Building facility intelligent monitoring method and system based on internet of things

Abstract

The application discloses a kind of building facility intelligent monitoring method and system based on Internet of Things, and the abnormal mode of facility response is formed by periodic deviation detection and persistence deviation analysis to facility operation data and control instruction response data;Based on the time sequence of abnormal event in turn relationship, the fault cause-effect chain is constructed, and the cause-effect correlation between facilities is extracted to form the correlation matrix of facility;According to the correlation matrix, deploy distributed monitoring nodes, and construct the health degree distribution of facility by linkage collection and coordination analysis;The deterioration rate analysis and spread detection are carried out on health degree, high-risk facility group is identified, and hierarchical monitoring task sequence is generated;The remaining safe operation cycle of high-risk facility is evaluated to determine the warning level parameter;Based on warning level, the sensitivity classification and time-sharing threshold configuration of monitoring node are carried out, and dynamic alarm threshold and monitoring benchmark are generated;Facility state monitoring report is output by fusing monitoring benchmark and health degree distribution, and early warning of fault risk is realized.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) technology, and in particular to an IoT-based intelligent monitoring method and system for building facilities. Background Technology

[0002] With the development of IoT technology, the operation monitoring of building electromechanical facilities is gradually shifting from manual inspection to automated data collection. Currently, most buildings have deployed sensors and data acquisition systems that can acquire real-time operating parameters such as current, temperature, pressure, and vibration of facilities like air conditioners, water pumps, fans, and power distribution cabinets, and trigger alarms for exceeding preset thresholds. However, existing monitoring systems mainly monitor individual facilities independently, using uniform fixed thresholds as the basis for alarms. They fail to consider the operational differences of facilities during different load periods and lack the ability to analyze the operational correlations between facilities, resulting in monitoring effectiveness that is insufficient to meet the needs of refined operation and maintenance.

[0003] In actual operation, building electromechanical facilities do not operate in isolation, but rather form an interconnected system through pipelines, circuits, and control signals. When an anomaly occurs in one facility, it often propagates along physical connections or control links to other facilities, creating a cascading effect. Existing monitoring technologies struggle to identify such fault propagation relationships between facilities, cannot distinguish between the source of the fault and the affected facilities, and cannot dynamically adjust monitoring strategies based on the actual risk status of each facility. Furthermore, the even distribution of monitoring resources fails to reflect the differences in facility importance and risk levels; monitoring accuracy for critical facilities is insufficient, while monitoring of general facilities is redundant, resulting in overall monitoring efficiency that needs improvement. Summary of the Invention

[0004] This invention discloses an intelligent monitoring method and system for building facilities based on the Internet of Things. By detecting periodic offsets in facility operation data to identify abnormal response patterns, analyzing the temporal causal relationships of abnormal events to construct a facility correlation matrix, deploying distributed monitoring nodes based on the correlation matrix and assessing the distribution of facility health, analyzing the deterioration trend of health to identify high-risk facility groups, assessing the remaining safe operating cycle to determine the warning level, performing adaptive threshold calibration on monitoring nodes based on the warning level, and integrating the health distribution to output a facility status monitoring report, the invention achieves intelligent monitoring and differentiated management of building electromechanical facilities.

[0005] The first aspect of this invention proposes an intelligent monitoring method for building facilities based on the Internet of Things, comprising the following steps:

[0006] Collect sensor operation data and control command response data of building electromechanical facilities, and perform operation cycle offset detection on the operation data and the command response data to form an abnormal facility response mode;

[0007] Extract the time sequence of abnormal events from the abnormal response pattern of the facility, arrange the time sequence of the abnormal events in chronological order to construct a fault time sequence causal chain, and form a facility association matrix based on the fault time sequence causal chain.

[0008] Based on the facility association matrix, a distributed monitoring node is established. Based on the distributed monitoring node, the working condition linkage of the building electromechanical facilities is collected to generate a facility status feature set. The facility status feature set is used to construct the facility health distribution.

[0009] The process involves analyzing the facility deterioration rate within the facility health distribution to determine the deterioration rate, grouping building electromechanical facilities into high-risk and low-risk groups based on the deterioration rate, and dynamically allocating monitoring resources to these groups to construct a tiered monitoring task sequence. The grouping of building electromechanical facilities into high-risk and low-risk groups based on the deterioration rate includes: time-series tracking of the deterioration rate to form a deterioration rate curve; analyzing the deterioration spread of adjacent facilities using the deterioration rate curve to identify deterioration spread source facilities; determining facility risk classification identifiers based on the deterioration spread source facilities and the deterioration rate; and identifying high-risk and low-risk groups based on the facility risk classification identifiers.

[0010] Based on the hierarchical monitoring task sequence, the remaining safe operating cycle of the high-risk group is assessed to identify fault risk points, and the fault risk points are used to determine the early warning level parameters.

[0011] Based on the warning level parameters, the distributed monitoring nodes are adaptively calibrated to generate a monitoring benchmark. Based on the monitoring benchmark and the facility health distribution, a facility status monitoring report is output.

[0012] Optionally, the step of detecting operational cycle offsets between the operational data and the instruction response data to form a facility response anomaly pattern includes: extracting periodic features from the operational data to obtain the actual operational cycle; comparing the actual operational cycle with a preset operational cycle in the instruction response data to calculate the cycle offset; performing positive and negative offset asymmetry analysis on the cycle offset to identify persistent offset events; and forming a facility response anomaly pattern based on the persistent offset events.

[0013] Optionally, the step of arranging the abnormal event sequence in chronological order to construct a fault time-series causal chain includes: timestamping the abnormal event sequence to obtain an event occurrence sequence; arranging the events in chronological order to form an ordered event chain; performing cross-level event progression association on the ordered event chain to identify indirect related events across devices and systems; and constructing a fault time-series causal chain based on the indirect related events and the ordered event chain.

[0014] Optionally, the step of collecting data on the working conditions of building electromechanical facilities based on the distributed monitoring nodes to generate a facility status feature set includes: identifying the linkage triggering relationship between building electromechanical facilities based on the distributed monitoring nodes to form a linkage facility group; capturing the response time difference of each facility based on the linkage facility group to obtain a linkage time difference sequence; performing deviation analysis on the linkage time difference sequence to identify linkage coordination characteristics; and generating a facility status feature set through the linkage coordination characteristics.

[0015] Optionally, the step of assessing the remaining safe operating period of the high-risk group and identifying fault risk points based on the hierarchical monitoring task sequence includes: obtaining the operating load records and rated service life of each facility in the high-risk group based on the hierarchical monitoring task sequence; performing load fluctuation amplitude analysis on the operating load records to obtain a fluctuation amplitude sequence; performing weighted conversion based on the fluctuation amplitude sequence to obtain the equivalent operating time; and assessing the remaining safe operating period and identifying fault risk points based on the equivalent operating time and the rated service life.

[0016] Optionally, the step of generating a monitoring benchmark by adaptive threshold calibration of the distributed monitoring nodes based on the warning level parameters includes: obtaining node sensitivity levels by performing sensitivity classification on the distributed monitoring nodes based on the warning level parameters; determining the monitoring area type according to the deployment location of the distributed monitoring nodes; determining a dynamic alarm threshold by performing time-sharing load benchmark matching based on the node sensitivity level and the monitoring area type; and generating a monitoring benchmark based on the dynamic alarm threshold.

[0017] Optionally, the step of performing adjacent facility degradation propagation analysis on the degradation rate curve to identify degradation propagation source facilities includes: performing inter-facility degradation correlation analysis through the degradation rate curve to obtain facility adjacency relationships; performing synchronicity detection based on the facility adjacency relationships to obtain degradation synchronicity; determining a propagation threshold based on the degradation synchronicity to obtain suspected propagation facility pairs; and performing degradation time sequence analysis on the suspected propagation facility pairs to identify degradation propagation source facilities.

[0018] Optionally, the step of determining the dynamic alarm threshold based on the time-sharing load benchmark matching of the node sensitivity level and the monitoring area type includes: obtaining a regional load characteristic curve based on the monitoring area type; dividing the regional load characteristic curve into peak and valley periods to obtain a time period load benchmark; configuring the time period load benchmark with threshold floating based on the node sensitivity level to obtain a time-sharing threshold; and determining the dynamic alarm threshold according to the time-sharing threshold.

[0019] A second aspect of this invention provides an intelligent monitoring system for building facilities based on the Internet of Things, comprising:

[0020] The data acquisition module is used to collect sensor operation data and control command response data of building electromechanical facilities, and to detect the operation cycle offset between the operation data and the command response data to form an abnormal response mode of the facility.

[0021] The fault tracing module is used to extract the time sequence of abnormal events from the abnormal response pattern of the facility, arrange the time sequence of abnormal events in chronological order to construct a fault time sequence causal chain, and form a facility association matrix based on the fault time sequence causal chain.

[0022] The status assessment module is used to establish distributed monitoring nodes based on the facility association matrix, collect data on the working conditions of building electromechanical facilities based on the distributed monitoring nodes to generate a facility status feature set, and use the facility status feature set to construct a facility health distribution.

[0023] The risk grading module is used to analyze the facility deterioration rate in the facility health distribution to determine the deterioration rate, group the building electromechanical facilities according to the deterioration rate to identify high-risk and low-risk groups, and dynamically allocate monitoring resources to the high-risk and low-risk groups to construct a graded monitoring task sequence. The grouping of building electromechanical facilities according to the deterioration rate to identify high-risk and low-risk groups includes: time-series tracking of the deterioration rate to form a deterioration rate curve; analysis of the deterioration spread of adjacent facilities on the deterioration rate curve to identify deterioration spread source facilities; determining facility risk grading identifiers based on the deterioration spread source facilities and the deterioration rate; and identifying high-risk and low-risk groups based on the facility risk grading identifiers.

[0024] The life assessment module is used to assess the remaining safe operating cycle of the high-risk group based on the graded monitoring task sequence, identify fault risk points, and use the fault risk points to determine the early warning level parameters.

[0025] The early warning output module is used to perform adaptive threshold calibration on the distributed monitoring nodes based on the early warning level parameters to generate a monitoring benchmark, and to output a facility status monitoring report based on the fusion of the monitoring benchmark and the facility health distribution.

