Building equipment and building information model matching method

Abstract

The invention discloses a building equipment and building information model matching method, particularly relates to the field of intelligent operation and maintenance of building equipment, and is used for solving the problem of dynamic matching of equipment time sequence behaviors and a BIM model. In combination with correlation strength calculation of transfer entropy and time sequence closeness two-parameter fusion, a time sequence dependency relationship between equipment behaviors and model attributes is quantified, and explicit expression of a complex causal chain is realized; extracting a core migration path through a maximum flow-minimum cut theorem of the weighted directed graph, generating an executable state migration rule, binding the executable state migration rule to a BIM entity dynamic attribute, and completing structured matching of equipment actual operation logic and model parameters; dynamically adjusting a confidence interval threshold value through the weight accumulation deviation degree; and finally, the deep matching capability of the equipment full-life-cycle behavior and the BIM model is formed, and cross-generation upgrading of the building equipment and the information model from static mapping to dynamic collaboration is promoted.

Description

Technical Field

[0001] The present invention relates to the field of intelligent operation and maintenance of construction equipment, and more specifically, to a method for matching construction equipment with a building information model. Background Art

[0002] In the field of intelligent operation and maintenance of building equipment, equipment failures are often triggered by time-dependent events, such as continuous overload. Traditional BIM models are limited by static attribute architectures and can only record instantaneous equipment state parameters. They lack the ability to model historical event sequences, making it difficult to trace the root cause of failures.

[0003] However, existing equipment and BIM matching methods rely on periodic snapshot updates or fixed threshold rule mapping. They are unable to capture nonlinear causal relationships between events, such as the lag effect of vibration exceeding the standard and sealing failure, and are also difficult to adapt to rule drift caused by equipment aging and environmental disturbances. This leads to systematic deviations between model data and physical equipment behavior, seriously restricting the application value of BIM in complex operation and maintenance scenarios.

[0004] In order to solve the above problems, a technical solution is now provided. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method for matching building equipment with a building information model. The method collects equipment operating parameters and event sequences in full based on the space-time matrix, and combines the calculation of correlation strength by fusing the two parameters of transfer entropy and temporal closeness to quantify the temporal dependency between equipment behavior and model attributes, and realizes the explicit expression of complex causal chains. The method extracts the core migration path through the maximum flow-minimum cut theorem of the weighted directed graph, generates executable state migration rules and binds them to the dynamic attributes of the BIM entity, and completes the structured matching between the actual operation logic of the equipment and the model parameters. The closed-loop feedback mechanism based on the real-time event stream dynamically adjusts the confidence interval threshold through the weighted cumulative deviation to ensure the dynamic synchronization of the equipment state evolution and the model rules, and adapt to the aging of the equipment and the fluctuation of the working conditions. Finally, the deep matching capability of the equipment's full life cycle behavior and the BIM model is formed, and the cross-generational upgrade of the building equipment and information model from static mapping to dynamic collaboration is promoted to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] S1. Collect real-time status data of building equipment through the device control bus, encode each data point into a three-dimensional vector containing a timestamp, event type, and value interval, and construct a spatiotemporal matrix based on the time axis, event type axis, and value axis for storage;

[0008] S2. Calculate the transfer entropy and temporal closeness of adjacent events based on the spatiotemporal matrix, fuse the two through asymmetric cross entropy to generate a correlation strength value, and select continuous event sequences whose strength values exceed the confidence interval threshold as key sequences;

[0009] S3. Map the key sequence into a weighted directed graph, with event types as nodes and association strength values as edge weights. Apply the maximum flow-minimum cut theorem to extract the core migration path, convert the path into state migration rules, and bind them to the dynamic rule attributes of the BIM entity.

[0010] S4. Continuously monitor the real-time event stream and match the current event sequence with the core migration path in the weighted directed graph. If the accumulated path weight exceeds the judgment standard value and deviates from the historical path topology, update the BIM model status attributes, trigger an alarm, and adjust the confidence interval threshold based on the deviation.

