CN117541221A - An intelligent inspection method for Internet of Things equipment suitable for network-connected intelligent control of centralized heating - Google Patents

Abstract

The invention relates to an intelligent inspection method of an Internet of things device suitable for Internet intelligent control central heating, which is characterized in that fault information of the Internet of things device is analyzed from real historical data of a heating system, a mechanism and data hybrid drive is performed, a fault information base comprising connection faults, offset faults and mechanism faults is established, various faults to be detected are associated with sampling information, and a detection strategy is formulated; the method comprises the steps that an entry detection module detects the faults of the Internet of things equipment of the heating system in real time, and periodically detects whether the supply and demand of the system are matched according to a certain frequency, if the faults of the equipment and the supply and demand of the system are detected to be unmatched, a fault maintenance work order and a supply and demand matching adjustment work order are generated and sent to an operation and maintenance management platform; the operation and maintenance management platform timely maintains the physical network of the heating system according to the fault maintenance worksheet, adjusts worksheets to schedule and optimize the management control cloud platform according to supply and demand matching, realizes supply and demand matching, and saves energy to the greatest extent for the whole heating system on the premise of meeting the thermal comfort of users.

Description

An intelligent inspection method for Internet of Things equipment suitable for network-connected intelligent control of centralized heating

Technical field

The invention belongs to the technical field of centralized heating, and specifically relates to an intelligent inspection method of Internet of Things equipment suitable for network-connected intelligent control of centralized heating.

Background technique

At present, our country will increase its nationally determined contributions and adopt more powerful policies and measures. Among them, the carbon emissions from buildings nationwide account for 22% of the total emissions, while the carbon emissions generated by the heating system in the northern region account for 25% of the building emissions. Therefore, it is particularly important to realize building energy conservation and reduce building carbon emissions while promoting efficient, clean and stable central heating.

According to statistics [1] (Tsinghua University Building Energy Efficiency Research Center. China Building Energy Efficiency Annual Development Research Report. 2020[M]. Beijing: China Building Industry Press, 2020), the heating area in northern my country exceeds 20 billion square meters. At present, The average energy consumption is about 15kg of standard coal per square meter, which can theoretically reach 5kg of standard coal, which has huge energy-saving potential. Therefore, the optimization and intelligent adjustment of the central heating system can improve the energy efficiency of the system and ensure the safety and stable operation of the system.

The central heating system is a very complex multi-variable control system. Due to the large heating area, many influencing factors, strong internal correlation, long lag time and severe non-linearity, operation and maintenance managers have limitations in regulating based on engineering experience. , phenomena such as supply and demand mismatch and energy waste appear to varying degrees. With the development of information technology, intelligent control of heating systems has become one of the research hotspots. Its essence is a smart heating system that combines the physical heating system and the information system to integrate system operation and intelligent control. The basis of smart heating is to install IoT devices in the heating system to detect the operating data of the heating system and adjust the operating parameters of the heating system. Smart heating takes data analysis as the entry point, discovers, analyzes and solves problems through the system's operating data. Based on the thermal data of the heating system's operation, it forms an operation adjustment strategy through training mechanism models or artificial intelligence models to replace traditional operations. Managers rely on their own experience to roughly regulate the system, achieve refined control of the system, and achieve a balance between supply and demand. On the premise of satisfying users' thermal comfort, the system can operate efficiently, energy-saving, and stably. The prerequisite for smart heating systems to match supply and demand is reliable operation and regulation strategies, which requires sufficient reliable thermal data to train the model. However, various failures may occur during the operation of centralized heating IoT equipment, resulting in loss or abnormality of the final thermal data. This will lead to unreliability of the operation adjustment strategy, which will lead to a certain degree of supply and demand mismatch in the heating system and energy problems. waste phenomenon.

Relevant literature on fault detection is as follows:

Yu Dan et al. [2] disclosed the "Equipment Fault Detection and Analysis Method and System Based on the Internet of Things Platform" (CN114726750B), which performs fault detection on all terminal equipment and labels the fault detection information and uploads it to the Internet of Things platform. Cluster analysis is used to divide the information into several information sub-sets, extract common fault performance characteristic information, determine the common fault cause and fault occurrence range of the terminal equipment of each fault detection information sub-set, and send maintenance notification messages to maintenance personnel. The system can improve the speed and accuracy of fault detection of terminal equipment, and the processing and resolution rate of faulty terminal equipment.

Yang Heming et al. [3] disclosed the "Internet of Things Equipment Network Fault Analysis Method and Device" (CN113923101A), which determines the target faulty IoT device, obtains the network information of the target faulty IoT device, and determines the target faulty device according to the operator area to which it belongs. The failure rate of IoT equipment, the number of signal receptions of the target faulty equipment and the online rate during the same period are used to determine the cause of the faulty IoT equipment. It is helpful to improve the stability of the IoT system and achieve more intelligent fault analysis.

Huang Shengfeng et al. [4] disclosed "A smart gas valve IoT device and cloud platform system and method" (CN114811170A). A valve intelligent connection unit is set above the gas valve body. This system can realize data collection, data cleaning, data The analysis and data statistics functions can collect data regularly and monitor the abnormal status and data behavior of various valve components. It can solve the problems of existing pipelines buried underground and mostly avoiding residential areas, poor environment, traditional operation and maintenance management that relies on manual inspections and manual operations on a daily basis, long data acquisition cycle, inability to achieve real-time monitoring and analysis, and inability to prevent risks. .