[0026] The beneficial effects of this invention are reflected in the following points: 1. By forming anomaly patterns in facility response through periodic offset detection and persistent offset analysis, and constructing a fault causal chain based on the temporal sequence of abnormal events, the causal relationships between facilities are extracted to form a facility correlation matrix. This reveals the transmission path of faults between different facilities, distinguishes the fault source from the affected facilities, and provides a technical means to prevent cascading failures. 2. In terms of facility status assessment, distributed monitoring nodes are deployed according to the facility correlation matrix to identify linked facility groups with triggering relationships. By collecting linkage time difference sequences and analyzing coordination characteristics, a facility health distribution is constructed. The linkage response characteristics are transformed into health scores, forming a status view covering all facilities, which facilitates maintenance personnel to grasp the overall facility operation status. 3. The system performs degradation rate analysis and spread detection on facility health, identifies high-risk and low-risk groups, constructs a graded monitoring task sequence, assesses the remaining safe operating cycle to identify fault risk points, determines the early warning level, and generates monitoring benchmarks by classifying the sensitivity of monitoring nodes and configuring time-based thresholds. It integrates health distribution to output facility status monitoring reports. Compared with the static monitoring strategies of existing technologies, this system enables dynamic adjustment of monitoring thresholds with time and risk level, tilting monitoring resources towards high-risk facilities, improving the utilization efficiency of monitoring resources and the timeliness of fault early warning. Attached Figure Description

[0027] The accompanying drawings illustrate specific examples of the technical solutions described in this invention and, together with the detailed embodiments, form part of the specification, serving to explain the technical solutions, principles, and effects of this invention.

[0028] Figure 1 This is a flowchart illustrating an intelligent monitoring method for building facilities based on the Internet of Things (IoT) according to the present invention.

[0029] Figure 2 This is a structural block diagram of an intelligent monitoring system for building facilities based on the Internet of Things (IoT) according to the present invention. Detailed Implementation

[0030] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0031] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0032] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0033] The technical solutions of the embodiments of this application will be described below.

[0034] like Figure 1 As shown, this embodiment of the invention provides a smart monitoring method for building facilities based on the Internet of Things, including the following steps S110-S160:

[0035] Step S110: Collect sensor operation data and control command response data of building electromechanical facilities, and perform operation cycle offset detection on the operation data and command response data to form an abnormal facility response mode.

[0036] Specifically, sensor operation data and control command response data of building electromechanical facilities are collected. Current sensors, temperature sensors, vibration sensors, and pressure sensors are deployed on the building electromechanical facilities. These sensors continuously collect the facility's operating status parameters, including motor operating current, equipment temperature, vibration amplitude, and pipeline pressure. These parameters are stored in chronological order to form operation data. The operation data records the complete operation process of the facility from start to stop. The operation data of air conditioning units shows a regular start-stop waveform, and the operation data of fresh air units adjusts with changes in indoor air quality. At the same time, control commands issued by the building automation system and the facility's response results are collected, including the time when the start-stop command is issued and the time when the facility actually acts, the target parameters of the adjustment command, and the parameters actually achieved. This information constitutes command response data. The command response data includes the facility's preset operating parameters, the preset operating cycle which specifies the standard start-stop interval of the facility, and the preset response delay which specifies the allowable time from command issuance to facility action. The command response data of exhaust fans records their preset start-stop mode at fixed time intervals, and the command response data of chillers records their response requirements to load changes.

[0037] In some embodiments, the step of detecting operational cycle offsets between the operational data and the instruction response data to form a facility response anomaly pattern includes: extracting periodic features from the operational data to obtain the actual operational cycle; comparing the actual operational cycle with a preset operational cycle in the instruction response data to calculate the cycle offset; performing positive and negative offset asymmetry analysis on the cycle offset to identify persistent offset events; and forming a facility response anomaly pattern based on the persistent offset events.

[0038] The actual operating cycle is obtained by extracting periodic features from the operating data. Parameters with periodic variation characteristics in the operating data are selected as the analysis objects: motor current parameter reflects the start-stop status and load changes of the facility, fan speed parameter reflects the operating rhythm of the facility, and water pump pressure parameter reflects the working cycle of the facility. A Fast Fourier Transform (FFT) is applied to the operating data to convert the time-domain signal into a frequency-domain representation. In the operating data of a building's air conditioning unit, the motor current exhibits regular start-stop fluctuations. After Fourier transform, the spectrum shows a peak with concentrated energy at a frequency of 0.000278Hz. The period corresponding to this peak is the dominant frequency component of the facility's operation. The period corresponding to the dominant frequency component is the actual operating cycle, calculated using the formula T = 1 / f, where T is the actual operating cycle in seconds, and f is the dominant frequency in Hertz. Substituting 0.000278Hz into the formula, the actual operating cycle is calculated to be 3600 seconds, or 1 hour, which matches the unit's design parameters of starting and stopping on the hour. For facilities with irregular start-stop patterns, autocorrelation function analysis is used to analyze the repetitive patterns of operating data. The time interval between the peak values ​​of the autocorrelation curve is the actual operating cycle. Precision air conditioners operate continuously without stopping, but the compressor loading and unloading process in their operating data exhibits periodic characteristics. After autocorrelation analysis, the operating data of a precision air conditioner in a data center showed a correlation coefficient peak of 0.92 at a time lag of 1800 seconds. Based on this, the actual operating cycle of the equipment was determined to be 1800 seconds, or 30 minutes, corresponding to the compressor loading and unloading cycle.

[0039] The cycle offset is calculated by comparing the actual operating cycle with the preset operating cycle in the command response data. The actual operating cycle is aligned one-to-one with the preset operating cycle of the corresponding facility in the command response data to ensure comparison of cycle data for the same facility and the same time period. The preset operating cycle recorded in the command response data is a standard parameter set by the building automation system based on energy-saving strategies and comfort requirements. The formula for calculating the cycle offset is ΔT = T_a - T_p, where ΔT is the cycle offset, T_a is the actual operating cycle, and T_p is the preset operating cycle in the command response data, all in seconds. For example, the actual operating cycle of an air conditioning unit in a building, extracted using Fourier transform, is 3720 seconds. The preset operating cycle for this unit in the command response data is 3600 seconds. Substituting these two values ​​into the formula yields a cycle offset of 120 seconds, indicating that the actual operating rhythm of the unit is 2 minutes slower than the preset requirement. A positive cycle offset indicates that the actual operating rhythm of the facility is slower than the preset requirement, which may be caused by mechanical wear leading to extended start-up time, delayed response of the control system, or excessive load on the facility. A negative cycle offset indicates that the actual operating rhythm of the facility is faster than the preset requirement, which may be caused by control parameter drift, sensor measurement error, or the facility operating under no-load conditions. If the exhaust fan's motor bearings age, causing a decrease in speed, the time required to complete the predetermined ventilation volume will be extended, and the cycle offset will also show a positive shift. For example, if the actual operating cycle of a certain exhaust fan is 1920 seconds while the preset operating cycle is 1800 seconds, the calculated cycle offset is 120 seconds.

[0040] Asymmetric analysis of positive and negative offsets in periodic offsets is used to identify persistent offset events. The positive and negative distribution of periodic offsets over multiple consecutive operating cycles is statistically analyzed. For normally operating facilities, the periodic offset should fluctuate randomly in both positive and negative directions with limited amplitude, exhibiting a symmetrical distribution centered at zero. When the periodic offset consistently deviates in the same direction for more than five consecutive cycles, it indicates a systematic deviation from the facility's operating status rather than a random disturbance; this unidirectional, continuous offset breaks the positive-negative symmetry. The cumulative value of the periodic offset over the observation period is calculated. For a building's central air conditioning unit, the periodic offsets over seven consecutive days were +85 seconds, +92 seconds, +78 seconds, +105 seconds, +98 seconds, +110 seconds, and +125 seconds, respectively. All seven cycles showed positive offsets with a cumulative value of 693 seconds, continuously increasing in the same direction. When the absolute value of the cumulative offset exceeds a preset threshold of 500 seconds and the offset direction remains consistent, this period is identified as a continuous offset event. The cumulative offset of this air conditioning unit exceeded the threshold by 693 seconds, thus identifying a continuous offset event. The continuous offset event record shows the start time of the offset as day 1 of monitoring, the duration as 7 days, the cumulative offset as 693 seconds, and the offset direction as positive. The initial assessment is that the cooling capacity is nearing saturation, leading to a prolonged start-up process. The circulating water pump's increased flow resistance due to scaling on the inner wall of the pipes, with a continuously positive periodic offset forming a continuous offset event, indicates that the pipes require cleaning and maintenance.

[0041] A facility response anomaly pattern was formed based on persistent offset events. The characteristic parameters of each persistent offset event were analyzed, including offset direction, cumulative offset amplitude, duration, and temporal pattern. The offset direction reflects the nature of the anomaly, the cumulative amplitude reflects the severity of the anomaly, and the duration reflects the stability of the anomaly. Multiple persistent offset events occurring within similar time periods of the same facility were correlated. In June, three persistent offset events were identified for an air conditioning unit in a building: the first had a cumulative offset of 520 seconds, the second 680 seconds, and the third 890 seconds. The offset amplitude showed a progressively increasing trend, and the interval between occurrences decreased from 15 days to 10 days, indicating a continuous deterioration in the facility's operational status. Typical anomaly manifestations were extracted based on the common characteristics of the persistent offset events, generating a facility response anomaly pattern. The common characteristics of the three persistent offset events were positive offset, increasing amplitude, and triggering during high-temperature periods, thus generating a facility response anomaly pattern. The facility response anomaly mode includes an anomaly type of extended positive cycle, involves the air conditioning unit number, has a characteristic parameter of an average deviation of 697 seconds, and occurs when the outdoor temperature exceeds 35°C. This facility response anomaly mode describes typical behavioral characteristics of a facility deviating from normal operating conditions. It reveals a recurring issue where the compressor start-up time is prolonged during high-load periods due to increased condensing pressure, necessitating condenser cleaning or increased cooling capacity before the summer season.

[0042] Step S120: Extract the time sequence of abnormal events from the facility response anomaly mode, arrange the time sequence of abnormal events in chronological order to construct the fault time sequence causal chain, and form a facility correlation matrix based on the fault time sequence causal chain.

[0043] Specifically, the time sequence of anomalous events is extracted from facility response anomaly patterns. Each facility response anomaly pattern may occur multiple times at different times. The occurrence instances of each anomaly pattern are expanded along the time dimension, with each occurrence constituting an independent anomaly event. Each anomaly event inherits the type identifier and characteristic parameters of its respective facility response anomaly pattern, while also attaching the specific time of occurrence. For example, the air conditioning system of a certain building exhibited anomaly response patterns five times during the observation period, occurring on June 5th at 14:00, June 12th at 15:30, June 18th at 14:20, June 25th at 13:50, and July 2nd at 14:10. These five occurrence instances are labeled as anomaly events E1 to E5, and each anomaly event records its occurrence time and inherited characteristic parameters. All anomaly events are labeled according to their occurrence time, forming a set of events with timestamps; this set constitutes the anomaly event time sequence. The anomaly event time series records the temporal distribution of all anomalies in the building's electromechanical systems during the observation period. Events involving response lag in the air conditioning system, pressure fluctuations in the water pump system, and timeouts in the elevator system are all included in the anomaly event time series according to their respective occurrence times. The anomaly event time series shows that air conditioning anomalies are concentrated in the afternoon period from 13:00 to 15:30, coinciding with peak outdoor temperature and peak population density. This temporal clustering suggests that there may be common triggering factors or causal transmission relationships among the anomalies.