[0011] In a preferred embodiment, step S1 includes the following contents:

[0012] Real-time status data of building equipment is collected through the device control bus, and each data point is encoded as a three-dimensional vector containing a timestamp, event type, and numerical interval. A space-time matrix is constructed according to the time axis, event type axis, and numerical axis for storage. Data collection includes operating parameters, fault alarms, maintenance records, and discrete states. The timestamp records the moment of data collection, the event type is represented by a predefined code, and the numerical interval discretizes the continuous parameters according to the equipment specifications. The space-time matrix uses sparse storage technology to record the number of event occurrences and processes new data through an incremental update strategy.

[0013] In a preferred embodiment, step S2 includes the following:

[0014] The transfer entropy and temporal closeness of adjacent events are calculated based on the space-time matrix. The transfer entropy and temporal closeness are fused through asymmetric cross entropy to generate the correlation strength value. The adjacent event pairs whose correlation strength values exceed the confidence interval threshold are screened, and the screened adjacent event pairs are connected to form a continuous event sequence. The corresponding continuous event sequence is defined as a key sequence.

[0015] In a preferred embodiment, step S2 further includes the following:

[0016] By counting the frequencies of adjacent event pairs, a joint probability distribution is constructed. The joint probability of the preceding and subsequent events occurring simultaneously, the conditional probability of the subsequent event under given conditions of the preceding event, and the marginal probability of the subsequent event are calculated. These probability values are used to evaluate the predictive ability of the preceding event on the occurrence of the subsequent event. The specific method is to take the logarithm of the ratio of the joint probability to the conditional probability and the marginal probability, and then perform a weighted summation on all event pairs to finally obtain the transfer entropy value.

[0017] In a preferred embodiment, step S2 further includes the following:

[0018] Determine the time periods in which the predecessor and successor events occur, then calculate the difference between the two time periods to obtain the time interval between them. Then, take the inverse of the corresponding time interval to generate a closeness index that increases as the time interval decreases, and finally obtain the time series closeness.

[0019] In a preferred embodiment, step S3 includes the following contents:

[0020] The screened key event sequences are mapped into a weighted directed graph, with event types as nodes and the association strength values of adjacent event pairs as edge weights. The maximum flow-minimum cut theorem is applied to extract the core migration path in the weighted directed graph. The source point is set as the initial event type and the sink point is set as the target event type. The main information flow from the source point to the sink point is calculated and the minimum cut set is identified. The connection path is traced back from the minimum cut set to extract the core migration path.

[0021] In a preferred embodiment, step S3 further includes the following:

[0022] The core migration path is converted into state transition rules, the event type sequence in the path is decomposed into adjacent event pairs to generate migration rules, and the migration rules and association strength values are bound to the dynamic rule attributes of the building information model entity.

[0023] In a preferred embodiment, step S4 includes the following contents:

[0024] The real-time status data of building equipment is continuously monitored through the device control bus, and the monitored data is encoded into data units containing timestamps, event types and numerical intervals. The event stream is constructed in the order of timestamps. A time window is set from the event stream to extract the current event sequence, and the corresponding path is matched in the weighted directed graph. The sum of the association strength values between adjacent event types in the path is calculated as the weight accumulation value.

[0025] In a preferred embodiment, step S3 further includes the following:

[0026] A judgment standard value is set based on the historical weight cumulative distribution of the core migration path. When the weight cumulative value does not exceed the judgment standard value, it is determined that the current event sequence deviates from the core migration path, the building information model status attribute is updated as abnormal, and an alarm is triggered.

[0027] In a preferred embodiment, step S3 further includes the following:

[0028] The confidence interval threshold of the correlation strength calculation is adjusted according to the degree of deviation to ensure that the model adapts to changes in the operating status of building equipment.