Ding Yan et al. [5] disclosed "A Fault Diagnosis Method for Heat Exchanger Operation in Heating System" (CN112051082A), which uses the primary side flow rate, supply and return water temperature difference and secondary side flow rate of the heat exchanger during operation. and the temperature difference between supply and return water, develop a gradient boosting regression tree (GBRT) model to calculate the fixed heat transfer coefficient of the heat exchanger, and use the heat transfer principle of the heat exchanger to develop a white box model that reflects the performance of the heat exchanger to calculate the total heat transfer coefficient of the heat exchanger. , use the fixed heat transfer coefficient and the total heat transfer coefficient of the heat exchanger to calculate the generalized fouling thermal resistance of the heat exchanger, and determine whether the heat exchanger has a fault and needs to be cleaned or repaired. It can reflect the performance of the heat exchanger in a targeted manner, diagnose the operating status of the heat exchanger without affecting normal work, and ensure the operational stability of the heating system.

Gao Yang [6] disclosed "Fault Detection Method, Device, Equipment and Storage Medium of Temperature Sensor" (CN116242506A). When the door of the target compartment of the rail transit vehicle is closed and the temperature of the target compartment is stable, the target compartment is obtained. When the difference between the temperature data collected by the two temperature sensors to be detected and the compartment temperature data corresponding to the target compartment of the rail transit vehicle exceeds the first set deviation threshold, Based on the temperature data of the two temperature sensors to be detected and the cabin temperature data, the difference is calculated according to the preset comparison rules, and the target temperature sensor with offset fault among the two temperature sensors to be detected is determined based on the difference calculation result. Temperature sensors with offset faults in rail transit carriages can be accurately and quickly detected when the number of temperature sensors in the carriage is limited.

According to the existing literature on equipment fault detection, fault detection of IoT devices generally only considers the connection fault between the IoT device and the cloud platform, while equipment fault detection in the heating field generally only involves offset faults caused by equipment performance degradation. In fact, the thermal data collected by the central heating system may contain some information that violates the mechanism. At this time, the system has already experienced a mechanical failure. If it cannot be discovered and repaired in time, it will affect the training of the adjustment model and the normal operation of the system.

references:

[1] Jiang Yi. Tsinghua University Building Energy Efficiency Research Center. China Building Energy Efficiency Annual Development Research Report 2020 [M] Beijing: China Building Industry Press, 2020.

[2] Yu Dan, Lan Yuqing, Zhang Tenghuai, Ge Yutong, Wang Danxing. Equipment fault detection and analysis method and system based on Internet of Things platform [P]. Beijing: CN114726750B, 2023-01-13.

[3] Yang Heming, Chen Dequan, Li Pengfei, Lin Tao, Li Bin, Li Chaoming, Chen Huarong, Mai Hongyong, Liang Binghao, Zhang Liangdong, Wang Bin. Internet of Things equipment network fault analysis methods and devices [P]. Guangdong Province: CN113923101A, 2022-01- 11. [4] Huang Shengfeng, Jin Keyu, Shang Yulai, Jin Ruijian, Jin Kadi, Hu Daozhong, Zhang Hailan, Cai Yuefa, Xiao Qing, Huang Zhiqiang. A smart gas valve IoT device and cloud platform system and method [P]. Zhejiang Province: CN114811170A, 2022-07-29.

[5] Ding Yan, Liu Luheng, Su Hao, Tian Zhe. A fault diagnosis method for heat exchanger operation in heating systems [P]. Tianjin: CN112051082A, 2020-12-08.

[6] Gao Yang. Fault detection method, device, equipment and storage medium of temperature sensor [P]. Shanghai: CN116242506A, 2023-06-09.

Contents of the invention

The purpose of the present invention is to overcome the shortcomings of the existing technology and provide an intelligent inspection method for Internet of Things equipment suitable for network-connected intelligent control of centralized heating. On the one hand, it can comprehensively and real-time detect faults that occur during the operation of the system, and provide real-time detection and feedback. Timely repair of the operation and maintenance management platform can ensure sufficient and reliable thermal data for training the thermal model, which is conducive to efficient system operation and reliable model; on the other hand, supply and demand matching detection can detect whether the system's supply and demand match at a certain frequency based on the thermal model. Feedback is given to the operation and maintenance management platform for scheduling optimization to achieve supply and demand matching, maximizing energy savings for the entire heating system on the premise of meeting user thermal comfort.

The technical problems solved by the present invention are achieved through the following technical solutions:

An intelligent inspection method for Internet of Things equipment suitable for network-connected intelligent control of centralized heating. The method is characterized in that: the method relies on the Internet of Things equipment installed in the heating system to collect data. The Internet of Things equipment includes a thermal unit installed on the user side. The heat meters, flow sensors, water temperature sensors, valve controllers and IoT devices on the inlet supply and return water pipes collect thermal data during the operation of the heating system and synchronously transmit it to the host computer through the communication module and save it on the management and control cloud platform; The intelligent inspection system used in the above method includes a database module, a fault and sampling information association module, a threshold module, a detection module and an output module that are connected in sequence;

The steps of the method are:

1) Fault analysis: Enter the database module. The database module includes the equipment information database, fault information database and historical sampling database. The equipment information database includes the location, number and data collection of the heating system IoT equipment; the historical sampling database is from The management and control cloud platform reads historical sampling information, including all historical sampling information of the system's IoT devices; the fault information database is analyzed based on the historical sampling information in the historical information database, including connection faults, offset faults and mechanism faults, among which Connection failure means that the IoT device goes offline and loses connection with the management and control cloud platform; offset failure refers to the performance degradation of the device due to long-term use or other reasons; mechanism failure means that the information collected by the IoT device violates the mechanism;

The mechanism failure analysis is as follows:

The heat supply power of the heating system is determined by the supply and return water temperature and instantaneous flow rate. The formula is as follows:

P＝ρGc(T _g -T _h ) (1)

Among them: P is the thermal power supplied by the heating system, kW;

ρ represents the density of working fluid, kg/m ³ ;

G represents the instantaneous flow rate of working fluid, m ³ /s;

c represents the specific heat capacity of the working fluid, kJ/(kg·℃);

T _g is the water supply temperature;

T _h is the return water temperature, ℃;

The cumulative load of the heating system is the integral of the supplied heat power over time. The formula is as follows:

Among them: Q _accumulation is the cumulative load of the heating system, kJ;

τ is time, s;

During the operation of the heating system, the water supply temperature is greater than the return water temperature. According to formula (2), it can be seen that the cumulative load of the heating system increases with time. Similarly, according to formula (3), it can be seen that the cumulative flow of the heating system increases with time.

Among them: G _accumulation is the cumulative flow rate of the heating system, m ³ ;

To sum up, the analysis shows the mechanism failure:

Fault 1: The return water temperature is greater than the water supply temperature;

Fault 2: Accumulated heat or accumulated flow rate decreases;

2) Establish a correlation between real-time sampling information and three types of faults; enter the fault and information correlation module, and when performing fault detection, obtain real-time sampling information from the management and control cloud platform. The real-time sampling information is all data in the current time window, including two Two types - instantaneous data and cumulative data. Instantaneous data is data that does not accumulate on the original basis as the system runs. It fluctuates with the system's operation. Accumulated data is data that accumulates on the original basis as the system runs. It increases or remains unchanged as the system runs. , the current time window is a time period whose end time is the current time and the time span is equal to the preset time length; if the data amount of the current time window is less than the set data amount, it is determined that a connection failure has occurred; only instantaneous data will have offset Fault, if the amount of instantaneous data exceeding the offset fault threshold in the current time window is greater than the set data amount, it is determined that an offset fault has occurred; if the amount of data that violates the mechanism in the current time window is greater than the set data amount, it is determined that a mechanism failure has occurred;

3) Determine the offset fault detection threshold and the supply and demand matching detection threshold: the offset fault detection threshold is combined with the system normal operation reference value given by the operation and maintenance manager to obtain the first threshold interval; use the isolated forest algorithm to remove anomalies from the historical sampling library The minimum and maximum values after the value take into account the margin to form the second threshold interval; the first threshold interval and the second threshold interval are intersected to obtain the third threshold interval, where the first threshold interval is the theoretical threshold interval of the system's normal operation reference value , represents the absolute upper and lower limits of the operation of the entire system; the second threshold interval is the actual value of the operation of each unit in the system, which is targeted; the third threshold interval is the intersection of the first threshold interval and the second threshold interval, taking into account the normal operation of the entire system The theoretical value and the actual value of the operation of different units of the system are the intervals for ultimately detecting whether there is an offset fault;

4) Fault and supply and demand matching detection: Enter the detection module to complete fault detection and supply and demand matching detection.

Fault detection: Read real-time sampling information from the management and control cloud platform, obtain the operating data of the current time window, and detect faults in sequence according to the order of connection fault, offset fault, and mechanism fault. If an equipment fault is detected, an alarm will prompt the location of the fault and the fault location. The type and reason, the fault detection frequency setting is related to the preset time length, specifically how long the preset time is and how often it is detected;

Supply and demand matching detection: Read real-time sampling information from the management and control cloud platform, obtain the operating data of the current time window, determine the supply and demand matching detection threshold interval of parameters based on the peripheral module thermal model, and calculate the offset fault detection threshold interval within the current time window. Parameter average value is used to determine the relationship between the average value and the supply and demand matching detection threshold interval. When the parameter average value is within the supply and demand matching detection threshold interval, it is deemed that supply and demand match; when the parameter average value is outside the supply and demand matching detection threshold interval, it is determined that supply and demand do not match. , because the heating system should not be adjusted too frequently, and the adjustment frequencies of different systems are different, supply and demand matching testing should be carried out regularly according to the system adjustment frequency;

5) Output fault maintenance work orders and supply and demand matching adjustment work orders: When the fault detection part of the detection module detects a fault, an alarm prompts, and the output module is entered to generate a fault maintenance work order and sent to the operation and maintenance management platform. The operation and maintenance management platform performs maintenance according to the fault Timely repair and maintenance of the physical network of the heating system to ensure sufficient and reliable thermal data for training the thermal model, which is conducive to efficient operation of the system; when the supply and demand matching detection part of the detection module is regularly performed according to the system adjustment frequency, if the system supply and demand is detected If there is no match, an alarm will be prompted. Enter the output module to generate a supply and demand matching adjustment work order and send it to the operation and maintenance management platform. The operation and maintenance management platform will optimize the scheduling of the management and control cloud platform based on the supply and demand matching adjustment work order to achieve supply and demand matching and meet the user's thermal comfort. Under the premise, the entire heating system can save energy to the greatest extent.