[0044] In some embodiments, the step of arranging the abnormal event sequence in chronological order to construct a fault time-series causal chain includes: timestamping the abnormal event sequence to obtain an event occurrence sequence; arranging the event occurrence sequence in chronological order to form an ordered event chain; performing cross-level event progressive association on the ordered event chain to identify indirect related events across devices and systems; and constructing a fault time-series causal chain based on the indirect related events and the ordered event chain.

[0045] The time sequence of abnormal events is obtained by timestamping the event occurrence sequence. Each abnormal event in the time sequence undergoes precise timestamp calibration to eliminate clock deviations between different acquisition devices, with clock synchronization accuracy controlled within 100 milliseconds. Events with similar timestamps in the time sequence are clustered; events with time intervals less than a set threshold are considered to have occurred at the same time and assigned the same sequence number. The abnormal event time sequence of a building's chiller station includes chiller unit startup anomalies, chilled water pump pressure fluctuations, and air conditioning terminal temperature anomalies. After timestamp calibration, these three events are arranged sequentially: the chiller unit event is assigned sequence number 1, the chilled water pump event is assigned sequence number 2, and the air conditioning terminal event is assigned sequence number 3, forming the event occurrence sequence. The event occurrence sequence retains the original attribute information of each event, including event type, involved facilities, and characteristic parameters, while also adding a sequence position number. The event occurrence sequence of the above chiller station shows that the time interval between the chiller unit event and the chilled water pump event is 37 seconds. This time interval matches the transmission time of chilled water from the unit to the pump, initially indicating a sequential correlation between the two events due to media transmission.

[0046] Events are arranged chronologically to form an ordered event chain. Adjacent events in the sequence are paired to analyze potential correlations between them. The criteria for determining correlation include the physical connection between facilities, the direction of media flow, and the transmission path of control signals. In a building's refrigeration system, the chiller unit startup anomaly is the first event in the sequence, followed by the chilled water pump pressure fluctuation. The chiller unit and chilled water pump are physically connected via chilled water pipes, and the media flow is from the chiller unit to the pump. This indicates a potential correlation between the two events, and the chiller unit event is linked to the chilled water pump event in a chain. When multiple events form a continuous transmission relationship, they are chained together to form an ordered event chain. In the above sequence, the three events are linked sequentially to form an ordered event chain: chiller unit startup anomaly → chilled water pump pressure fluctuation → air conditioning terminal temperature anomaly. The ordered event chain records the type, facility, and time of occurrence of each event, as well as the time interval and transmission direction between adjacent events. The ordered event chain of the heating system shows that the boiler outlet water temperature drops first, followed by the terminal radiator temperature being insufficient. The time interval corresponds to the duration of hot water transportation in the pipe network. The ordered event chain describes the path of the anomaly from the source facility to the downstream facility level by level.

[0047] This study identifies indirectly related events by performing cross-level event progression association on ordered event chains. It analyzes whether event transmission across multiple facility levels exists within the ordered event chain. Directly related events occur between physically adjacent facilities, while indirectly related events occur between facilities that are not physically adjacent but connected through intermediate facilities. An ordered event chain for a building's central air conditioning system contains five nodes: cooling tower fan failure → cooling water temperature rise → chiller unit high-pressure alarm → chilled water supply temperature rise → terminal air supply temperature rise. The cooling tower fan failure event and the terminal air supply temperature rise event are not directly adjacent but are connected through three intermediate nodes, thus constituting an indirectly related event. The transmission path crosses two levels: the cooling system and the air conditioning system. The transmission strength of the indirectly related events is evaluated. Transmission strength is negatively correlated with path length and positively correlated with the correlation tightness of each link. The indirectly related events in this study involve direct transmission through media, resulting in a high correlation tightness. The overall calculated transmission strength is 0.72. Indirectly correlated events are recorded, including source events, end events, intermediate transmission nodes, transmission path length, and transmission intensity. The building power supply and distribution system identifies indirectly correlated events between voltage fluctuations and frequent start-stop of precision air conditioning compressors. The transmission path passes through the distribution cabinet and UPS system, and the transmission intensity of 0.65 indicates that power supply quality has a significant impact on the operational stability of precision air conditioning.

[0048] A fault-sequential causal chain is constructed based on indirectly related events and ordered event chains. The direct relationships in the ordered event chains and the indirect relationships in the indirect events are integrated to construct a complete event causal network. The causal paths in the network are analyzed to identify the complete transmission chain from the root event to the terminal event. Each chain constitutes a fault-sequential causal chain. For example, the ordered event chain of a building's refrigeration system includes the direct relationship between the cooling tower, chiller unit, and air conditioning terminal. Indirectly related events record the cross-level relationships between the cooling tower and the air conditioning terminal. The fault-sequential causal chain constructed after integrating these two is: cooling tower packing blockage (root cause) → increased cooling water temperature → decreased chiller unit efficiency → increased chilled water supply temperature → insufficient cooling at the air conditioning terminal (symptom). The fault-sequential causal chain clearly reveals the complete path of the fault's transmission from the cold source to the terminal. The fault-sequential causal chain records the temporal position, causal direction, and transmission delay of each link. The starting point of the chain is the root cause of the fault, the ending point is the symptom of the fault, and the intermediate nodes are the transmission links. The fault sequence causal chain constructed based on indirect related events and ordered event chains of the fire protection system reveals the fault transmission process of fire water tank level drop → fire pump frequent start → pipeline pressure fluctuation → insufficient pressure of terminal sprinkler heads. The root cause points to the abnormality of the fire water tank water replenishment system, and the working status of water replenishment valves and level controllers should be checked first.

[0049] A facility association matrix is ​​formed based on the fault-sequence causal chains. All facilities involved in the fault-sequence causal chains and their relationships are statistically analyzed. Adjacent facilities in each fault-sequence causal chain are paired, and the causal relationships between them are recorded. A two-dimensional matrix is ​​constructed with facilities as rows and columns. The rows of the matrix represent causal facilities, the columns represent result facilities, and the matrix elements represent the association strength from the causal facility to the result facility. Multiple fault-sequence causal chains were identified for a certain building. Among them, the facility pairing of chiller unit → air conditioning unit appears in multiple fault-sequence causal chains and has a short transmission delay. The overall association strength is calculated to be 0.85, and this 0.85 is filled into the intersection of the chiller unit row and the air conditioning unit column in the facility association matrix. The association strength is calculated based on the frequency of the facility pair's appearance in all fault-sequence causal chains and the transmission delay; the higher the frequency and the shorter the transmission delay, the stronger the association strength. The facility correlation matrix is ​​an asymmetric matrix reflecting the directionality of fault propagation between facilities. The correlation strength between a chiller unit failure and an air conditioning unit malfunction is 0.85, while the correlation strength between an air conditioning unit failure and a chiller unit malfunction is only 0.12, showing a significant difference in correlation strength between the two directions. The diagonal elements of the facility correlation matrix are zero, indicating that there is no causal relationship between the facility and itself. The matrix also shows that the correlation strength between the standby generator and critical equipment reaches 0.78. When the standby generator malfunctions, the operating status of all critical equipment with a correlation strength exceeding 0.5 should be monitored simultaneously.

[0050] Step S130: Establish distributed monitoring nodes based on the facility association matrix, collect data on the working conditions of building electromechanical facilities based on the distributed monitoring nodes to generate facility status feature sets, and construct facility health distribution using the facility status feature sets.

[0051] Specifically, a distributed monitoring node is established based on the facility association matrix. The distribution of association strength among facilities in the facility association matrix is ​​analyzed to identify key facilities with high association strength and convergence nodes of association paths. In the facility association matrix, facilities with a larger sum of row elements are the main outputters of faults, and their anomalies are easily propagated to multiple downstream facilities. The facility association matrix of a certain building shows that the sum of elements in the row containing the chiller unit is much higher than that of other facilities. When the chiller unit fails, it will simultaneously affect multiple downstream facilities such as chilled water pumps, air conditioning terminals, and fresh air units, making it a key fault output node. Facilities with a larger sum of column elements are the main bearers of faults. The column containing the air conditioning terminal units has a high sum of elements, making it susceptible to the influence of multiple upstream facilities such as chiller units, fresh air units, and air supply ducts. The deployment locations of monitoring nodes are determined based on the correlation strength distribution in the facility correlation matrix. Chiller units, as critical nodes for fault output, are equipped with high-precision, high-frequency distributed monitoring nodes, featuring vibration, temperature, and current sensors for simultaneous multi-parameter acquisition. Lighting circuits have lower correlation strength, and their faults have limited impact on other facilities; therefore, deploying distributed monitoring nodes with conventional accuracy suffices. The distributed monitoring nodes employ an edge computing architecture, with each node possessing local data acquisition, preprocessing, and temporary storage capabilities. Nodes are interconnected via industrial Ethernet or wireless networks. The deployment density of distributed monitoring nodes is positively correlated with the correlation strength in the facility correlation matrix. High-density distributed monitoring nodes are deployed in the densely correlated refrigeration station area, while low-density distributed monitoring nodes are deployed in the relatively independent general office area, forming focused monitoring coverage for critical facilities.

[0052] In some embodiments, the step of collecting data on the working conditions of building electromechanical facilities based on the distributed monitoring nodes to generate a facility status feature set includes: identifying the linkage triggering relationship between building electromechanical facilities based on the distributed monitoring nodes to form a linkage facility group; capturing the response time difference of each facility based on the linkage facility group to obtain a linkage time difference sequence; performing deviation analysis on the linkage time difference sequence to identify linkage coordination characteristics; and generating a facility status feature set through the linkage coordination characteristics.

[0053] Based on distributed monitoring nodes, the linkage triggering relationships between building electromechanical facilities are identified to form linkage facility groups. Distributed monitoring nodes collect real-time data on the operational status changes of each facility, identifying the triggering relationships where a change in one facility's status leads to changes in the status of other facilities. When a distributed monitoring node detects the start or stop of a facility, it scans other distributed monitoring nodes within a set time window to see if they have detected status changes in related facilities. For example, when a building's chiller plant starts its cooling mode in the morning, the chiller unit receives the start command from the building automation system and begins operation. The distributed monitoring nodes then detect the start of the cooling water pump, cooling tower fan, and chilled water pump. The start signals of all four facilities are captured within the set time window, indicating a stable linkage triggering relationship between these facilities. Facilities with stable triggering relationships are grouped into linkage facility groups. The above four facilities form the chiller plant's linkage facility group. The facilities within the linkage facility group cooperate with each other during operation; the action of one facility triggers a response action from the other facilities in the group. The scale of the linkage facility group varies from a simple pairing of two facilities to a complex linkage of multiple facilities. The linkage facility group of the air conditioning system includes air conditioning units, fresh air units and exhaust units, while the linkage facility group of the fire protection system includes smoke detectors, fire pumps and smoke exhaust fans.