[0029] The technical effects and advantages of the method for matching construction equipment with a building information model of the present invention are as follows:

[0030] The beneficial effect of the present invention is that it constructs a dynamic matching mechanism between building equipment and building information model (BIM). Through time-series data encoding and event correlation analysis, it solves the technical bottleneck of static separation between equipment status and model attributes in traditional methods. Based on the full acquisition of equipment operating parameters and event sequences in the space-time matrix, combined with the calculation of correlation strength by fusing the two parameters of transfer entropy and time series tightness, the temporal dependency between equipment behavior and model attributes is accurately quantified, and the explicit expression of complex causal chains is achieved. The core migration path is extracted through the maximum flow-minimum cut theorem of the weighted directed graph, and executable state migration rules are generated and bound to the dynamic attributes of the BIM entity to complete the structured matching of the actual operation logic of the equipment and the model parameters. Based on the closed-loop feedback mechanism of the real-time event stream, the confidence interval threshold is dynamically adjusted through the weighted cumulative deviation to ensure the dynamic synchronization of equipment state evolution and model rules, and adapt to equipment aging and working condition fluctuations. Finally, the deep matching capability of the equipment's full life cycle behavior and the BIM model is formed, which improves the accuracy of equipment status tracing, abnormal warning and maintenance decision-making, and promotes the cross-generational upgrade of building equipment and information models from static mapping to dynamic collaboration. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 The figure is a flow chart of a method for matching building equipment with a building information model according to the present invention. DETAILED DESCRIPTION

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0033] Example 1: Figure 1 The present invention provides a method for matching building equipment with a building information model, comprising:

[0034] S1. Collect real-time status data of building equipment through the device control bus, encode each data point into a three-dimensional vector containing a timestamp, event type, and value interval, and construct a spatiotemporal matrix based on the time axis, event type axis, and value axis for storage;

[0035] S2. Calculate the transfer entropy and temporal closeness of adjacent events based on the spatiotemporal matrix, fuse the two through asymmetric cross entropy to generate a correlation strength value, and select continuous event sequences whose strength values exceed the confidence interval threshold as key sequences;

[0036] S3. Map the key sequence into a weighted directed graph, with event types as nodes and association strength values as edge weights. Apply the maximum flow-minimum cut theorem to extract the core migration path, convert the path into state migration rules, and bind them to the dynamic rule attributes of the BIM entity.

[0037] S4. Continuously monitor the real-time event stream and match the current event sequence with the core migration path in the weighted directed graph. If the accumulated path weight exceeds the judgment standard value and deviates from the historical path topology, update the BIM model status attributes, trigger an alarm, and adjust the confidence interval threshold based on the deviation.

[0038] In the intelligent operation and maintenance scenario of building equipment, the temporal dependence of equipment operating status and failure modes is significant. For example, the protective shutdown of an elevator may be triggered by the accumulation of continuous overload events. Traditional building information models (BIM) are limited to recording the instantaneous status of equipment and lack the ability to temporally model historical event sequences, resulting in data gaps during root cause analysis of faults. Existing methods mostly rely on periodic snapshots or fixed threshold rules, and are unable to capture nonlinear causal relationships between events or adapt to dynamic changes in the operating environment. The present invention proposes a method for matching building equipment with BIM. Step S1 collects real-time status data and constructs a spatiotemporal matrix to provide a temporal data basis for subsequent event correlation analysis and model adaptive updates, thereby improving decision support capabilities in complex operation and maintenance scenarios.

[0039] Step S1 includes the following contents:

[0040] Real-time status data of building equipment is collected through the device control bus. The collected content includes the equipment's operating parameters, fault alarms, maintenance records, and discrete states. Operating parameters cover physical quantities such as temperature, pressure, current, and vibration; fault alarms include abnormal conditions such as overload and overheating; maintenance records contain information such as repair time and component replacement; discrete states refer to discrete equipment operating states such as switch status and operating mode. Each data point is accompanied by a precise timestamp to record the specific moment of data collection. The sampling frequency of the collection is pre-set according to the type of equipment and monitoring requirements to ensure the real-time nature of the data. The collected data is transmitted to the central processing unit via the device control bus for subsequent encoding and storage processing.