Moreover, the specific steps of step 3) are:

Offset fault detection threshold:

(1) Combined with the normal operation reference value of the system given by the operation and maintenance manager, the first threshold interval [x _1min , x _1max ] is obtained;

(2) Read the historical sampling information from the historical sampling library, obtain historical instantaneous data, use the isolated forest algorithm to detect and remove outliers, and obtain the anomaly-free data set X={x ₁ , x ₂ , x ₃ ,..., x _p } ;

The principle of the isolation forest algorithm is as follows:

A. Sampling in the training set and taking out a subset containing n samples;

B. Each time a certain threshold is randomly selected for division, 2 child nodes are generated. When the number of samples in a node is 1 or the sample characteristics in the same node are the same or the specified depth is reached, the division is stopped. At this time, 1 itree is generated. ;

C. Repeat the operations of A and B to generate m itree;

D. Calculate the depth of the given sample in m itree;

E. Determine whether it is an outlier based on the outlier scoring function. The scoring function is shown in Equation (4);

Among them: h(x) represents the average depth of sample x in m itree;

E represents mathematical expectation;

c(n) represents the average path length of unsuccessful binary sorting tree search;

(3) The second threshold interval [x _2min ,x _2max ] is obtained by considering the margin from the maximum value of the anomaly removal data set. The formula is as follows:

x _2min = (minX)·ξ ^-1 (5)

x _2max =(maxX)·ξ (6)

Among them: ξ is the margin factor, which is determined according to the length of the first threshold interval. The recommended value range is 1.1~1.5;

(4) The third threshold interval [x _3min ,x _3max ] is obtained by taking the intersection of the first threshold interval and the second threshold interval, which is the interval finally used for offset fault detection. The formula is as follows:

[x _3min ,x _3max ]＝[x _1min ,x _1max ]∩[x _2min ,x _2max ] (7)

Supply and demand matching detection threshold: determined based on the thermal model of the peripheral module. When the parameters are within the threshold range, supply and demand are deemed to match; when the parameters are outside the threshold range, supply and demand are deemed to be mismatched. The thermal model is determined from the historical sampling library. The data is obtained through data processing, data segmentation, model training, and model verification.

The advantages and beneficial effects of the present invention are:

1. The present invention is an intelligent inspection method for Internet of Things equipment suitable for network-connected intelligent control of centralized heating. On the one hand, it comprehensively detects faults that occur during the operation of the system in real time, detects them in real time and feeds them back to the operation and maintenance management platform for timely repair, which can Ensure sufficient and reliable thermal data is used to train the thermal model, which is conducive to efficient operation of the system and reliability of the model; on the other hand, supply and demand matching detection can detect whether the system's supply and demand match at a certain frequency based on the thermal model, and feedback to the operation and maintenance management platform for scheduling optimization. Achieve supply and demand matching, and maximize energy saving for the entire heating system on the premise of satisfying user thermal comfort.

2. This invention is based on the network-connected intelligent control of the heating system, analyzes the fault information of the Internet of Things equipment from the real historical data of the heating system, drives the mechanism and data hybridly, and establishes a fault information database including connection faults, offset faults, and mechanism faults. , associate various faults that need to be detected with sampling information, and formulate detection strategies; enter the detection module to detect IoT equipment faults in the heating system in real time, and regularly detect whether the system supply and demand match a certain frequency. If an equipment fault is detected and the system supply and demand are not matched, Match, generate fault maintenance work orders and supply and demand matching adjustment work orders and send them to the operation and maintenance management platform; the operation and maintenance management platform promptly repairs and maintains the heating system physical network according to the fault maintenance work orders, and schedules the management and control cloud platform according to the supply and demand matching adjustment work orders. Optimize to achieve supply and demand matching, and maximize energy saving for the entire heating system on the premise of meeting user thermal comfort.

3. The database module of the present invention analyzes fault information based on the real historical data of the heating system operation. It not only considers connection faults and offset faults, but also considers mechanism faults that violate the mechanism. The mechanism and data are mixed and driven to comprehensively and targetedly analyze the fault information. Discover fault information and establish a fault information database. The detection strategy can be comprehensively formulated based on the characteristics of the actual system, and the corresponding fault causes can be output to ensure that the detected faults are comprehensive and clear, improving maintenance efficiency and reducing maintenance costs.

4. The fault and sampling information correlation module of the present invention establishes the correlation between real-time sampling information and various faults. The actual data collection frequency of IoT devices in heating systems is different. Systems with high collection frequency can reach minute-by-minute or even second-by-second levels. When there is a problem with a certain data, it does not affect the system operation and data training. Therefore, this module determines whether there is a fault based on the comparison between the number of fault conditions in the current time window and the set number. The fault is determined only when there are multiple problematic data. Filtering out basically harmless fault types can avoid frequent manual attention on system changes.

5. The threshold module of the present invention is divided into two parts: offset fault detection threshold and supply and demand matching detection threshold. Offset fault detection threshold: taking into account the theoretical value for normal operation of the entire system and the actual value for the operation of different units of the system. Supply and demand matching detection threshold: It can be used to detect whether the system's supply and demand match. It can adjust work orders according to the supply and demand matching for optimal scheduling, achieve supply and demand matching, and maximize energy saving for the system on the premise of satisfying user thermal comfort.