[0054] The linkage time difference sequence is obtained by capturing the response time differences of each facility in the linkage facility group. The response time of each facility in the linkage facility group to a trigger signal is monitored. The trigger signal can be the start / stop action of the main facility, a regulation command, or an alarm signal. The time difference between the response time of each facility in the linkage facility group and the time of the trigger signal is calculated; this time difference is the response time difference of that facility. In the morning startup process of a building's chiller station linkage facility group, the moment the chiller unit receives the start command is taken as the trigger signal issuance time. The cooling water pump starts immediately after the trigger signal is issued, with a short response time difference; the cooling tower fan starts slightly later, with a slightly longer response time difference; the chiller unit compressor starts after completing pre-lubrication, further extending the response time difference; the chilled water pump starts after the chiller unit has stabilized, with the longest response time difference. Each facility starts sequentially according to a preset delay order. The response time differences of all facilities in the linkage facility group are arranged in the order of response to form a linkage time difference sequence. The length of the linkage time difference sequence is equal to the number of facilities in the linkage facility group, and each element corresponds to the response time difference value of one facility. The linkage time difference sequence describes the temporal distribution of the responses of each facility during the linkage process. When the linkage facilities of the air conditioning system are activated, the fresh air unit responds first to introduce fresh outdoor air, the air conditioning unit then starts to adjust the temperature and humidity, and the exhaust fan unit finally starts to exhaust the indoor stale air, forming a complete linkage time difference sequence.

[0055] Deviation analysis is performed on the linkage time difference sequence to identify linkage coordination characteristics. A baseline value for the linkage time difference sequence is established, derived from the standard response time difference recorded during facility commissioning or the response time difference requirements specified in the design documents. The deviation between the current linkage time difference sequence and the baseline value is calculated using the formula Δt_i = t_i - t_base_i, where Δt_i is the response time difference deviation of the i-th facility, t_i is the response time difference of the i-th facility in the current linkage time difference sequence, and t_base_i is the baseline response time difference of that facility, all in seconds. When a building's chiller plant starts up in the morning, the cooling water pumps and cooling tower fans respond normally according to the preset timing, with a response time difference deviation close to zero. However, the chiller compressor starts significantly later than the baseline value, with a positive delay in the response time difference deviation. Consequently, the chilled water pumps exhibit a slight response delay. Investigation revealed that aging of the chiller compressor's starting capacitor caused a slow start-up current, which in turn affected the linkage timing of the downstream chilled water pumps. Analyzing the distribution characteristics of deviations in the linkage time difference sequence reveals that deviations concentrated in a single facility indicate abnormal performance of that facility, while deviations dispersed across multiple facilities indicate a decline in overall linkage coordination. Facilities with deviations exceeding the allowable range in the linkage time difference sequence are identified, and the direction and magnitude of these deviations are extracted to form a comprehensive linkage coordination characteristic. This characteristic describes the degree of synchronization and stability of the coordination among facilities within the linkage group. Synchronization reflects the consistency of responses from each facility, while stability reflects the fluctuation range of response time differences. The linkage coordination characteristic of the fire protection system shows that the response time difference deviations of each facility are within the allowable range, indicating good linkage coordination and the ability to meet the requirements of rapid response after a fire alarm.

[0056] A facility status feature set is generated based on the linkage and coordination characteristics. The linkage and coordination characteristics of all linked facility groups are aggregated, integrating the feature data scattered across various linkage groups. A status feature file is established for each facility, recording its linkage and coordination characteristics within its respective linked facility group. For example, a building chiller unit belongs to both the chiller plant linkage facility group and the cold / heat source switching linkage facility group. In the chiller plant linkage scenario, the unit's linkage and coordination characteristics show a slow start-up response; in the cold / heat source switching scenario, the unit's linkage and coordination characteristics show a normal switching response. Based on the combined linkage and coordination characteristics of the two scenarios, it is determined that the unit's start-up performance has degraded, but its steady-state operation is normal. Key status indicators of the facilities are extracted based on the linkage and coordination characteristics, including average response time deviation, maximum response time deviation, deviation fluctuation amplitude, and deviation change trend. These indicators comprehensively reflect the current status and evolution trend of the facilities. The status indicators of all facilities are aggregated to form a facility status feature set, which is a data collection indexed by facilities and with status indicators as fields. The facility status feature set covers all electromechanical facilities in the building that are included in the linkage monitoring. Each facility corresponds to a feature record, which contains the multi-dimensional status indicators of the facility. The facility status feature set of the building electromechanical system includes the status features of various facilities such as air conditioning units, chillers, water pumps, and fans.

[0057] A facility health distribution is constructed using a facility status feature set. Based on the status indicators of each facility in the feature set, a facility health score is calculated. The health score comprehensively considers three dimensions: response time deviation, deviation fluctuation amplitude, and deviation trend. The health calculation formula is H = 100 - α × Δt - β × σ - γ × k, where H is the facility health score, with a maximum score of 100; Δt is the normalized value of the average response time deviation, obtained by dividing the average response time deviation of the facility in the feature set by a reference threshold (divided by a dimensionless value); σ is the normalized value of the standard deviation of the response time deviation, reflecting the fluctuation amplitude (divided by a dimensionless value); k is the normalized value of the slope of the deviation trend, reflecting the rate of deterioration (divided by a dimensionless value); and α, β, and γ are dimensionless weighting coefficients reflecting the relative importance of the three factors in the health assessment. A building chiller unit operated at full load continuously during the high-temperature summer period. The facility status feature set recorded an average response time deviation that was positively delayed and showed a weekly increasing trend. Substituting this into the health score formula, the calculated health score was in the yellow warning range, indicating that the compressor's starting performance had deteriorated due to prolonged high-load operation, requiring compressor overhaul after the cooling season. The health scores of all facilities were grouped according to spatial location or system affiliation to form a facility health score distribution. This distribution was presented as a heatmap or list, showing the health status of each facility: high health facilities were displayed in green, medium health facilities in yellow, and low health facilities in red.

[0058] Step S140: Analyze the facility deterioration rate in the facility health distribution to determine the deterioration rate, group the building electromechanical facilities according to the deterioration rate to identify high-risk groups and low-risk groups, and dynamically allocate monitoring resources to the high-risk groups and low-risk groups to construct a hierarchical monitoring task sequence.

[0059] Specifically, the degradation rate is determined by analyzing the facility health distribution. A historical health score time series for each facility in the facility health distribution is established. For example, during a period of continuous high temperatures, the operating efficiency of a chiller unit in a building continuously decreased due to poor condenser heat dissipation. The health score of this unit in the facility health distribution gradually decreased from the green healthy range to the yellow warning range during the observation period. These data points are arranged chronologically to form a health time series. Linear regression analysis is performed on the health time series of each facility in the facility health distribution to fit a trend line of health degradation over time. The degradation rate is calculated using the formula v = ΔH / Δt, where v is the degradation rate, ΔH is the change in health score during the observation period, Δt is the observation time span, and the unit of degradation rate is minutes / day. A negative value indicates a decline in health. The above chiller unit, when calculated using the formula, has a negative degradation rate, indicating that the unit is in an accelerated degradation stage and requires attention to condenser cleaning and maintenance. The larger the absolute value of the degradation rate, the faster the facility's health deteriorates, requiring earlier maintenance. The degradation rate of all facilities in the facility health distribution is calculated to form a degradation rate dataset corresponding to each facility. The fresh air handling unit operates stably within the same observation period, with a degradation rate close to zero, and its health status is relatively stable. The circulating water pump experiences increased vibration due to bearing wear, resulting in a larger absolute value of the degradation rate and a faster degradation speed, requiring priority maintenance.

[0060] In some embodiments, the step of grouping building electromechanical facilities into high-risk and low-risk groups according to the degradation rate includes: performing time-series tracking of the degradation rate to form a degradation rate curve; performing adjacent facility degradation propagation analysis on the degradation rate curve to identify degradation propagation source facilities; determining facility risk classification identifiers based on the degradation propagation source facilities and the degradation rate; and identifying high-risk and low-risk groups based on the facility risk classification identifiers.

[0061] A degradation rate curve was generated by time-series tracking of the degradation rate. The degradation rate of each facility was continuously calculated using a sliding time window, with a window width of 7 days and a sliding step of 1 day. After each sliding window, the degradation rate within the window was recalculated, and the calculation time and corresponding rate value were recorded. A building chiller unit operated relatively well in the early summer, with a low degradation rate. However, as the outdoor temperature continued to rise and the condensing pressure increased, the unit's operating load increased, and the degradation rate gradually worsened. The absolute value of the degradation rate calculated after each window sliding increased. Connecting the degradation rates from multiple consecutive moments in chronological order formed the degradation rate curve. The degradation rate curve for this chiller unit showed an overall downward slope, indicating that degradation was accelerating. The slope of the degradation rate curve reflects whether degradation is accelerating or decelerating; a downward slope indicates accelerating degradation, while a flatter curve indicates a stable degradation rate. Analyzing the fluctuation characteristics of the degradation rate curves, the inflection point of the sudden increase in degradation rate was identified. The degradation rate curves of the aforementioned chiller units showed a clear inflection point at a certain time. Before the inflection point, the degradation rate was relatively flat; after the inflection point, the degradation rate rapidly deteriorated. This inflection point corresponded to the moment when the outdoor temperature first exceeded the 35°C high-temperature warning level, indicating that extreme high temperature was the key factor triggering the accelerated degradation of the unit. The degradation rate curves of the cooling tower fans remained stable within the same observation period, without significant changes.

[0062] For example, the step of performing adjacent facility degradation spread analysis on the degradation rate curve to identify degradation spread source facilities includes: performing inter-facility degradation correlation analysis through the degradation rate curve to obtain facility adjacency relationships; performing synchronicity detection based on the facility adjacency relationships to obtain degradation synchronicity; determining a spread threshold based on the degradation synchronicity to obtain suspected spread facility pairs; and performing degradation time sequence analysis on the suspected spread facility pairs to identify degradation spread source facilities.

