: Figure 3

[0098] The cloud decision platform is deployed on the heat company's cloud server (Aliyun ECS instance), with functions of single household accurate analysis, whole building batch decision, dynamic response, and data archiving. The core modules are as follows:

[0099] Data input: receive single household / whole building regular / abnormal feature parameters uploaded by edge computing gateway, synchronize cross-system associated data (property registration status, water and electricity consumption, access control record, communication address matching degree, heat payment record), such as room 101 "property not occupied, water meter 0 m 3 , electricity meter 2 kWh, communication mismatch, not paid", room 3-1-102 "property not occupied, water meter 0 m 3 , electricity meter 3 kWh, not paid, communication mismatch".​

[0100] Multi-dimensional collaborative detection engine: For single or full building residents, perform five-dimensional parallel analysis of spatial correlation, time series, operation behavior, environmental impact, and multi-source data cross-validation, calculate standardized anomaly score, contribution weight and vacancy state confidence.

[0101] 1) Single unit analysis (take 1-1-101 room and 3-2-202 room as an example)

[0102]

[0103] 2) Full building batch analysis (take 60 households in Building 3 as an example)

[0104] Take 3-1-102 room as an example:

[0105] Spatial correlation dimension: Similarity with 90% of the same unit residents (17 households) is 0.12 (lower than the threshold of 0.15), standardized anomaly score = 1-0.12 = 0.88;

[0106] Time series dimension: Turned on 3 times (single time <3 minutes) at 2-4 am in the past 10 days, standardized anomaly score = 3 ÷ 10 = 0.3;

[0107] Operation behavior dimension: Turned on 5 times (single time <2 minutes) within 1 hour, temperature gradient change rate 2.5℃ / min (exceeds mode four threshold), deviation rate = (1.5 ÷ 5) × 100% + 30% = 100%, standardized anomaly score = 1.0;

[0108] Environmental impact dimension: Actual heat loss 0.042W, theoretical heat loss 81.6W, standardized anomaly score = 0.9995;

[0109] Multi-source data cross-validation dimension: 5 types of data all point to vacancy, standardized anomaly score = 1.0.

[0110] 3) Contribution weight and confidence calculation

[0111] Weight calculation steps (take 3-1-102 room as an example):

[0112] i. Preset basic weight: space 0.2, time 0.2, operation 0.2, environment 0.15, multi-source 0.25 (total 1);

[0113] ii. Calculate the initial weight of each dimension: the spatial dimension is reduced by 0.03 (0.15-0.12) due to similarity, according to the rule of increasing by 10% for every 0.05 reduction, no increase needed → keep 0.2; the time dimension is reduced to 0.15 due to consecutive abnormal days < 3 days; the operation dimension is increased by 40% → 0.2 x 1.4 = 0.28 due to pattern four camouflage; the environment dimension is increased to 0.25 due to actual ratio < 0.6; the multi-source dimension is increased to 0.3 due to item number ≥ 3;

[0114] iii. Dynamic balance: the sum of initial values = 0.2 + 0.15 + 0.28 + 0.25 + 0.3 = 1.18, correction factor = 1 ÷ 1.18 ≈ 0.8475;

[0115] iv. Final weight: space 0.2 x 0.8475 ≈ 0.1695, time 0.15 x 0.8475 ≈ 0.1271, operation 0.28 x 0.8475 ≈ 0.2373, environment 0.25 x 0.8475 ≈ 0.2119, multi-source 0.3 x 0.8475 ≈ 0.2542 (sum ≈ 1).

[0116] Vacancy confidence calculation results:

[0117] 101 room: (0.91 x 0.1667 + 1.0 x 0.1818 + 1.0 x 0.1969 + 0.9995 x 0.2273 + 0.8 x 0.2273) x 100% ≈ 93.94%;

[0118] 202 room: (0.89 x 0.22 + 0.30 x 0.06 + 0.90 x 0.10 + 0.99 x 0.15 + 1.0 x 0.20) x 100% ≈ 65%;

[0119] 3-1-102 room: (0.88 x 0.1695 + 0.3 x 0.1271 + 1.0 x 0.2373 + 0.9995 x 0.2119 + 1.0 x 0.2542) x 100% ≈ (0.1492 + 0.0381 + 0.2373 + 0.2118 + 0.2542) x 100% ≈ 89.06%.