6. The detection module of the present invention is divided into two parts: fault detection and supply and demand matching detection. In the fault detection part, the detection frequency is set according to the preset time period during detection, which can monitor system equipment operation faults in real time. If there is a faulty device, an alarm will be prompted, and a fault maintenance work order will be output and sent to the operation and maintenance management platform to repair the faulty device in a timely manner to ensure stable system operation, complete and reliable operation data, and reliable adjustment models, so as to meet the premise of user thermal comfort. Maximum energy savings.

7. The detection module of the present invention is divided into two parts: fault detection and supply and demand matching detection. The supply and demand matching detection part is tested according to the system adjustment frequency, and can detect whether the supply and demand match each location in the system. If there is a mismatch between supply and demand, an alarm will be prompted, and a supply and demand matching adjustment work order will be output and sent to the operation and maintenance management platform to optimize scheduling, achieve supply and demand matching, and maximize energy savings for the system on the premise of satisfying user thermal comfort.

8. This method is easy to implement, has a wide range of uses, is highly operable, and has controllable costs. The heating pipe network does not require large-scale changes and does not involve civil engineering or other modifications.

9. The intelligent inspection system and method proposed by the present invention based on the network-connected intelligent control IoT equipment of the heating system have good extrapolation properties, and the results and methods can be applied to scenarios with similar characteristics.

10. The intelligent inspection system and method proposed by the present invention based on network-connected intelligent control of IoT equipment in the heating system integrates practicality, applicability, advancement and demonstration, and is useful for low-carbon construction in the construction field in the context of achieving dual carbon goals. Efficient and clean heating is of great significance.

Description of drawings

Figure 1 is a technical roadmap of the present invention;

Figure 2 is a schematic diagram of the central heating system pipe network of the present invention;

Figure 3 is a schematic diagram of the heating system according to the embodiment of the present invention;

Figure 4 is a schematic diagram of the historical sampling library of the present invention;

Figure 5 is a schematic diagram of a fault case of the present invention;

Figure 6 is a fault detection flow chart of the present invention;

Figure 7 is a flow chart of supply and demand matching detection according to the present invention;

Figure 8 is a fault maintenance work order diagram of the present invention;

Figure 9 is a work order diagram of supply and demand matching adjustment according to the present invention.

Explanation of reference signs

1-boiler; 2-water temperature sensor; 3-water pump; 4-flow sensor; 5-heat meter; 6-communication module; 7-valve controller; 8-building complex; 9-room temperature sensor.

Detailed ways

The present invention will be further described in detail below through specific examples. The following examples are only descriptive, not restrictive, and cannot be used to limit the scope of the present invention.

The present invention provides an intelligent inspection method for Internet of Things equipment suitable for network-connected intelligent control of centralized heating. It is applied to a centralized heating system of a university as shown in Figure 2. The heating system includes a boiler 1 and a device connected to the boiler. Water supply pipe and return water pipe. The water supply pipe is connected to a water temperature sensor 2, a flow sensor 4 and a building group 8 in sequence. The return water pipe is connected to a temperature sensor, a water pump 3 and a building group 8 in sequence. The water supply pipe and the return water pipe are connected to The temperature sensors on the water pipes are all connected to the heat meter 5. The heat meters are respectively connected to the valve controller 7 and the flow sensor 4. The valve controller is connected to the communication module 6. The communication module is connected to the meteorological module. A room temperature sensor 9 is provided in the building complex 8. The room temperature sensor 9 and the valve controller 7 are both wirelessly connected to the external monitoring platform through the communication module.

The heating system currently supplies a heating area of 267,975 square meters, including high areas and low areas, which provide heat to different building areas. A total of 5 gas boilers provide direct supply to the low and high areas. The heating buildings include scientific research office buildings, experimental factories, student dormitories, teaching buildings, canteens and many other types of users. The heating system was renovated in 2019 and consists of three levels, namely the energy center, thermal inlet and end room. The energy center includes two parts, the low zone and the high zone. The transformed thermal inlets are all in the low zone, with a total of 30 indivual. A water temperature sensor, flow sensor, and heat meter are installed in the energy center to detect the supply and return water temperature, instantaneous flow, cumulative flow, and accumulated heat of the energy center; a water temperature sensor, flow sensor, heat meter, and valve controller are installed at the thermal inlet. , used to detect the supply and return water temperature, instantaneous flow, cumulative flow, cumulative heat and valve position of the thermal inlet; a room temperature sensor is installed in the end room to detect the room temperature of the end room. The ends of the source network of the heating system are equipped with communication modules connected to the cloud platform. The actual operating data involved in the present invention are obtained from the data collection modules of the energy center, thermal entrance, and end room, as shown in Figure 1. The data collection frequency of the energy center and thermal entrance is once every 6 minutes, and the data collection frequency of the end room is every 10 minutes.

As shown in Figure 1, the following mainly takes this heating system as an example to illustrate the specific implementation and beneficial effects of this patented technology. The heating system of this embodiment is shown in Figure 3 and has undergone intelligent transformation in 2019. The operating parameters of the Internet of Things equipment monitoring system are installed at the three levels of energy center, thermal entrance and terminal room, and uploaded to the cloud control platform. Among them, a) is the transformation diagram of the energy center, b) is the transformation diagram of the thermal entrance and terminal rooms, c) is the thermal entrance and building distribution diagram of the low area of the energy center, d) is the smart heating network cloud control platform, and e) is the operation and maintenance management platform.