[0063] Correlation analysis of deterioration rates between facilities was conducted to obtain facility adjacency relationships. The Pearson correlation coefficient was calculated between the deterioration rate curves of different facilities. The correlation coefficient reflects the consistency of the changing trends of the two curves. A correlation coefficient close to 1 indicates a strong correlation between the deterioration rates of the two facilities, with both rising and falling in tandem. For example, a chiller unit in a building experienced a continuous decline in heat exchange efficiency due to evaporator scaling, resulting in an accelerated deterioration trend in its deterioration rate curve. Simultaneously, a chilled water pump exhibited a similar accelerated deterioration due to increased pipe flow resistance. The Pearson correlation coefficient of the two deterioration rate curves was close to 1, showing a highly consistent trend and indicating a correlation between their deterioration processes. A correlation coefficient close to 0 indicates that the deterioration rate changes of the two facilities are independent. Pairwise correlation calculations were performed on the deterioration rate curves of all facilities. Facility pairs with correlation coefficients exceeding a set threshold were marked as having a facility adjacency relationship. For instance, the correlation coefficient between the chiller unit and the chilled water pump exceeded the threshold, also marking them as having a facility adjacency relationship. Facility adjacency describes the pairing of facilities that exhibit correlation during the degradation process. The adjacency relationship between air conditioning units and fresh air handling units is weak, with correlation coefficients below the threshold, indicating that their degradation processes are relatively independent. Cooling towers and cooling water pumps, operating in series within the refrigeration system, show similar degradation rate curves and exhibit a strong facility adjacency relationship.

[0064] Synchronization detection based on facility adjacency relationships is used to obtain the degree of degradation synchronization. For facility pairs with facility adjacency relationships, the temporal synchronization degree of their degradation rate changes is further analyzed. The cross-correlation function of the degradation rate curves of the two facilities is calculated, and the peak position of the cross-correlation function reflects the time delay between the two curves. The formula for calculating the degree of degradation synchronization is S=R_max×τ_ref / (τ+τ_0), where S is the degree of degradation synchronization, dimensionless, ranging from 0 to 1; R_max is the peak value of the cross-correlation function, reflecting the correlation strength between the two curves, ranging from 0 to 1; τ is the time delay corresponding to the peak value, in days; τ_ref is the reference time delay, which is taken as the typical fault propagation cycle of the system, set to 1 day based on historical data statistics; τ_0 is a smoothing constant, with a value of 0.1 days, used to avoid the division problem when τ is zero and to ensure the upper limit of S. A building's chiller unit and chilled water pump are adjacent facilities. Cross-correlation analysis of their degradation rate curves revealed that the chilled water pump's degradation trend lags behind the chiller unit by several days. This time delay aligns with the system transmission characteristics of chilled water flowing from the unit to the pump. Calculations using the formula show a high degree of synchronicity in degradation between the two facilities, indicating that the chiller unit's degradation may be spreading to the chilled water pump. A higher degree of synchronicity indicates that the degradation processes of the two facilities are more synchronized in time, and the degradation of one facility is more likely to affect the other. Degradation synchronicity was calculated for all adjacent facility pairs, creating a table mapping facility pairs to synchronicity. The cooling tower and air conditioning terminal exhibited a low degree of synchronicity, with a long time delay, reflecting that the fault transmission from the cold source side to the terminal side requires multiple intermediate steps.

[0065] To identify potential spreading facility pairs, a spread threshold is determined based on the degree of degradation synchronicity. A spread threshold for degradation synchronicity is set, typically based on confirmed degradation spread cases in historical data, usually the 75th percentile of the synchronicity distribution. When the degradation synchronicity of a facility pair exceeds the spread threshold, it is marked as a potential spreading facility pair, indicating a possible mutual transmission of degradation between the two facilities. In a building, a chiller unit and a chilled water pump were marked as a potential spreading facility pair because their degradation synchronicity exceeded the spread threshold. These two units are directly connected in the refrigeration system via chilled water pipes. Scale buildup on the chiller unit's evaporator caused an increase in outlet water temperature, and the chilled water pump's long-term delivery of excessively hot chilled water accelerated the aging of seals, creating a transmission path of degradation spreading from upstream to downstream. Further analysis of the causal direction is needed for potential spreading facility pairs to determine which facility is the source of degradation and which is the recipient. Degradation synchronicity is ranked, with facility pairs with higher synchronicity being prioritized for causal analysis due to their greater likelihood of spreading. Cooling towers and cooling water pumps were also marked as potential spreading facility pairs. The air conditioning unit and the lighting system deteriorated at a rate below the threshold, and did not constitute a suspected spread of the deterioration of the facilities. The deterioration of the two was determined to be an independent event.

[0066] For suspected spreading facility pairs, a sequence analysis of degradation was conducted to identify the source of degradation. The order in which the degradation rates of the two facilities in the suspected spreading facility pair began to accelerate was analyzed, with the facility showing accelerated degradation first identified as the source of degradation. For each facility in the suspected spreading facility pair, the inflection point on its degradation rate curve was identified, corresponding to the starting point of accelerated degradation. In a suspected spreading facility pair of a chiller unit and a chilled water pump in a certain building, the chiller unit's degradation rate curve showed an inflection point and began to accelerate degradation early in the observation period due to scaling on the evaporator tube bundle, while the chilled water pump's degradation rate curve showed a significantly delayed inflection point. Analysis showed that this delay was consistent with the thermal inertia of the chilled water system, indicating that the chiller unit deteriorated first, followed by the chilled water pump. By comparing the inflection points of the two facilities in the suspected spreading facility pair, the facility with the earlier inflection point was marked as the source of degradation, and the chiller unit was identified as the source of degradation. The source of degradation is the root cause of degradation in related facilities. Prioritizing the treatment of the source of degradation can prevent the degradation from spreading to other facilities. Among suspected sources of degradation, cooling towers and cooling water pumps are identified as the source of degradation. The cooling tower's heat dissipation capacity decreases due to packing blockage, and its inflection point occurs earlier than that of the cooling water pump. The cooling tower is thus identified as the source of degradation. The increased cooling water temperature causes the mechanical seal of the cooling water pump to experience higher thermal stress, accelerating wear.

[0067] Facility risk classification is determined by considering both the facility's own degradation rate and whether it is a source of degradation spread. Each facility is assessed for risk level by comprehensively considering both its degradation rate and whether it is a source of degradation spread. The facility risk scoring formula is R = w1 × |v| / v_ref + w2 × I_source, where R is the risk score, v is the facility's degradation rate (in minutes / day), v_ref is the reference degradation rate threshold (set by the system as a baseline for early warning) (in minutes / day), and |v| / v_ref is the normalized degradation degree index (dimensionless). I_source is the spread source identifier, taking a value of 1 if the facility is a source of degradation spread, and 0 otherwise (dimensionless). w1 and w2 are weighting coefficients, both dimensionless, with w1 + w2 = 1, reflecting the relative importance of degradation rate and spread source factors in risk assessment. R is the dimensionless comprehensive risk score, and risk levels are determined based on the R value. During the sustained high temperatures of summer, a chiller unit in a building experienced a decrease in heat exchange efficiency due to scaling on its evaporator tube bundles. The degradation rate continued to worsen, and the degradation had spread to downstream chilled water pumps and air conditioning terminals. Identified as a source of degradation propagation, the calculated risk score based on the degradation rate and the source identification exceeded the high-risk threshold, classifying the facility as high-risk. Chemical cleaning of the evaporator should be prioritized after the cooling season ends. Risk levels are categorized based on risk scores: facilities with scores above the upper threshold are marked as high-risk, those below the lower threshold are marked as low-risk, and those in between are marked as medium-risk. A fresh air handling unit in another building operated stably during the same observation period, maintaining a low degradation rate without affecting other related facilities. As a non-source of degradation propagation, its risk score was below the lower threshold, classifying it as low-risk. Maintaining the usual inspection frequency is sufficient.

[0068] High-risk and low-risk groups are identified based on facility risk classification labels. All facilities classified as high-risk are grouped into the high-risk group, and all facilities classified as low-risk are grouped into the low-risk group. Medium-risk facilities are not included in either group. A building has several electromechanical facilities. Facilities classified as high-risk are concentrated in the chiller station and fire pump room areas, while facilities classified as low-risk are distributed in air conditioning rooms and public areas on each floor. Facilities in the high-risk group deteriorate rapidly or have a spreading effect, requiring priority monitoring and maintenance. The high-risk group of the building's refrigeration system includes chillers, cooling towers, and circulating water pumps; these facilities are concentrated in the chiller station area and exhibit a deterioration-spreading relationship with each other. Facilities in the low-risk group are relatively stable in operation, and monitoring frequency can be appropriately reduced to conserve resources. The low-risk group includes fresh air units, exhaust fans, and terminal fan coil units, distributed in air conditioning rooms and ceiling spaces on each floor. The number and spatial distribution of facilities in high-risk and low-risk groups were statistically analyzed to determine whether there was a regional concentration of high-risk facilities. The high-risk group of fire protection systems included fire pumps and pressure stabilizing equipment, which were mainly distributed in the basement fire pump room, while the low-risk group included fire hydrants and sprinkler heads on each floor.

[0069] A hierarchical monitoring task sequence is constructed by dynamically allocating monitoring resources for high-risk and low-risk groups. Differentiated monitoring strategies are developed based on the differences in monitoring needs between high-risk and low-risk groups. Facilities in high-risk groups are configured with high-frequency monitoring tasks, shorter sampling intervals, higher data reporting frequencies, and stricter alarm response time requirements. For example, a chiller unit in a building, belonging to the high-risk group and a source of deterioration spread, is configured with high-frequency sampling, rapid reporting, and immediate alarms, focusing on monitoring the evaporator inlet and outlet temperature difference and compressor operating current. Any abnormality triggers an alarm immediately to notify maintenance personnel. Facilities in low-risk groups are configured with routine monitoring tasks, maintaining longer sampling intervals, moderate data reporting frequencies, and more lenient alarm response time requirements. For example, a fresh air handling unit in a building, belonging to the low-risk group, is configured with routine sampling and timed reporting, primarily monitoring the fan operating status and filter pressure difference; periodic inspections are sufficient to meet its monitoring needs. Monitoring tasks are prioritized, with high-risk group tasks at the beginning of the sequence and low-risk group tasks at the end, forming a hierarchical monitoring task sequence. The tiered monitoring task sequence specifies the monitoring order, frequency, and resource quotas for each facility, guiding the work scheduling of distributed monitoring nodes. When monitoring resources are scarce, resources are allocated according to the priority order of the tiered monitoring task sequence to ensure that the monitoring quality of high-risk groups is not affected.

[0070] Step S150: Based on the hierarchical monitoring task sequence, assess the remaining safe operating cycle of high-risk groups to identify fault risk points, and use the fault risk points to determine the early warning level parameters.