[0120] Dynamic response module: generate customized instructions according to the confidence interval ( < 60%: continuous monitoring; 60%-85%: manual verification; ≥ 85%: automatic valve closing), specific rules as follows:

[0121] 1) High confidence (≥ 85%, such as 101 room 93.94%, 3-1-102 room 89.06%):

[0122] Instruction generation: both 101 and 3-1-102 rooms are issued valve closing instructions within 3 minutes, and the monitoring period is adjusted from 15 minutes / time to 1 hour / time after closing, focusing on monitoring abnormal opening of valves and sudden heat loss changes. For the mode four camouflage features of 3-1-102 room, additional temperature gradient mutation monitoring (threshold set to 2°C / min) is added, and once triggered, the abnormal event is immediately uploaded.

[0123] Result push:

[0124] Push full data to the heat company: including 101 room confidence 93.94%, dimension score / weight details, trigger control instruction basis (mode one camouflage + four multi-source data); 3-1-102 room confidence 89.06%, high weight proportion of operation behavior dimension 0.2373, mode four camouflage feature parameters.

[0125] Push classification tips to property: 101 room does not need to be checked (high confidence + multi-source data fully supported); 3-1-102 room auxiliary check (need to confirm whether there is a tenant short-term departure situation, no need to repeat verify heat camouflage).

[0126] Push desensitization results to residents: uniformly prompt "your house is confirmed to be empty by the system detection, the heating valve has been closed, if you need to restore heating, you can submit an application through the heat APP", and the specific detection dimension details are shielded.

[0127] 2) Confidence (60%-85%, such as 202 room 65%):

[0128] Instruction generation: generate artificial check instructions, the core content includes:

[0129] Objective: confirm whether 202 room is short-term empty (owner short-term business trip) or camouflage empty (deliberately avoid fees)

[0130] Key abnormal feature list (sorted by weight): mode two camouflage of operation behavior dimension (0.10 weight), heat loss anomaly of environmental impact dimension (0.15 weight)

[0131] Check requirements: take pictures of the entrance door status (whether there are recent entry and exit marks), record the valve physical position (whether it is consistent with the system state), take pictures of the water and electricity meter readings (verify the accuracy of system data collection)

[0132] Feedback time limit: 72 hours

[0133] Instruction flow: issued through MQTT protocol to edge computing gateway, directed to the property APP's to-be-checked task module, and simultaneously pushed to the heat company's check progress board (real-time display of task status: not ordered / processing / finished).

[0134] Result closed loop: Property on-site verification result uploaded ("no sealing tape on the door, found recent express packaging, confirmed short-term business trip"), cloud combined with new data to recalculate confidence (room 202 from 65% to 58%), trigger to remove key monitoring instructions, restore normal sampling frequency (30 minutes / time); To multiple ends, the final result is synchronized, such as the heat company end showing "room 202: not intentionally disguised, recommend keeping heating".

[0135] Abnormal plan:

[0136] Property overdue feedback: For room 3-2-501 (confidence 78%) more than 72 hours without feedback, automatically send a warning to the heat company, generate a heat specialist verification instruction, with higher priority than regular tasks.

[0137] Owner objection processing: Room 3-1-102 owner submits "objection application + recent occupancy certificate", cloud starts secondary verification: cross-verify access records (12 entries in the past 7 days), increase time series dimension weight to 0.25, recalculate confidence to 52%, immediately trigger valve opening instruction and cancel key monitoring.