Specifically, it includes the following steps:

(1) Enter the database module, which includes three parts: equipment information database, fault information database and historical sampling database.

The device information database includes information such as the location, number, and data collected of the system's IoT devices.

The historical sampling library is read from the management and control cloud platform and contains the sampling information of all historical heating seasons of the system's IoT devices. See Figure 4. The system is divided into three levels: energy center, thermal entrance, and end room. , where Figure 4a) is the historical heating season sampling information of the energy center, Figure 4b) is the historical heating season sampling information of the thermal inlet, and Figure 4c) is the historical heating season sampling information of the end room.

The fault information database is analyzed based on the historical sampling information in the historical information database, including connection faults, offset faults and mechanism faults. After analysis, the fault statistics of the system are shown in Table 1.

There are 6 types of faults, of which 101 faults do not affect the normal use of data and should be filtered during detection. The schematic diagram of the six fault cases is shown in Figure 5. A to f in Figure 5 correspond to the fault numbers in Table 1 in sequence.

Table 1 Fault statistics table

(2) Enter the fault and information correlation module. When performing fault detection, real-time sampling information is obtained from the management and control cloud platform. Real-time sampling information is all data in the current time window, including two types - instantaneous data and cumulative data. Instantaneous data is data that does not accumulate on the original basis with the operation of the system. It fluctuates with the operation of the system, such as supply and return water temperature, instantaneous flow rate, etc. Accumulated data is data accumulated on the original basis as the system operates. It increases or remains unchanged as the system operates, such as accumulated flow, accumulated heat, etc. The current time window is a time period in which the end time is the current time and the time span is equal to the preset time length.

For the 6 types of faults analyzed above:

If the data amount of the current time window is less than the set data amount, it is determined that a connection failure 000 has occurred;

If the instantaneous data amount exceeding the offset fault threshold in the current time window is greater than the set data amount and the data is not 0, it is determined that an offset fault 001 has occurred;

If the amount of data with 0 data in the current time window is greater than the set data amount, it is determined that a mechanical fault 010 has occurred;

If the accumulated data difference in the current time window is less than 0 and the amount of data is greater than the set amount of data, it is determined that a mechanical fault 011 has occurred;

If the amount of data that the return water temperature in the current time window is greater than the water supply temperature is greater than the set data amount, it is determined that a mechanical failure has occurred 100;

By comparing the amount of data that meets the problem conditions with the set amount of data, the mechanism fault 101 can be filtered out.

3) Enter the threshold module.

Offset fault detection threshold: The first threshold interval is obtained based on the normal operation reference value of the system given by the operation and maintenance manager; the minimum and maximum values after removing outliers using the isolated forest algorithm from the historical sampling library are considered to form the second threshold. Threshold interval; the intersection of the first threshold interval and the second threshold interval obtains the third threshold interval. The offset fault detection thresholds of some parameters of the system are shown in Table 2. The margin factor settings of each parameter are shown in Table 3.

Table 2 Partial parameter threshold interval table

Table 3 Margin factors of each parameter

parameter margin factor Energy center water supply temperature 1.1 Energy center return water temperature 1.1 Energy Center Instantaneous Traffic 1.2 Thermal inlet water temperature 1.1 Thermal inlet return water temperature 1.1 Thermal inlet instantaneous flow 1.2 Thermal inlet valve position 1.4 End room room temperature 1.3

Supply and demand matching detection threshold: determined based on the thermal model of the peripheral module. When the parameters are outside the threshold range, it is determined that supply and demand do not match. Among them, the thermal model is derived from the data in the historical sampling database through data processing, data segmentation, model training, and model verification. In this system, the thermal model refers to the energy center water supply temperature model and the thermal inlet valve position model.

(4) Enter the detection module.

Fault detection part: The detection flow chart is shown in Figure 6.

Read real-time sampling information from the management and control cloud platform, and obtain the operating data of the current time window with a set time length of 1 hour, including the energy center water supply temperature, return water temperature, instantaneous flow, accumulated flow, accumulated heat, thermal inlet water supply temperature, return water temperature, etc. Water temperature, instantaneous flow, cumulative flow, cumulative heat, valve position, and terminal room room temperature.

First, detect connection failures. If the data amount of the current time window is less than the set data amount, it is determined that a connection failure 000 has occurred. If all equipment in the energy center, thermal entrance, and end room have a connection failure 000, an alarm will occur - all equipment is not online, and the system may be lost.

Second, detect offset faults. For all instantaneous data (including energy center water supply temperature, return water temperature, instantaneous flow, thermal inlet water supply temperature, return water temperature, instantaneous flow, valve position, terminal room room temperature), if the current time window exceeds the data of the offset fault threshold If the amount is greater than the set data amount and the data is not 0, it is determined that offset fault 001 has occurred.

Finally, detect mechanical failures. If the amount of data with 0 data in the current time window is greater than the set data amount, it is determined that a mechanism fault 010 has occurred; if the difference in the accumulated data in the current time window (including accumulated flow, accumulated heat of the energy center, accumulated flow, and accumulated heat of the thermal inlet) is less than 0 If the amount of data is greater than the set data amount, it is determined that a mechanism failure 011 has occurred; if the return water temperature of the energy center and thermal inlet in the current time window is greater than the water supply temperature and the amount of data is greater than the set data amount, it is determined that a mechanism failure 100 has occurred;

If none of the above conditions are met, the device is considered normal.