[0071] In some embodiments, the step of assessing the remaining safe operating period of the high-risk group and identifying fault risk points based on the hierarchical monitoring task sequence includes: obtaining the operating load records and rated service life of each facility in the high-risk group based on the hierarchical monitoring task sequence; performing load fluctuation amplitude analysis on the operating load records to obtain a fluctuation amplitude sequence; performing weighted calculation based on the fluctuation amplitude sequence to obtain the equivalent operating time; and assessing the remaining safe operating period and identifying fault risk points based on the equivalent operating time and the rated service life.

[0072] The operational load records and rated service life of each facility in the high-risk group were obtained based on a tiered monitoring task sequence. The tiered monitoring task sequence recorded the monitoring priority and data acquisition configuration for each facility. Historical operational data for each facility in the high-risk group was extracted from the monitoring database based on the facility number in the tiered monitoring task sequence. A certain building chiller unit was listed as a high-priority monitoring object in the tiered monitoring task sequence. Its operational load record showed that the unit operated at high load for extended periods during the high-temperature summer months due to bearing the cooling load of the entire building, with the load rate consistently maintained at a high level. The operational load record included the facility's instantaneous load rate, cumulative operating time, and number of start-ups and shutdowns. The load rate was expressed as a percentage of the facility's rated power. Simultaneously, the rated service life of each facility in the high-risk group was obtained. The rated service life was derived from the equipment nameplate or manufacturer's technical manual, expressed in hours or years. The rated service life is the design service life of the facility under standard operating conditions; the actual service life may vary depending on different operating conditions. The circulating water pumps in the tiered monitoring task sequence belong to the high-risk group. Their operating load records show that due to the parallel operation of the pipeline network, the pumps require frequent adjustments to the outlet valves, resulting in numerous load fluctuations and a shorter rated service life compared to the chiller units. The operating load records of the cooling tower fans show that they use variable frequency control to automatically adjust the speed according to cooling demand, and their rated service life falls between that of the chiller units and the circulating water pumps.

[0073] Load fluctuation amplitude analysis is performed on the operating load records to obtain the fluctuation amplitude sequence. The characteristics of load rate changes over time in the operating load records are analyzed to identify load abrupt changes and load fluctuations. Load abrupt changes refer to a significant change in load rate within a short period, such as a sudden increase from low load to high load or a sudden drop from high load to low load. The load change between adjacent sampling points in the operating load records is calculated. The operating load record of a building's circulating water pump shows a drastic change in load rate when the pipeline valves are activated. When maintenance personnel close a branch valve for pipeline maintenance, the remaining operating pumps need to handle a larger flow rate, and the load rate rises sharply within seconds. The absolute value of the load change is the fluctuation amplitude at that moment. The fluctuation amplitudes at each moment are arranged chronologically to form a fluctuation amplitude sequence, which describes the temporal distribution and intensity changes of facility load fluctuations. High-value periods in the fluctuation amplitude sequence correspond to moments when the facility experiences significant impact; frequent large fluctuations accelerate facility aging. The fluctuation amplitude sequence of the chiller unit shows that its load fluctuations are relatively stable. This unit uses variable frequency control to gradually load and unload, avoiding the impact of load abrupt changes on the compressor. The fluctuations in the cooling tower fan speed exhibit regular patterns at night. As cooling demand decreases at night, the frequency converter frequently adjusts the fan speed to maintain a stable cooling water temperature, resulting in periodic load fluctuations.

[0074] Equivalent runtime is obtained through weighted conversion based on the fluctuation amplitude sequence. A conversion factor is assigned to each time period in the fluctuation amplitude sequence, reflecting the degree of wear and tear on the facility's lifespan during that period. The larger the fluctuation amplitude, the higher the conversion factor, indicating that the wear and tear on the facility during that period is greater than during stable operation. The conversion factor is determined by the formula k = 1 + α × A, where k is the conversion factor, A is the percentage of fluctuation amplitude for that period, and α is the fluctuation loss factor, typically 0.02. The equivalent runtime is obtained by multiplying the actual runtime of each time period corresponding to the fluctuation amplitude sequence by the conversion factor and summing the results. The equivalent runtime reflects the true wear and tear on the facility. The formula for calculating the equivalent runtime is T_eq = Σ(t_i × k_i), where T_eq is the equivalent runtime, t_i is the actual runtime of the i-th time period, and k_i is the conversion factor for the i-th time period, with all units in hours. A building chiller unit employs a variable frequency soft start and gradual loading strategy. The fluctuation amplitude is relatively small for most periods in the fluctuation amplitude sequence, with a conversion factor close to 1. The equivalent running time is not significantly different from the actual running time, indicating that the stable operation mode effectively protects the equipment's lifespan. However, the circulating water pump, due to parallel operation of the pipeline network and frequent valve adjustments, exhibits several periods of high fluctuation in the fluctuation amplitude sequence. The conversion factor for these periods is significantly greater than 1, and the cumulative equivalent running time is significantly higher than the actual running time. This indicates that frequent load shocks accelerate the aging process of the pump, necessitating optimization of the pipeline network operation strategy to reduce the frequency of valve adjustments.

[0075] The remaining safe operating period is assessed based on the comparison of equivalent operating time and rated service life to identify potential failure points. The lifespan consumption ratio of the facility is calculated as the ratio of equivalent operating time to rated service life. The formula for calculating the remaining safe operating period is T_remain=(1-T_eq / T_life)×T_life / v_eq, where T_remain is the remaining safe operating period in days, T_eq is the equivalent operating time, T_life is the rated service life, and v_eq is the current equivalent operating rate, i.e., the equivalent operating time per day. The remaining safe operating period represents the time remaining before the facility reaches its lifespan limit under the current operating mode. A building chiller unit has been operating for many years, and its equivalent operating time is close to the design value of its rated service life. Furthermore, the unit operates at a high load continuously during the summer, resulting in a high daily equivalent operating rate. Substituting these values ​​into the formula, the calculated remaining safe operating period is short, below the safety threshold for critical facilities. This chiller unit is marked as a potential failure point, and the compressor bearings and motor windings are nearing their design life limits, posing a high risk of sudden failure. The remaining safe operating period is compared with a safety threshold, which is set according to the importance of the facility; critical facilities have longer safety thresholds, while general facilities have shorter thresholds. Facilities with a remaining safe operating period below the safety threshold are marked as failure risk points, representing facilities that are about to enter a period of high failure incidence. Although the fluctuation range sequence of cooling tower fans shows periodic fluctuations, their equivalent operating time still has a large margin before their rated service life, and their remaining safe operating period is much higher than the safety threshold; therefore, they are not marked as failure risk points. Circulating water pumps, due to frequent load fluctuations leading to a rapid accumulation of equivalent operating time, have a short remaining safe operating period and are marked as failure risk points; the wear condition of their mechanical seals and bearings needs to be monitored.

[0076] Early warning level parameters are determined using fault risk points. The number, location, and remaining safe operating period of all facilities marked as fault risk points are statistically analyzed. Early warning levels are assigned based on the urgency of the fault risk points: Level 1 (red) corresponds to a short remaining safe operating period, Level 2 (orange) to a medium remaining safe operating period, and Level 3 (yellow) to a relatively long remaining safe operating period. The early warning level parameters include four fields: warning level, affected facilities, remaining period, and recommended measures. Level 1 warnings require immediate shutdown for maintenance or activation of backup facilities; Level 2 warnings require a maintenance plan to be arranged in the short term; and Level 3 warnings require inclusion in the monthly maintenance plan. For example, a building chiller unit is considered a fault risk point. Due to the compressor's long service life and continuous high-load operation during summer, its remaining safe operating period falls within the Level 2 warning range. Therefore, the warning level parameter is set to Level 2, and the recommended measures are to arrange special inspections of the compressor bearings and motor windings, and to contact the manufacturer to prepare critical spare parts. As a potential point of failure, the circulating water pump, due to the frequent pressure shocks subjected to its mechanical seal over a long period, also triggers a Level 2 warning. Recommended measures include checking the wear of the sealing surface and preparing spare sealing components. Similarly, the fire pump, also a potential point of failure, triggers a higher-level warning due to its connection to life safety, even with a relatively long remaining safe operating period. The warning level parameters require prior consultation with a professional maintenance company for a comprehensive inspection.

[0077] Step S160: Based on the early warning level parameters, adaptive threshold calibration is performed on the distributed monitoring nodes to generate a monitoring benchmark. Based on the monitoring benchmark and the facility health distribution, a facility status monitoring report is output.

[0078] In some embodiments, the step of generating a monitoring benchmark by adaptive threshold calibration of the distributed monitoring nodes based on the warning level parameters includes: obtaining node sensitivity levels by performing sensitivity classification on the distributed monitoring nodes based on the warning level parameters; determining the monitoring area type according to the deployment location of the distributed monitoring nodes; determining a dynamic alarm threshold by performing time-sharing load benchmark matching based on the node sensitivity level and the monitoring area type; and generating a monitoring benchmark based on the dynamic alarm threshold.

[0079] Based on the warning level parameters, the sensitivity of distributed monitoring nodes is graded to obtain node sensitivity levels. According to the warning level of each facility in the warning level parameters, the distributed monitoring nodes monitoring that facility are assigned corresponding sensitivity levels, establishing a mapping relationship between warning levels and sensitivity levels. For example, a building chiller unit has a level 2 warning warning. Due to the long service life of its compressor, this unit has a high risk of failure. The distributed monitoring nodes monitoring this unit are set to a medium node sensitivity level, maintaining sensitivity to abnormal signals while avoiding false alarms due to normal load fluctuations. Distributed monitoring nodes corresponding to level 1 warning facilities are set to a high node sensitivity level. Nodes at this level are highly sensitive to even small changes in parameters and can detect early abnormal signals. Nodes corresponding to level 2 warning facilities are set to a medium node sensitivity level, striking a balance between sensitivity and stability. Nodes corresponding to level 3 warning facilities and facilities without warnings are set to a standard node sensitivity level, using industry-standard monitoring methods. The node sensitivity level determines the node's responsiveness to abnormal signals; nodes with a high node sensitivity level will trigger an alarm even for minor anomalies, while nodes with a standard node sensitivity level will only trigger an alarm for obvious anomalies. For fresh air handling units and terminal fan coil units without warning levels, the node sensitivity level of their corresponding distributed monitoring nodes is set to normal sensitivity. Fire pumps, due to their involvement in life safety, although the warning level parameter is Level 3, have their node sensitivity level increased to medium sensitivity to ensure timely detection of any abnormalities in fire-fighting equipment.