[0138] Equipment failure response: Room 101 feedbacks fault code 0x01 (card jam), system pushes maintenance instructions to heat operation end within 24 hours, uses temporary control during maintenance: valve opens ≤1 minute at a time, ≤5 minutes per day, to avoid invalid heat loss.

[0139] (4) Hierarchical access interface

[0140] As shown in Figure 4 , the hierarchical access interface is based on a three-level mapping mechanism of "role-permission-data", which assigns differentiated permissions to three types of roles: heat company management end, property execution end, and user query end, realizes data security isolation and accurate flow, and supports multi-end collaborative control.

[0141] 1. Core control module: permission management center

[0142] Deployed in the cloud decision platform, the permission templates of the three types of roles are preset and stored in the permission database, and the specific mapping rules are as follows:

[0143]

[0144] Data distribution logic: After receiving the full-floor heat curve uploaded by the edge computing gateway, valve opening rate statistics, and the cloud decision platform's vacancy determination result, automatically filter sensitive data according to role permissions (such as shielding other resident information, hiding abnormal feature calculation details), and push to the corresponding terminal.

[0145] 2. Specific implementation of each role access terminal

[0146] 1) Thermal Company Management End (PC Management System)

[0147] Data Access:

[0148] i. Single Household Details: View the full-dimensional data of Room 1-1-101, including regular sensor data (average daily water supply temperature 28°C, return water temperature 22°C), abnormal characteristic parameters (flow micro-pulse 0.03L / h, temperature gradient 0.8°C), and calculation details of confidence 93.94% (dimension score x weight); pattern four heat camouflage behavior record of Room 3-1-102 (opened 5 times in 1 hour, temperature gradient change rate 2.5°C / min).

[0149] ii. Building-wide Statistics: View the confidence distribution table of 60 vacant households in Building 3 (7% households ≥ 85%, 20% households 60%-85%, 73% households < 60%), heat camouflage behavior classification statistics (pattern one 8 households, pattern four 4 households), and heat loss control effect (valves closed in 12 households, daily average heat loss saved about 800W).

[0150] Operation Authority:

[0151] i. Instruction Issuance: Directly issue valve closing instructions to Room 1-1-101 and Room 3-1-102 (confidence ≥ 85%), and view the instruction execution status in real time (e.g. Room 101 closed successfully at 15:03); issue heat specialist verification instructions to Room 3-2-501 (overdue for verification).

[0152] ii. System Configuration: Adjust the five-dimensional basic weight (e.g. fine-tune the multi-source cross-validation dimension basic weight from 0.25 to 0.28), modify the confidence threshold interval (e.g. adjust the lower limit of the artificial verification threshold from 60% to 55%), and export the building-wide monthly heating data report (including vacancy rate, heat loss amount, and abnormal event statistics).

[0153] 2) Property Execution End (Mobile APP)

[0154] Data Access:

[0155] i. Task List: Only show information of 12 households in Building 3 that need artificial verification, each task marked with core abnormal characteristics (e.g. Room 202 "Pattern Two Camouflage + Multi-source Data Contradiction", Room 3-1-102 "Pattern Four Camouflage + Heat Loss Abnormality"), verification time limit (72 hours), and basis for triggering control instructions (confidence 65%-73%).

[0156] ii. Progress Tracking: View the results of 8 households that have completed verification (e.g. Room 202 "confirmed short-term business trip, not intentional camouflage"), and countdown of 4 households awaiting feedback (e.g. Room 3-2-501 remaining 24 hours).

[0157] Operation Authority:

[0158] i. Task processing: Click "Start Verification" to upload on-site photos (e.g., no seal on the entrance door of Room 202, water and electricity meter readings 0m 3 ), fill in the verification conclusion ("Short-term vacancy, recommend retaining basic heating"), and submit. After submission, the cloud automatically updates the corresponding resident confidence (Room 202 from 65% to 58%).

[0159] ii. Abnormal feedback: Found that the physical state of the valve in Room 3-1-102 is inconsistent with the system display (system shows closed, actual on-site opening), upload photos through the "Device Abnormality" portal, trigger the cloud "Valve State Calibration" instruction.