The fault detection frequency is once every 1 hour.

Supply and demand matching detection part: The detection flow chart is shown in Figure 7.

Read real-time sampling information from the management and control cloud platform, and obtain the operating data of the current time window with a set time length of 1 hour, including the energy center water supply temperature and thermal inlet valve position.

Calculate the average value of the energy center's water supply temperature and thermal inlet valve position in the offset fault detection threshold interval within the current time window, and determine the relationship between the parameter average and the supply and demand matching detection threshold interval.

If the average parameter value is within the supply and demand matching detection threshold range, supply and demand are deemed to match.

If the average parameter value is outside the supply and demand matching detection threshold range, it is determined that supply and demand do not match.

The supply and demand matching detection frequency is once every 1d.

(5) Enter the output module.

If the fault detection part of the detection module detects a fault, a fault maintenance work order is generated to the operation and maintenance management platform, and the operation and maintenance management platform repairs and maintains the physical network of the heating system according to the fault maintenance work order. It is conducive to the efficient operation of the system, ensuring the integrity and reliability of data, and ensuring the reliability of the thermal model, thereby achieving supply and demand matching.

The fault maintenance work orders for 2022-01-31 and 2023-07-01 are shown in Figure 8, as shown in Figure 8a) and b) respectively.

If the supply and demand matching detection part of the detection module detects a mismatch between system supply and demand, a supply and demand matching adjustment work order is generated. The operation and maintenance management platform optimizes the scheduling of the management and control cloud platform based on the supply and demand matching and adjustment work order. The management and control cloud platform can control the operation of the physical network of the heating system. . Achieve supply and demand matching, and maximize energy saving for the system on the premise of satisfying users' thermal comfort.

The supply and demand matching adjustment work orders for 2023-11-11 and 2023-11-25 are shown in Figure 9, as shown in Figure 9a) and b) respectively.

Although the embodiments and drawings of the present invention have been disclosed for illustrative purposes, those skilled in the art will understand that various substitutions, changes and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. , therefore, the scope of the present invention is not limited to the contents disclosed in the embodiments and drawings.

Claims (2)