[0041] Each collected data point is encoded as a vector consisting of three dimensions: timestamp, event type, and numerical interval vector. The timestamp is directly based on the moment of data collection. Event types are represented using predefined classification codes. For example, temperature data and fault alarms are categorized as different code values, ensuring the uniqueness and scalability of each event type. Numerical intervals are achieved by dividing continuous operating parameter values into predefined ranges. The division rules are determined by the equipment's technical specifications. For example, temperature values are divided into three ranges: low, medium, and high.

[0042] A multidimensional spatiotemporal matrix is constructed based on the encoded vectors to organize and store data. The spatiotemporal matrix is indexed along three axes: the time axis, the event type axis, and the value axis. The time axis divides continuous time into time periods with fixed intervals, such as every minute or every hour as a time unit; the event type axis lists all predefined event types, such as temperature anomalies and fault alarms; and the value axis lists the numerical intervals corresponding to each event type, such as low temperature, medium temperature, and high temperature. Each element of the spatiotemporal matrix records the number of events that occurred in a specific time period, a specific event type, and a specific numerical interval. For example, if a high temperature event occurs within a certain time period, the value of the corresponding element is recorded as 1.

[0043] Due to the sparse distribution of events, the spatiotemporal matrix uses sparse storage technology, recording only elements with an event occurrence count greater than zero, while omitting zero-valued elements to reduce storage space and computing resource consumption. Data updates employ an incremental strategy: each time a new data point is collected, it is mapped to the corresponding element in the spatiotemporal matrix and the value of that element is incremented by 1. This update method avoids recalculating historical data and processes only newly added data.

[0044] Step S1 collects real-time status data via the device control bus, encodes it into a three-dimensional vector containing a timestamp, event type, and numerical interval, and constructs a spatiotemporal matrix, laying the data foundation for subsequent time-series analysis. However, the spatiotemporal matrix only provides the time and type distribution of events and cannot directly reveal the causal relationships and evolutionary patterns between events, making it difficult to support fault prediction and maintenance decisions. Step S2 aims to quantify the nonlinear causal relationships and temporal regularities between events based on the spatiotemporal matrix, screen key event sequences, and provide a basis for extracting state transition rules in step S3.

[0045] In order to extract the deep relationship between adjacent events from the space-time matrix, calculating the transfer entropy and temporal closeness becomes a necessary and reasonable quantitative means.

[0046] Step S1 collects real-time status data via the device control bus, encodes it into a three-dimensional vector containing a timestamp, event type, and numerical interval, and constructs a spatiotemporal matrix to store the temporal changes in device status, laying the data foundation for subsequent analysis. However, the spatiotemporal matrix only provides a raw record of event sequences and cannot directly reveal the causal relationships and temporal patterns between events. Therefore, step S2 requires further quantification of the correlation strength between adjacent events and identification of key event sequences to support fault prediction and maintenance decision-making.

[0047] In building equipment operation and maintenance scenarios, the evolution of equipment operating states and failure modes exhibits significant temporal dependencies. This characteristic provides the theoretical basis for the rationality of calculating the transfer entropy and temporal closeness of adjacent event pairs. Equipment failures or abnormal conditions are often not isolated events, but rather the result of a series of cumulative operational events. For example, a protective elevator shutdown may result from the ongoing impact of multiple overload events. This temporal correlation demonstrates that the direct relationship between adjacent events is crucial for revealing the cause of the failure. By analyzing adjacent event pairs, we can effectively capture the dynamic evolution of events, quantify the impact of the preceding event on the subsequent event, and the temporal continuity between the two. Specifically, transfer entropy, based on the principles of information theory, reveals potential causal relationships by evaluating the amount of information transferred between events. Temporal closeness, by measuring the time intervals between events, reflects their temporal proximity. This analytical approach not only conforms to the physical laws of equipment operation but also adapts to the complex temporal data characteristics of building operation and maintenance, providing a scientific basis for subsequent fault prediction and maintenance decisions.