[0080] The monitoring area type is determined based on the deployment location of distributed monitoring nodes. The physical location and functional area of ​​each distributed monitoring node are analyzed, and nodes with similar locations or functions are grouped into the same monitoring area type. Monitoring area types include cold source areas, air conditioning room areas, power distribution areas, fire protection equipment areas, and general equipment areas, each with unique load characteristics and operating patterns. For example, distributed monitoring nodes in a building's chiller plant monitor chiller units, chilled water pumps, cooling water pumps, and cooling towers. These facilities are tightly connected through pipelines to form a complete refrigeration cycle; any abnormality in any facility can affect the entire refrigeration system. Therefore, the distributed monitoring nodes in this area are classified as a cold source area. The air conditioning room area monitoring type includes air conditioning units and fresh air units. Its load is closely related to the activity of people in the area it serves, increasing during peak hours and decreasing after get off work. The power distribution area monitoring type includes power supply facilities such as transformers and distribution cabinets, which have a fundamental impact on the operation of the entire building. The monitoring area type influences the alarm threshold setting strategy; critical areas use stricter thresholds, while general areas use standard thresholds. The basement electrical distribution room is classified as a monitoring area, which, although it has fewer facilities, can cause a large-scale power outage due to a single point of failure, affecting a wide range of areas.

[0081] For example, the step of determining the dynamic alarm threshold by matching the time-sharing load benchmark based on the node sensitivity level and the monitoring area type includes: obtaining a regional load characteristic curve based on the monitoring area type; dividing the regional load characteristic curve into peak and valley periods to obtain a time period load benchmark; configuring the time period load benchmark with a threshold floating based on the node sensitivity level to obtain a time-sharing threshold; and determining the dynamic alarm threshold according to the time-sharing threshold.

[0082] Regional load characteristic curves were obtained based on monitoring area types. Historical load data for facilities corresponding to each monitoring area type were statistically analyzed, covering a complete operating cycle of nearly 30 days. A day was divided into 24-hour periods, and the average load rate of facilities for that monitoring area type was calculated for each period. The load rates of the 24 periods were connected chronologically to form the regional load characteristic curve. The regional load characteristic curve visually displays the daily variation pattern of the facility load in the area, including peak and off-peak periods. The regional load characteristic curve corresponding to the monitoring area type of a building's cooling source area shows that air conditioning demand gradually increases after employees arrive at work in the morning, leading to a rise in chiller unit load. During the lunch break, the load slightly decreases due to reduced air conditioning demand in some areas. The load rises again during the afternoon work hours and gradually decreases to a low point after get off work in the evening. Different monitoring area types exhibit different morphological characteristics in their regional load characteristic curves. The regional load characteristic curve for the power distribution area shows two peak electricity consumption periods in the morning and afternoon, corresponding to the concentrated use of lighting, computers, and office equipment during peak office hours. The regional load characteristic curve of the air conditioning room area is similar to the overall trend of the cold source area, but the peak time is slightly delayed because it takes a certain amount of time for chilled water to be transported from the chiller station to the terminal.

[0083] The regional load characteristic curve is divided into peak and valley periods to obtain time-period load benchmarks. The morphological characteristics of the regional load characteristic curve are analyzed, and statistical methods are used to identify the boundaries of peak load periods, stable load periods, and low load periods. Periods with load rates higher than the regional load characteristic curve mean plus one standard deviation are classified as peak periods; periods with load rates lower than the regional load characteristic curve mean minus one standard deviation are classified as valley periods; and all other periods are classified as stable periods. A corresponding time-period load benchmark is set for each period, which is a load reference value for normal facility operation within that period. The time-period load benchmark for peak periods is set as the average value of the regional load characteristic curve for that period plus an appropriate margin, and the time-period load benchmark for valley periods is set as the average value for that period. After analysis, the regional load characteristic curve of a building's cooling source area is divided into peak periods during working hours, with the time-period load benchmark set at a higher level to accommodate peak cooling demand; nighttime is classified as valley periods, with the time-period load benchmark set at a lower level. Abnormal load increases at night may indicate equipment failure or abnormal startup. The time-period load benchmark is used for subsequent alarm threshold calculations; an alarm is triggered when the actual load deviates significantly from the time-period load benchmark. The time-based load benchmark for the air conditioning room area is close to that of the cold source area during peak hours, and slightly higher than that of the cold source area during off-peak hours because some areas need to maintain basic temperature and humidity control.

[0084] Time-sharing thresholds are obtained by configuring threshold fluctuations based on the load baseline during a given time period according to the node sensitivity level. The threshold fluctuation coefficient is set according to the node sensitivity level; the fluctuation coefficient is smaller for high-sensitivity nodes, moderate for medium-sensitivity nodes, and a standard value for conventional-sensitivity nodes. The formula for calculating the time-sharing threshold is Th=B×(1±f×k), where Th is the time-sharing threshold, B is the load baseline for the given time period, f is the basic fluctuation ratio, and k is the fluctuation coefficient corresponding to the node sensitivity level. The positive and negative signs correspond to the upper and lower thresholds, respectively. The smaller the fluctuation coefficient, the narrower the threshold range, the lower the tolerance for anomalies, and the easier it is to trigger an alarm. For example, the distributed monitoring nodes of a building chiller unit, due to their medium-sensitivity level, have a narrower time-sharing threshold range during peak hours compared to conventional nodes. Even slight anomalies in parameters such as compressor current and evaporator temperature difference can trigger an early warning, allowing maintenance personnel to detect signs of compressor performance degradation early. Time-sharing thresholds are calculated for each monitoring parameter of each distributed monitoring node in each time period, forming a complete threshold configuration table. The time-sharing threshold configuration for the circulating water pump is also calculated according to the mid-node sensitivity level, with a correspondingly tighter monitoring range for pressure and flow parameters, enabling timely alarms when there are abnormal fluctuations in outlet pressure or flow. The time-sharing threshold for the fresh air handling unit is calculated according to the conventional node sensitivity level, with a relatively loose monitoring range to avoid false alarms caused by normal fluctuations in air volume due to changes in indoor occupancy.

[0085] The dynamic alarm threshold is determined based on time-sharing thresholds. The time-sharing thresholds of each distributed monitoring node for each time period are integrated into a time-varying threshold curve, which is the dynamic alarm threshold. It automatically switches between different time-sharing thresholds over time. The dynamic alarm threshold uses the peak-hour time-sharing threshold configuration during peak hours and the off-hour time-sharing threshold configuration during off-hour hours, with a gradual transition at the time-segment boundaries to avoid abrupt threshold changes. Compared to fixed thresholds, dynamic alarm thresholds are better adapted to the daily variation patterns of facility load. For example, after implementing dynamic alarm thresholds on the distributed monitoring nodes in a building's cooling source area, during the daytime working hours, due to high cooling demand, the high-load operation of the chiller units is normal, and the upper limit of the dynamic alarm threshold is correspondingly widened to avoid false alarms. At night, during low-load periods, if there is an abnormal increase in load, the dynamic alarm threshold will trigger an alarm promptly, as a nighttime load increase usually indicates abnormal equipment startup or control system failure. The distributed monitoring nodes compare the collected data with the current dynamic alarm threshold in real time; if the data exceeds the threshold range, an alarm is triggered. The dynamic alarm threshold also supports differentiated configuration by date type, with different threshold curves for weekdays, weekends and holidays. The dynamic alarm threshold for the air conditioning room area is configured for off-peak hours throughout the day on holidays to avoid false alarms for normal standby states.

[0086] A monitoring baseline is generated based on dynamic alarm thresholds. The dynamic alarm thresholds of all distributed monitoring nodes are aggregated and coded according to a unified data format to form a monitoring baseline configuration file that can be directly loaded and executed by the monitoring system. The monitoring baseline contains four core fields: node number, monitoring parameters, time period division, and corresponding threshold, fully defining the alarm conditions for each parameter at each time period. For example, the monitoring baseline for a building configures dynamic alarm thresholds for multiple monitoring parameters such as compressor current, evaporator inlet and outlet temperature difference, and condensing pressure for the chiller unit. When any parameter exceeds the threshold range for the corresponding time period, the distributed monitoring node immediately reports an alarm, allowing maintenance personnel to quickly determine the direction of the fault based on the alarm parameter type. The monitoring baseline is stored in a standard format, facilitating download and local parsing by distributed monitoring nodes. When the warning level parameters are updated, the dynamic alarm thresholds of the affected nodes are recalculated, and the monitoring baseline is synchronously updated and pushed to the relevant nodes. The version number of the monitoring baseline increments with each update, ensuring that each node loads the latest configuration. The monitoring baseline also includes alarm priority settings to ensure that alarm information from critical facilities is processed first. The monitoring baseline for building electromechanical systems covers all distributed monitoring nodes and includes time-division threshold configurations for monitoring parameters such as current, temperature, pressure, and vibration.

[0087] A facility status monitoring report is generated by fusing monitoring benchmarks and facility health distribution. Alarm records from the monitoring benchmarks are linked and integrated with health scores from the facility health distribution to construct a multi-dimensional facility status assessment view. The number of alarms, alarm types, and alarm durations for each facility within the reporting period are statistically analyzed. Real-time monitoring indicators are extracted from the monitoring data corresponding to the monitoring benchmarks, and the current health score, health trend, and risk level of each facility are extracted from the facility health distribution. For example, a building's chiller unit triggered several alarms this week for high compressor current and abnormal evaporator temperature difference. Monitoring benchmark records show that the alarms were mainly concentrated during the afternoon peak hours. Combined with the declining health trend of the unit in the facility health distribution, it is determined that the unit's compressor load has increased due to a decrease in evaporator heat exchange efficiency, requiring evaporator cleaning and maintenance. Alarm statistics and health information are matched and fused according to facility number to generate a comprehensive assessment result including real-time status and long-term trends. The facility status monitoring report is presented in a hierarchical structure: the first layer is an overall building overview, the second layer is a system-level summary, and the third layer is facility details. Facility status monitoring reports are presented in a combination of charts and lists. Facility health distribution is displayed as a heat map, alarm trends are displayed as a line graph, and key facilities are highlighted in a list. Facility status monitoring reports are automatically generated and pushed to the operation and maintenance management platform every day to provide data support for operation and maintenance decisions.

[0088] To implement the above-described method embodiments, a smart monitoring method for building facilities based on the Internet of Things (IoT) is proposed to achieve the corresponding functions and technical effects. See also... Figure 2 , Figure 2 This diagram illustrates a structural block diagram of an IoT-based intelligent building facility monitoring system 200 according to an embodiment of this application. For ease of explanation, only the parts relevant to this embodiment are shown. The IoT-based intelligent building facility monitoring system 200 provided in this embodiment includes:

[0089] The data acquisition module 201 is used to collect sensor operation data and control command response data of building electromechanical facilities, and to perform operation cycle offset detection on the operation data and the command response data to form an abnormal facility response mode;

[0090] The fault tracing module 202 is used to extract the time sequence of abnormal events from the abnormal response mode of the facility, arrange the time sequence of abnormal events in chronological order to construct a fault time sequence causal chain, and form a facility association matrix based on the fault time sequence causal chain.