1. An intelligent inspection method for Internet of Things equipment suitable for network-connected intelligent control of centralized heating. It is characterized in that: the method relies on the Internet of Things equipment installed in the heating system to collect data. The Internet of Things equipment includes installation on the user's computer. The heat meter, flow sensor, water temperature sensor, valve controller, and IoT equipment on the side heat inlet supply and return pipes collect thermal data during the operation of the heating system and synchronously transmit it to the host computer through the communication module and save it on the management and control cloud platform ; The intelligent inspection system used in the method includes a database module, a fault and sampling information association module, a threshold module, a detection module and an output module that are connected in sequence; The steps of the method are: 1) Fault analysis: Enter the database module. The database module includes the equipment information database, fault information database and historical sampling database. The equipment information database includes the location, number and data collection of the heating system IoT equipment; the historical sampling database is from The management and control cloud platform reads historical sampling information, including all historical sampling information of the system's IoT devices; the fault information database is analyzed based on the historical sampling information in the historical information database, including connection faults, offset faults and mechanism faults, among which Connection failure means that the IoT device goes offline and loses connection with the management and control cloud platform; offset failure refers to the performance degradation of the device due to long-term use or other reasons; mechanism failure means that the information collected by the IoT device violates the mechanism; The mechanism failure analysis is as follows: The heat supply power of the heating system is determined by the supply and return water temperature and instantaneous flow rate. The formula is as follows: P＝ρGc(T _g -T _h ) (1) Among them: P is the thermal power supplied by the heating system, kW; ρ represents the density of working fluid, kg/m ³ ; G represents the instantaneous flow rate of working fluid, m ³ /s; c represents the specific heat capacity of the working fluid, kJ/(kg·℃); T _g is the water supply temperature; T _h is the return water temperature, ℃; The cumulative load of the heating system is the integral of the supplied heat power over time. The formula is as follows: Among them: Q _accumulation is the cumulative load of the heating system, kJ; τ is time, s; During the operation of the heating system, the water supply temperature is greater than the return water temperature. According to formula (2), it can be seen that the cumulative load of the heating system increases with time. Similarly, according to formula (3), it can be seen that the cumulative flow of the heating system increases with time. Among them: G _accumulation is the cumulative flow rate of the heating system, m ³ ; To sum up, the analysis shows the mechanism failure: Fault 1: The return water temperature is greater than the water supply temperature; Fault 2: Accumulated heat or accumulated flow rate decreases; 2) Establish a correlation between real-time sampling information and three types of faults; enter the fault and information correlation module, and when performing fault detection, obtain real-time sampling information from the management and control cloud platform. The real-time sampling information is all data in the current time window, including two Two types - instantaneous data and cumulative data. Instantaneous data is data that does not accumulate on the original basis as the system runs. It fluctuates with the system's operation. Accumulated data is data that accumulates on the original basis as the system runs. It increases or remains unchanged as the system runs. , the current time window is a time period whose end time is the current time and the time span is equal to the preset time length; if the data amount of the current time window is less than the set data amount, it is determined that a connection failure has occurred; only instantaneous data will have offset Fault, if the amount of instantaneous data exceeding the offset fault threshold in the current time window is greater than the set data amount, it is determined that an offset fault has occurred; if the amount of data that violates the mechanism in the current time window is greater than the set data amount, it is determined that a mechanism failure has occurred; 3) Determine the offset fault detection threshold and the supply and demand matching detection threshold: the offset fault detection threshold is combined with the system normal operation reference value given by the operation and maintenance manager to obtain the first threshold interval; use the isolated forest algorithm to remove anomalies from the historical sampling library The minimum and maximum values after the value take into account the margin to form the second threshold interval; the first threshold interval and the second threshold interval are intersected to obtain the third threshold interval, where the first threshold interval is the theoretical threshold interval of the system's normal operation reference value , represents the absolute upper and lower limits of the operation of the entire system; the second threshold interval is the actual value of the operation of each unit in the system, which is targeted; the third threshold interval is the intersection of the first threshold interval and the second threshold interval, taking into account the normal operation of the entire system The theoretical value and the actual value of the operation of different units of the system are the intervals for ultimately detecting whether there is an offset fault; 4) Fault and supply and demand matching detection: Enter the detection module to complete fault detection and supply and demand matching detection. Fault detection: Read real-time sampling information from the management and control cloud platform, obtain the operating data of the current time window, and detect faults in sequence according to the order of connection fault, offset fault, and mechanism fault. If an equipment fault is detected, an alarm will prompt the location of the fault and the fault location. The type and reason, the fault detection frequency setting is related to the preset time length, specifically how long the preset time is and how often it is detected; Supply and demand matching detection: Read real-time sampling information from the management and control cloud platform, obtain the operating data of the current time window, determine the supply and demand matching detection threshold interval of parameters based on the peripheral module thermal model, and calculate the offset fault detection threshold interval within the current time window. Parameter average value is used to determine the relationship between the average value and the supply and demand matching detection threshold interval. When the parameter average value is within the supply and demand matching detection threshold interval, it is deemed that supply and demand match; when the parameter average value is outside the supply and demand matching detection threshold interval, it is determined that supply and demand do not match. , because the heating system should not be adjusted too frequently, and the adjustment frequencies of different systems are different, supply and demand matching testing should be carried out regularly according to the system adjustment frequency; 5) Output fault maintenance work orders and supply and demand matching adjustment work orders: When the fault detection part of the detection module detects a fault, an alarm prompts, and the output module is entered to generate a fault maintenance work order and sent to the operation and maintenance management platform. The operation and maintenance management platform performs maintenance according to the fault Timely repair and maintenance of the physical network of the heating system to ensure sufficient and reliable thermal data for training the thermal model, which is conducive to efficient operation of the system; when the supply and demand matching detection part of the detection module is regularly performed according to the system adjustment frequency, if the system supply and demand is detected If there is no match, an alarm will be prompted. Enter the output module to generate a supply and demand matching adjustment work order and send it to the operation and maintenance management platform. The operation and maintenance management platform will optimize the scheduling of the management and control cloud platform based on the supply and demand matching adjustment work order to achieve supply and demand matching and meet the user's thermal comfort. Under the premise, the entire heating system can save energy to the greatest extent. 2. The intelligent inspection method of Internet of Things equipment suitable for network-connected intelligent control of centralized heating according to claim 1, characterized in that: the specific steps of step 3) are: Offset fault detection threshold: (1) Combined with the normal operation reference value of the system given by the operation and maintenance manager, the first threshold interval [x _1min , x _1max ] is obtained; (2) Read the historical sampling information from the historical sampling library, obtain historical instantaneous data, use the isolated forest algorithm to detect and remove outliers, and obtain the anomaly-free data set X={x ₁ , x ₂ , x ₃ ,..., x _p } ; The principle of the isolation forest algorithm is as follows: A. Sampling in the training set and taking out a subset containing n samples; B. Each time a certain threshold is randomly selected for division, 2 child nodes are generated. When the number of samples in a node is 1 or the sample characteristics in the same node are the same or the specified depth is reached, the division is stopped. At this time, 1 itree is generated. ; C. Repeat the operations of A and B to generate m itree; D. Calculate the depth of the given sample in m itree; E. Determine whether it is an outlier based on the outlier scoring function. The scoring function is shown in Equation (4); Among them: h(x) represents the average depth of sample x in m itree; E represents mathematical expectation; c(n) represents the average path length of unsuccessful binary sorting tree search; (3) The second threshold interval [x _2min ,x _2max ] is obtained by considering the margin from the maximum value of the anomaly removal data set. The formula is as follows: x _2min = (min X)·ξ ^-1 (5) x _2max = (max X)·ξ (6) Among them: ξ is the margin factor, which is determined according to the length of the first threshold interval. The recommended value range is 1.1~1.5; (4) The third threshold interval [x _3min ,x _3max ] is obtained by taking the intersection of the first threshold interval and the second threshold interval, which is the interval finally used for offset fault detection. The formula is as follows: [x _3min ,x _3max ]＝[x _1min ,x _1max ]∩[x _2min ,x _2max ] (7) Supply and demand matching detection threshold: determined based on the thermal model of the peripheral module. When the parameters are within the threshold range, supply and demand are deemed to match; when the parameters are outside the threshold range, supply and demand are deemed to be mismatched. The thermal model is determined from the historical sampling library. The data is obtained through data processing, data segmentation, model training, and model verification.