[0048] Step S2 includes the following contents:

[0049] S2.1. In a space-time matrix, adjacent events are two events that occur consecutively on the timeline. The first event is called the predecessor event, and the event that immediately follows it is called the successor event, with no other events intervening. Each event is represented by a three-dimensional vector consisting of a time period, an event type, and a numerical interval. Adjacent events are considered to be adjacent only if the successor event's time period is immediately adjacent to the predecessor's, or if both events fall within the same time period but have different event types.

[0050] S2.2, calculation of transfer entropy:

[0051] Transfer entropy measures the amount of information transferred from a preceding event to a subsequent event, reflecting the ability of a preceding event to predict the occurrence of a subsequent event. The calculation process first counts the frequencies of all pairs of adjacent events in the spatiotemporal matrix. Based on these frequencies, the joint probability of the preceding and subsequent events occurring simultaneously is calculated. Next, the conditional probability of the subsequent event occurring given the preceding event is calculated, as well as the marginal probability of the subsequent event occurring independently. The transfer entropy value is then derived by comparing the relationship between the joint, conditional, and marginal probabilities. To avoid zero probability, a small smoothing factor is added to the frequency statistics, using Laplace smoothing techniques for adjustment.

[0052] Transfer entropy quantifies the causal relationships between events, identifying which preceding events significantly influence subsequent events. This information-theoretic quantification method outperforms traditional correlation analysis by capturing nonlinear dependencies, thereby improving the accuracy and depth of analysis and providing a reliable basis for device status prediction. By counting the frequencies of adjacent event pairs and constructing probability distributions, the direction and intensity of information flow between events can be assessed.

[0053] For example, the transfer entropy is calculated as follows:

[0054] In the space-time matrix, adjacent events refer to two events E that occur consecutively on the time axis. A and E B , that is, E B Following E A occurs, with no other events intervening. Each event is represented by a three-dimensional vector: E A =(t A ,e A ,v A ) and E B =(t B ,e B ,v B ),in:

[0055] t is the timestamp, indicating the time period when the event occurs;

[0056] e is the event type;

[0057] v is a numerical interval, which indicates the specific measurement value range of the event.

[0058] Adjacent condition requirement t B =t A +1 (adjacent time period) or t B =t A (different events within the same time period).

[0059] Transfer entropy is used to measure E A To E B The amount of information transmitted, reflecting EA To E B The predictive power of occurrence. The calculation formula is:

[0060]

[0061] in:

[0062] P(e A ,e B ) is E A and E B Joint probability of simultaneous occurrence;

[0063] P(e B |e A ) is given by E A When E occurs B Conditional probability of occurrence;

[0064] P(e B ) is E B The marginal probability of occurrence.

[0065] Count the adjacent event pairs (E A ,E B ) and calculate the joint probability:

[0066]

[0067] Conditional probability: in

[0068] Marginal probability:

[0069] In order to avoid calculation failure caused by zero probability, Laplace smoothing is introduced:

[0070]

[0071] where N e is the total number of event types.

[0072] Transfer entropy The larger the value, the more E A To E B The more significant the causal influence.

[0073] S2.3, calculation of timing tightness:

[0074] Temporal closeness measures the temporal proximity of a preceding and succeeding event, reflecting the strength of their temporal association. The calculation method first determines the time interval between the preceding and succeeding events, expressed in time units. Then, the reciprocal of this time interval is converted to a temporal closeness value. The smaller the time interval, the higher the temporal closeness.

[0075] Temporal closeness directly reflects the temporal continuity of events and helps identify closely related event sequences. This quantification method is simple and intuitive, effectively capturing temporal patterns between events. By calculating the reciprocal of the time interval, temporal closeness is converted into a comparable numerical value.