[0160] 3) User query end (WeChat mini-program / APP)

[0161] Data access (desensitization processing):

[0162] i. Basic information: After logging in, the owner of Room 1-1-101 only views the daily average water temperature 28℃, valve state "closed", and vacancy confidence 93.94%; the owner of Room 3-1-102 views the confidence 89.06% and the "valve closed" prompt, shielding sensitive abnormal parameters such as flow micro-pulse and temperature gradient.

[0163] ii. History record: View the heating state record of the house in the past 1 month (e.g., "vacancy monitoring, valve closed" from January 10 to January 15), and the processing progress of the objection application / recovery of heating application (e.g., "recovery application on January 12, opened on January 13").

[0164] Operation authority:

[0165] i. Objection appeal: The owner of Room 3-1-102 has objections to the vacancy determination, submits "objection application" and uploads the residence certificate (access records of the past 7 days, water and electricity payment vouchers), and the cloud receives it and starts the second verification.

[0166] ii. Service application: The owner of Room 1-1-101 plans to move in, submits "recovery of heating application", selects the opening time (e.g., January 20, 9:00), and the system automatically issues "valve opening instruction" after passing the audit, and synchronously pushes the "opening success" notification.

[0167] 3. Data security guarantee

[0168] Desensitization rules: hide "abnormal feature parameters" (such as flow micro-pulse, temperature gradient) from the user query end, "other resident data" (such as neighborhood heating mode similarity calculation, only show "significant deviation", do not show specific resident data), and "determination details" (such as not showing the weight and score of each dimension, only showing the final confidence).

[0169] Permission check: When the property execution end tries to access the data of building No. 1, the system automatically intercepts and prompts "no permission to access non-jurisdiction building"; when the user query end tries to view the confidence of other residents, it displays "only the information of the own house can be viewed".

[0170] Operation traceability: all instructions, data modification, and check feedback operations are recorded in operation logs (including operator, time, and content), stored in the cloud database for 1 year, and convenient for subsequent traceability (such as checking the "3-1-102 room valve error" event, which can be traced back to the heat company instruction on January 10).

[0171] In the above embodiments, the description of each embodiment has its own focus, and the parts not described or recorded in a certain embodiment can be referred to the related description of other embodiments.

[0172] In each embodiment, the hardware implementation of the technology can directly use existing intelligent devices, including but not limited to industrial computers, PC computers, smart phones, handheld computers, floor-standing computers, etc. The input device is preferably a screen keyboard, the data storage and calculation module uses existing memory, calculator, controller, the internal communication module uses existing communication ports and protocols, and the remote communication uses existing GPRS network, World Wide Web, etc.

[0173] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is taken as an example for illustration, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The above integrated unit can be realized in the form of hardware or in the form of software functional unit. In addition, the specific name of each functional unit and module is only for easy distinction, and does not limit the protection scope of the present application. The specific working process of the units and modules in the above system can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0174] In the embodiments of the present application, it should be understood that the disclosed apparatus / terminal device and method can be implemented in other manners. For example, the embodiments of the apparatus / terminal device described above are merely schematic, and the division of the modules or units is merely logical function division, and there can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between different units can be indirect couplings or communication connections through some interfaces, devices or units, and can be electrical, mechanical or in other forms. The units described as separate components can or can not be physically separate, and components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. In actual implementation, some or all of the units can be selected according to the actual needs to achieve the purposes of the embodiments of the present application.