[0091] The status assessment module 203 is used to establish distributed monitoring nodes based on the facility association matrix, collect data on the working conditions of building electromechanical facilities based on the distributed monitoring nodes to generate a facility status feature set, and construct a facility health distribution using the facility status feature set.

[0092] The risk classification module 204 is used to analyze the facility deterioration rate in the facility health distribution to determine the deterioration rate, group the building electromechanical facilities according to the deterioration rate to identify high-risk groups and low-risk groups, and dynamically allocate monitoring resources to the high-risk groups and low-risk groups to construct a graded monitoring task sequence. The grouping of building electromechanical facilities according to the deterioration rate to identify high-risk groups and low-risk groups includes: performing time-series tracking of the deterioration rate to form a deterioration rate curve; performing adjacent facility deterioration spread analysis on the deterioration rate curve to identify deterioration spread source facilities; determining facility risk classification identifiers based on the deterioration spread source facilities and the deterioration rate; and identifying high-risk groups and low-risk groups based on the facility risk classification identifiers.

[0093] Life assessment module 205 is used to assess the remaining safe operating cycle of the high-risk group based on the graded monitoring task sequence, identify fault risk points, and use the fault risk points to determine the early warning level parameters.

[0094] The early warning output module 206 is used to perform adaptive threshold calibration on the distributed monitoring nodes based on the early warning level parameters to generate a monitoring benchmark, and to output a facility status monitoring report based on the fusion of the monitoring benchmark and the facility health distribution.

[0095] The aforementioned IoT-based intelligent building facility monitoring system 200 can implement an IoT-based intelligent building facility monitoring method according to the above-described method embodiments. The options in the above method embodiments are also applicable to this embodiment and will not be detailed here. The remaining content of this application's embodiments can be referred to the content of the above method embodiments, and will not be repeated in this embodiment.

[0096] The above embodiments are not an exhaustive list based on the present invention, and there may be many other embodiments not listed. Any substitutions and improvements made without departing from the concept of the present invention are within the protection scope of the present invention.

Claims

1. A smart monitoring method for building facilities based on the Internet of Things, characterized in that, include: Collect sensor operation data and control command response data of building electromechanical facilities, and perform operation cycle offset detection on the operation data and the command response data to form an abnormal facility response mode; Extract the time sequence of abnormal events from the abnormal response pattern of the facility, arrange the time sequence of the abnormal events in chronological order to construct a fault time sequence causal chain, and form a facility association matrix based on the fault time sequence causal chain. Based on the facility association matrix, a distributed monitoring node is established. Based on the distributed monitoring node, the working condition linkage of the building electromechanical facilities is collected to generate a facility status feature set. The facility status feature set is used to construct the facility health distribution. The facility deterioration rate is determined by analyzing the facility health distribution. The building electromechanical facilities are then grouped according to the deterioration rate to identify high-risk and low-risk groups. Monitoring resources are dynamically allocated to the high-risk and low-risk groups to construct a hierarchical monitoring task sequence. The step of grouping and identifying high-risk and low-risk groups of building electromechanical facilities according to the degradation rate includes: performing time-series tracking of the degradation rate to form a degradation rate curve; performing adjacent facility degradation spread analysis on the degradation rate curve to identify degradation spread source facilities; determining facility risk classification identifiers based on the degradation spread source facilities and the degradation rate; and identifying high-risk and low-risk groups based on the facility risk classification identifiers; wherein, the step of performing adjacent facility degradation spread analysis on the degradation rate curve to identify degradation spread source facilities includes: performing inter-facility degradation analysis on the degradation rate curve. Correlation analysis of degradation is performed to obtain the adjacency relationship of facilities; based on the adjacency relationship of facilities, synchronicity detection is performed to obtain the degradation synchronicity degree, which is calculated based on the degradation synchronicity degree formula S=R_max×τ_ref / (τ+τ_0), where R_max is the peak value of the cross-correlation function, τ is the time delay corresponding to the peak value, τ_ref is the reference time delay and takes the value of the typical fault propagation cycle of the system, and τ_0 is a smoothing constant; the degradation synchronicity degree is used to determine the propagation threshold to obtain suspected propagation facility pairs; the degradation time sequence analysis is performed on the suspected propagation facility pairs to identify the degradation propagation source facilities; Based on the hierarchical monitoring task sequence, the remaining safe operating cycle of the high-risk group is assessed to identify fault risk points, and the fault risk points are used to determine the early warning level parameters. Based on the warning level parameters, the distributed monitoring nodes are adaptively calibrated to generate a monitoring benchmark. Based on the monitoring benchmark and the facility health distribution, a facility status monitoring report is output.

2. The method according to claim 1, characterized in that, The step of detecting operational cycle offsets between the operational data and the instruction response data to form a facility response anomaly pattern includes: The actual operating cycle is obtained by performing periodic feature extraction on the operating data; The cycle offset is calculated by comparing the actual running cycle with the preset running cycle in the instruction response data. Perform positive and negative offset asymmetry analysis on the periodic offset to identify persistent offset events; The facility response anomaly pattern is formed based on the persistent offset events.

3. The method according to claim 1, characterized in that, The step of arranging the abnormal events in chronological order to construct a fault sequence causal chain includes: The abnormal event time sequence is timestamped to obtain the event occurrence sequence; The events are arranged in chronological order to form an ordered event chain. The ordered event chain is used to identify indirect related events across devices and systems by performing cross-level event progression association. A fault-sequence causal chain is constructed based on the indirectly related events and the ordered event chain.

4. The method according to claim 1, characterized in that, The process of collecting and generating a set of facility status features based on the distributed monitoring nodes for the coordinated operation of building electromechanical facilities includes: Based on the distributed monitoring nodes, the linkage triggering relationship between building electromechanical facilities is identified to form a linkage facility group; Based on the response time difference of each facility in the interconnected facility group, an interconnected time difference sequence is obtained; Deviation analysis is performed on the linked time difference sequence to identify linkage coordination characteristics; A facility status feature set is generated based on the aforementioned linkage and coordination characteristics.

5. The method according to claim 1, characterized in that, The step of assessing the remaining safe operating cycle of the high-risk group and identifying fault risk points based on the hierarchical monitoring task sequence includes: Based on the hierarchical monitoring task sequence, the operating load records and rated service life of each facility in the high-risk group are obtained; Perform load fluctuation amplitude analysis on the operating load records to obtain the fluctuation amplitude sequence; The equivalent runtime is obtained by weighted calculation based on the fluctuation amplitude sequence; Based on the equivalent operating time and the rated service life, assess the remaining safe operating cycle and identify potential failure points.

6. The method according to claim 1, characterized in that, The step of generating a monitoring benchmark by adaptive threshold calibration of the distributed monitoring nodes based on the early warning level parameters includes: Based on the aforementioned warning level parameters, the sensitivity levels of the distributed monitoring nodes are determined by classifying their sensitivity. The monitoring area type is determined based on the deployment location of the distributed monitoring nodes; Dynamic alarm thresholds are determined by time-sharing load benchmark matching based on the node sensitivity level and the monitoring area type. A monitoring benchmark is generated based on the dynamic alarm threshold.

7. The method according to claim 6, characterized in that, The determination of the dynamic alarm threshold based on time-sharing load benchmark matching according to the node sensitivity level and the monitoring area type includes: Obtain the regional load characteristic curve based on the monitoring area type; The load characteristic curve of the region is divided into peak and valley periods to obtain the time period load benchmark; Based on the node sensitivity level, the time-period load benchmark is configured with a threshold floating to obtain the time-sharing threshold; The dynamic alarm threshold is determined based on the time-sharing threshold.

8. An intelligent monitoring system for building facilities based on the Internet of Things, characterized in that, include: The data acquisition module is used to collect sensor operation data and control command response data of building electromechanical facilities, and to detect the operation cycle offset between the operation data and the command response data to form an abnormal response mode of the facility. The fault tracing module is used to extract the time sequence of abnormal events from the abnormal response pattern of the facility, arrange the time sequence of abnormal events in chronological order to construct a fault time sequence causal chain, and form a facility association matrix based on the fault time sequence causal chain. The status assessment module is used to establish distributed monitoring nodes based on the facility association matrix, collect data on the working conditions of building electromechanical facilities based on the distributed monitoring nodes to generate a facility status feature set, and use the facility status feature set to construct a facility health distribution. The risk classification module is used to analyze the facility deterioration rate in the facility health distribution to determine the deterioration rate, to group the building electromechanical facilities according to the deterioration rate to identify high-risk groups and low-risk groups, and to dynamically allocate monitoring resources to the high-risk groups and the low-risk groups to construct a graded monitoring task sequence. The step of grouping and identifying high-risk and low-risk groups of building electromechanical facilities according to the degradation rate includes: performing time-series tracking of the degradation rate to form a degradation rate curve; performing adjacent facility degradation spread analysis on the degradation rate curve to identify degradation spread source facilities; determining facility risk classification identifiers based on the degradation spread source facilities and the degradation rate; and identifying high-risk and low-risk groups based on the facility risk classification identifiers; wherein, the step of performing adjacent facility degradation spread analysis on the degradation rate curve to identify degradation spread source facilities includes: performing inter-facility degradation analysis on the degradation rate curve. Correlation analysis of degradation is performed to obtain the adjacency relationship of facilities; based on the adjacency relationship of facilities, synchronicity detection is performed to obtain the degradation synchronicity degree, which is calculated based on the degradation synchronicity degree formula S=R_max×τ_ref / (τ+τ_0), where R_max is the peak value of the cross-correlation function, τ is the time delay corresponding to the peak value, τ_ref is the reference time delay and takes the value of the typical fault propagation cycle of the system, and τ_0 is a smoothing constant; the degradation synchronicity degree is used to determine the propagation threshold to obtain suspected propagation facility pairs; the degradation time sequence analysis is performed on the suspected propagation facility pairs to identify the degradation propagation source facilities; The life assessment module is used to assess the remaining safe operating cycle of the high-risk group based on the graded monitoring task sequence, identify fault risk points, and use the fault risk points to determine the early warning level parameters. The early warning output module is used to perform adaptive threshold calibration on the distributed monitoring nodes based on the early warning level parameters to generate a monitoring benchmark, and to output a facility status monitoring report based on the fusion of the monitoring benchmark and the facility health distribution.

Images