[0076] For example, timing tightness is calculated as follows:

[0077] Timing tightness is used to measure E A and E B The degree of temporal proximity reflects the strength of temporal correlation. The calculation formula is:

[0078]

[0079] in:

[0080] For E A and E B The time interval between them, in time periods;

[0081] If t B =t A +1, then

[0082] If t B =t A ,but

[0083] Extract E from the space-time matrix A and E B Timestamp t A and t B ;

[0084] Calculated based on time period index

[0085] The range of is (0,1], and the larger the value, the closer the events are in time.

[0086] S2.4, Calculation of association strength:

[0087] Correlation strength is generated by fusing transfer entropy and temporal closeness using asymmetric cross-entropy. It comprehensively assesses the degree of association between a preceding and a subsequent event. The calculation multiplies the ratio of transfer entropy and the logarithm of temporal closeness to obtain the correlation strength. To avoid zero denominators, a small smoothing factor is introduced. A larger correlation strength value indicates a stronger association between the preceding and subsequent events.

[0088] For example, the strength of association is calculated as follows:

[0089] Generate association strength value by fusing transfer entropy and temporal closeness through asymmetric cross entropy The calculation formula is:

[0090]

[0091] in:

[0092] Provides causal information transfer;

[0093] Provides temporal proximity;

[0094] Mapping the temporal closeness to the real domain amplifies the impact of close events.

[0095] ∈ is a smoothing factor to avoid the denominator being zero.

[0096] S2.5. After calculating the correlation strength of all adjacent event pairs, set a confidence interval threshold and filter out adjacent event pairs whose correlation strength exceeds the threshold. These filtered adjacent event pairs are connected in chronological order to form a continuous event sequence, called a key event sequence, which is used to reflect the strong correlation pattern in the device state evolution.

[0097] For example, the screening process can be illustrated by:

[0098] For all adjacent event pairs in the space-time matrix (E A ,E B )calculate

[0099] Set the confidence interval threshold θ, take The upper quartile of

[0100] Filter out the satisfied Adjacent event pairs;

[0101] Connect adjacent event pairs that meet the conditions into a continuous event sequence, such as E A →E B →E C ,in and

[0102] These continuous event sequences are called key sequences, which reflect the strong correlation patterns in the evolution of device states.

[0103] The transfer entropy and temporal closeness of adjacent events are calculated based on the space-time matrix, where adjacent events refer to two consecutive events on the timeline. The transfer entropy is used to measure the amount of information transferred from the preceding event to the succeeding event, and the temporal closeness is used to measure the temporal proximity of the preceding and succeeding events. The transfer entropy and temporal closeness are fused through asymmetric cross-entropy to generate an association strength value. Adjacent event pairs with association strength values exceeding the confidence interval are screened, and the screened adjacent event pairs are connected to form a continuous event sequence. This continuous event sequence is defined as a key sequence for subsequent analysis.

[0104] Step S2 calculates the transfer entropy and temporal closeness of adjacent events based on the spatiotemporal matrix, generates correlation strength values, and screens for key event sequences, revealing strong correlation patterns between events. However, key event sequences only reflect the strength of correlations between events and do not directly reflect the specific paths and patterns of equipment state evolution. Therefore, step S3 further processes the key event sequences, constructs a state evolution model, and binds it to BIM entities to enable dynamic modeling and prediction of equipment states.

[0105] The following is a description of the specific processing technology logic of step S3. Step S3 takes the key event sequence selected in step S2 as input, maps it into a weighted directed graph, extracts the core migration path, converts the core migration path into a state transition rule, and finally binds it to the dynamic rule attributes of the building information model entity, thereby achieving modeling of the device state evolution.

[0106] Step S3 includes the following contents:

[0107] S3.1, map the key event sequence into a weighted directed graph:

[0108] The selected key event sequences are mapped into a weighted directed graph. Nodes in the weighted directed graph represent event types within the key event sequence, such as overload or temperature anomaly, with each event type corresponding to a separate node. Directed edges represent the relationships between adjacent events within the key event sequence, with edges pointing from the preceding event to the succeeding event, reflecting the order in which events migrate along the time series. Edge weights are derived from the calculated association strength value, which is calculated by fusing transfer entropy and temporal closeness via asymmetric cross entropy. This value quantifies the degree of association between adjacent events. Finally, a complete weighted directed graph is constructed by summarizing all nodes, directed edges, and their corresponding weights.