[0175] The various function units in the embodiments of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of software function unit. When the integrated module / unit is realized in the form of software function unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on such understanding, all or part of the flow of the method in the above embodiments can be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium, and when the processor executes the computer program, the steps of the various method embodiments described above can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable file or some intermediate form. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

Claims

1. A multi-terminal interconnected thermal control system based on NB-IoT on / off control valves, characterized in that, This includes dual-mode sensing terminals, edge computing gateways, cloud decision-making platforms, and hierarchical access interfaces; The dual-mode sensing terminal includes an NB-IoT on / off control valve, a dual-mode sensor group, and an encrypted communication unit. The NB-IoT on / off control valve is equipped with an electric actuator and a status monitoring module. The electric actuator is used to receive control commands and perform on / off operations on the heating pipeline, and the status monitoring module is used to collect valve status data. The dual-mode sensor group includes a public detection module and a hidden detection module, wherein the public detection module is used to collect conventional sensing data, and the hidden detection module is used to collect hidden sensing data. The encrypted communication unit classifies and encrypts the valve status data, conventional sensing data, and hidden sensing data before sending them to the edge computing gateway, and parses the received control commands to drive the NB-IoT on / off control valve to operate. The edge computing gateway is used to realize the communication connection between the dual-mode sensing terminal and the cloud decision-making platform; the cloud decision-making platform includes a multi-dimensional collaborative detection engine and a dynamic response module; the multi-dimensional collaborative detection engine is used to perform parallel analysis of the target household's regular and abnormal characteristic parameters in five dimensions: spatial correlation, time series, operational behavior, environmental impact, and multi-source data cross-validation, to obtain the vacancy confidence of the target household. The dynamic response module generates a corresponding control command based on the vacancy confidence level of the target resident and sends the control command through the edge computing gateway. The control command is used to drive the NB-IoT on / off control valve of the target resident to operate. At the same time, the decision result, including the vacancy confidence level of the target resident, the control command type, and the basis for triggering the control command, is synchronized to the hierarchical access interface.

2. The multi-terminal interconnected thermal control system according to claim 1, characterized in that, The encryption mechanism of the encrypted communication unit is as follows: For basic operating status information and conventional sensing data in the valve status data, the AES-128 symmetric encryption algorithm is used, with the last 16 bits of the unique device identifier of the dual-mode sensing terminal as a fixed key. After encryption, a conventional data frame is generated, with the data type identifier bit 0x01 marked in the frame header. For sensitive operation information and hidden sensing data in the valve status data, the AES-256 symmetric encryption algorithm is used, with the first 16 bits of the unique device identifier of the dual-mode sensing terminal and the real-time timestamp concatenated as a dynamic key. After encryption, a hidden data frame is generated, with the data type identifier bit 0x02 marked in the frame header. The encrypted data frame adopts a composite transmission structure of conventional data frame + hidden data frame. The hidden data frame embeds the check bit field of the conventional data frame. The two types of data are distinguished and parsed through the frame header identifier bit. All encrypted data is appended with an 8-bit CRC check code for integrity verification.

3. The multi-terminal interconnected thermal control system according to claim 1, characterized in that, The concealed detection module includes a miniature turbine flow sensor, three distributed temperature sensors, and an event-triggered wake-up unit. The miniature turbine flow sensor is used to acquire the circulating water flow rate in the heating pipes of the target household. The acquired data is processed by the encrypted communication unit and embedded as a concealed data frame into the checksum field of the regular data frame. The frame header identifier bit 0x02 is used to achieve disguised transmission. The three distributed temperature sensors are used to acquire the radial temperature difference of the pipe to identify false temperature changes caused by the brief opening of the valve. The event-triggered wake-up unit is used to automatically activate the miniature turbine flow sensor and the three distributed temperature sensors when a valve is detected to be opened for less than 5 minutes between 2-4 am. The system continuously acquires data for 1 hour, and the data transmission interval during the acquisition period is shortened to 1 / 5 of the regular transmission interval.

4. The multi-terminal interconnected thermal control system according to claim 1, characterized in that, The edge computing gateway is configured with a first data processing thread and a second data processing thread for parallel processing. The first data processing thread receives and parses the basic information in the regular sensor data and valve status data, generates a visualized thermal curve and synchronizes it to the hierarchical access interface, and extracts the regular characteristic parameters of the corresponding target household and sends them to the cloud decision platform. The second data processing thread receives and decrypts the operation behavior information in the hidden sensor data and valve status data, extracts the abnormal characteristic parameters of the corresponding target household and sends them to the cloud decision platform.

5. The multi-terminal interconnected thermal control system according to claim 1, characterized in that, The parallel analysis of the five dimensions specifically includes: for each dimension, extracting single-dimensional abnormal features pointing to vacancy under that dimension; then calculating a standardized anomaly score reflecting the degree of anomaly based on the single-dimensional abnormal feature; dynamically determining the contribution weight of each dimension according to the degree to which the single-dimensional abnormal feature deviates from the normal threshold; then calculating the weighted score of each dimension by multiplying the standardized anomaly score of that dimension with the corresponding contribution weight; finally, summing the weighted scores of the five dimensions to obtain the confidence level of the vacancy status of the target household.

6. The multi-terminal interconnected thermal control system according to claim 5, characterized in that, In the analysis of operational behavior, the consistency between concealed sensor data and regular sensor data is compared to determine whether heat camouflage behavior exists. If heat camouflage behavior exists, the behavior feature is regarded as the key abnormal feature of this dimension and its contribution weight is increased.

7. The multi-terminal interconnected thermal control system according to claim 6, characterized in that, The dynamic response module sends the control command and the detection result of the heat camouflage behavior together through the edge computing gateway. After receiving the detection result of the heat camouflage behavior, if the heat camouflage behavior exists, the edge computing gateway will trigger the dual-mode sensing terminal of the target household to execute the high-frequency acquisition command. The decision results synchronized to the hierarchical access interface also include the results of hot masquerading behavior detection.

8. The multi-terminal interconnected thermal control system according to claim 5, characterized in that, The single-dimensional anomaly features corresponding to the five dimensions of parallel analysis in the multi-dimensional collaborative detection engine are: Spatial correlation dimension: Taking the heating pattern data of neighboring residents of the target household as the processing object, spatial anomalies are identified by calculating the similarity of heating patterns between the target household and more than 90% of neighboring residents. The output single-dimensional anomaly feature is the deviation of neighboring heating patterns; the neighboring residents are residents of the same apartment type, the same unit, or the same building. Time series dimension: The data on valve opening and closing timestamps and single opening duration of the target household for more than 10 consecutive days are used as the processing objects. By comparing the three-level time thresholds of weekdays 8:00-18:00, holidays, and day and night 22:00-6:00, time anomalies are identified. The output single-dimensional anomaly feature is the frequency of operation during irregular time periods. Operational behavior dimension: Taking the real-time operation parameters of the target residents and the historical operation model trained based on LSTM neural network as the processing objects, operational anomalies are identified by calculating the deviation rate between the real-time operation parameters and the historical operation model, and the output single-dimensional anomaly feature is the operational behavior deviation rate. Environmental impact dimension: Taking the environmental data of the target residents and the temperature difference between the supply and return water as the processing objects, a heat loss theoretical model is constructed based on the outdoor temperature-wind speed-heat loss correlation formula. Environmental anomalies are identified by comparing the deviation ratio between the actual heat loss and the theoretical heat loss value. The output single-dimensional anomaly feature is the heat loss deviation rate. Multi-source data cross-validation dimension: The analysis also includes cross-system related data associated with the target household. The cross-system related data comes from the property information management system, the data platform of water and electricity supply companies, the billing system of heating companies, and the resident identity authentication system. It includes five basic data items: occupancy status of property registration, water meter readings in the past 30 days, electricity meter readings in the past 30 days, heating payment records for the past heating season, and the matching degree between the resident's real-name authentication communication address and the house address. The credibility of the anomaly is verified by statistically analyzing the number of items pointing to the vacancy status in the above cross-system related data. The output single-dimensional anomaly feature is the vacancy related data matching degree.