[0109] Mapping key event sequences into weighted directed graphs clearly demonstrates the temporal order and degree of correlation between events, facilitating subsequent path analysis and the extraction of state transition rules. The structured representation of weighted directed graphs supports complex network analysis methods, enhancing data processing flexibility and analytical depth.

[0110] S3.2, apply the maximum flow-minimum cut theorem to extract the core migration path:

[0111] After constructing a weighted directed graph, the maximum flow-minimum cut theorem is applied to extract core migration paths. First, the source and sink nodes of the weighted directed graph are set. The source node typically represents the initial state of the equipment before normal operation or failure, and the sink node represents the target failure state or critical abnormal state. Next, the Ford-Fulkerson method is used to calculate the main information flow from the source node to the sink node. The main information flow reflects the main channel of state migration. Subsequently, the minimum cut set in the weighted directed graph is identified. The minimum cut set consists of a set of directed edges whose sum of weights is equal to the main information flow, and these edges partition the weighted directed graph into two subsets: one containing the source node and the other containing the sink node. Finally, by tracing the connection paths of the directed edges in the minimum cut set, the main migration sequence from the source node to the sink node is extracted and recorded as the core migration path.

[0112] Applying the maximum flow-minimum cut theorem to extract core migration paths effectively identifies key connections and primary paths in weighted directed graphs, screening out the event transition sequences that have the greatest impact on device state evolution. Leveraging network flow theory, the most influential state transition paths are automatically extracted, providing a scientific basis for fault prediction and maintenance decision-making.

[0113] S3.3, convert the core migration path into state migration rules:

[0114] After extracting the core migration path, it is converted into state transition rules. The core migration path is represented as a sequence of event types. Each pair of adjacent events in the sequence constitutes a state transition rule. The rule is formulated as follows: if the current event is a predecessor event, then the next event is likely a successor event. Each rule is associated with a credibility value, which is the strength of the association between adjacent events in the core migration path. These state transition rules are stored as key-value pairs in the dynamic rule attributes of the building information model entity, allowing the building information model entity to query and update device status based on real-time event streams.

[0115] This approach converts core migration paths into state transition rules, simplifying complex path information into actionable rule sets. This facilitates dynamic updates of building information models and fault warnings. This rule-based representation supports real-time monitoring and rapid response, improving the efficiency and accuracy of operational and maintenance decisions.

[0116] Step S3 maps the key event sequence into a weighted directed graph, applies the maximum flow-minimum cut theorem to extract the core migration path, and converts this core migration path into state transition rules, binding them to the dynamic rule attributes of the building information model entities. This completes the dynamic modeling of equipment state evolution. The output of step S3 provides a precise rule basis for real-time monitoring and fault warning in step S4, fully demonstrating the technical advantages of this invention in equipment state analysis.

[0117] Step S3 maps the key event sequence into a weighted directed graph, extracts the core migration path, and converts it into state transition rules, completing the dynamic modeling of device state evolution. However, this only implements the construction of the state evolution model and does not involve real-time monitoring and fault warning. Step S4 aims to continuously monitor the real-time event stream and match the core migration path to determine device state deviations, update the building information model state attributes, trigger alarms, and dynamically adjust the association strength calculation parameters to achieve real-time monitoring of device status and fault warning.