9. The multi-terminal interconnected thermal control system according to claim 8, characterized in that, The standardized anomaly scores and corresponding contribution weights for each dimension are as follows: Spatial Relationship Dimension: Its standardized anomaly score = 1 - similarity of heating patterns between the target household and 90% of neighboring households; the base weight is 0.2; the normal similarity threshold is ≥0.

15. When the similarity is < this threshold, the weight increases by 10% on the base value for every 0.05 decrease in similarity, up to a maximum of 0.3; when the similarity is ≥ this threshold, the weight decreases to 0.

15. Time series dimension: its standardized anomaly score = number of days of abnormal operation during the target household ÷ total number of monitoring days; the base weight is 0.2, and when the number of consecutive abnormal days is ≥7 days, the weight increases by 20% on the base value, with a maximum of 0.24; When the number of consecutive abnormal days is less than 3 days, the weight is reduced to 0.15; Operational behavior dimension: its standardized anomaly score = deviation rate of the target household's valve operation parameters from the historical model ÷ 300%; when the deviation rate is ≥300%, the score = 1; the basic weight is 0.2; Environmental impact dimension: its standardized anomaly score = 1 - actual ratio, where actual ratio = actual heat loss of target household ÷ theoretical heat loss value; the basic weight is 0.15; when the actual ratio < 0.6, the weight increases linearly according to the rule of basic weight × [1 + (0.6 - actual ratio) ÷ 0.6], and increases to a maximum of 0.3; Multi-source data cross-validation dimension: its standardized anomaly score = number of vacant related data items pointed to by the target household ÷ total number of related data items; the basic weight is 0.25; when the number of vacant related data items is ≥3, the weight is increased to the maximum value of 0.3; When the number of items is less than 2, the weight drops to 0.15; First, calculate the initial value of the contribution weight for each dimension separately according to the above rules. Then, calculate the sum of all the initial values ​​of contribution weight. If the sum of the initial values ​​is not equal to 1, then the final contribution weight of each dimension is equal to the initial value of the contribution weight of that dimension × the correction coefficient, and the correction coefficient is equal to 1 ÷ the sum of the initial values.

10. An application of a multi-terminal interconnected thermal control system based on NB-IoT on / off control valves in the intelligent analysis of vacant houses, characterized in that, The multi-terminal interconnected thermal control system described in any one of claims 1-9 enables intelligent identification of the vacancy status of target residential houses within the heating area, abnormal heat consumption monitoring, and dynamic control, specifically including the following steps: S1: Data Acquisition and Encrypted Transmission: Through the dual-mode sensing terminal corresponding to the target household, the conventional sensing data, concealed sensing data and valve status data of the heating pipe of the target household are collected. After being classified and encrypted by the encrypted communication unit, the data is transmitted to the edge computing gateway through the NB-IoT network. S2: Data preprocessing and feature extraction: The edge computing gateway parses and processes the received data through two threads, extracts the regular and abnormal feature parameters of the target residents, and synchronizes these two types of parameters to the cloud decision platform. At the same time, when receiving the heat camouflage behavior detection results from the cloud decision platform, it triggers the dual-mode sensing terminal to perform high-frequency data acquisition. S3: Multi-dimensional vacancy analysis and confidence calculation: The multi-dimensional collaborative detection engine of the cloud decision-making platform calculates the confidence of the vacancy status of the target household based on the target household's regular characteristic parameters, abnormal characteristic parameters, and cross-system related data through five parallel analyses: spatial correlation, time series, operational behavior, environmental impact, and multi-source data cross-validation. S4: Dynamic Response and Result Output: The cloud-based decision-making platform generates corresponding control commands based on the vacancy status confidence level and sends them to the target resident's NB-IoT on / off control valve through the edge computing gateway. At the same time, the vacancy status confidence level, control command type, and the basis for triggering the control command are synchronized to the hierarchical access interface, allowing the heating company and property management to obtain and execute subsequent management and control actions according to their permissions. Users can obtain their own house's anonymized data and vacancy status confidence level through the user query terminal.

Images