[0118] Based on the weighted directed graph and state transition rules generated in step S3, step S4 continuously monitors the real-time event stream, matches the current event sequence with the core transition path, determines whether the device state deviates from the normal evolution path, updates the building information model state attributes based on the deviation, triggers an alarm signal, and dynamically adjusts the confidence interval threshold of the association strength calculation in step S2. The following is the detailed technical logic of step S4:

[0119] S4.1, continuously monitor the real-time event stream:

[0120] Real-time status data of building equipment is collected via the device control bus. Each data point is encoded into a data unit consisting of a timestamp, an event type, and a numerical range, consistent with the encoding method defined in step S1. The collected data units are arranged in chronological order by timestamp, forming a continuous event stream. The event type sequence in the event stream serves as the basis for subsequent path matching.

[0121] Continuously monitoring real-time event streams ensures timely access to the latest status data for building equipment, ensuring real-time and accurate monitoring. The construction of event streams provides a dynamically updated data foundation for subsequent path matching, enabling equipment status analysis and decision-making based on the latest operational information.

[0122] S4.2, matching the current event sequence with the core migration path:

[0123] Based on the event stream, a fixed-length time window is set, and several recent event types are extracted from the event stream to form the current event sequence. Next, the path corresponding to the current event sequence is searched in the weighted directed graph generated in step S3. The weighted directed graph uses event types as nodes, and the association strength values between adjacent event types serve as the edge weights connecting the nodes. The path matching process compares the current event sequence with the paths in the weighted directed graph and calculates the sum of the edge weights between all adjacent event types in the matching path, which is called the weight accumulation value. A judgment standard value is pre-set, which is determined based on the distribution of historical weight accumulation values of the core migration path in a similar time window in step S3. If the weight accumulation value exceeds the judgment standard value, the current event sequence is judged to be structurally similar to the core migration path, indicating that the device status has not deviated from the normal evolution path. If the weight accumulation value does not exceed the judgment standard value, the current event sequence is judged to have deviated from the structure of the core migration path and needs to enter the subsequent processing flow.

[0124] By extracting the current event sequence through a fixed-length time window and matching it with the core migration path, we can monitor the evolution of device status in real time. The calculation of weighted cumulative values and comparison with judgment criteria provide a quantitative basis for assessing the degree of device status deviation, ensuring the objectivity and accuracy of monitoring results.

[0125] S4.3, Update building information model status attributes and trigger alarms:

[0126] Based on the path matching results and the cumulative weight value, the state attributes of the corresponding entities in the building information model are updated. If the cumulative weight value exceeds the judgment standard, the state attribute of the building information model is updated to the latest event type in the current event sequence. If the cumulative weight value does not exceed the judgment standard, the state attribute of the building information model is marked as abnormal, and the current event sequence that caused the deviation is recorded. At the same time, when the cumulative weight value does not exceed the judgment standard, an alarm signal is triggered, notifying the operation and maintenance personnel. The alarm signal includes the current event sequence that caused the deviation, the degree of deviation, and the maintenance measures recommended based on the state transition rules.

[0127] Real-time updates to the BIM's status attributes ensure that model data remains consistent with the equipment's actual operating status, enhancing the model's practicality in real-world applications. Triggering alarms promptly notifies maintenance personnel to take necessary maintenance measures, preventing potential failures and ensuring the safe operation of building equipment.

[0128] S4.4, adjust the confidence interval threshold for the association strength calculation based on the degree of deviation:

[0129] When a device state is detected to have deviated from its normal evolution path, the confidence interval threshold used in the association strength calculation in step S2 is adjusted based on the degree of deviation. This adjustment process uses the ratio of the accumulated weight value to the judgment standard value as a measure of the degree of deviation, and an adjustment factor is determined based on this ratio. The size of the adjustment factor is proportional to the degree of deviation, and the new confidence interval threshold is calculated by multiplying the original confidence interval threshold by the adjustment factor. The adjusted confidence interval threshold is applied to the association strength screening process in step S2, updating the calculation logic of the key event sequence to adapt to the changes in the device's operating status.

[0130] Dynamically adjusting the confidence interval threshold allows the correlation strength calculation to adapt to changes in the equipment operating environment, thereby improving the robustness and accuracy of the entire analysis model. This threshold adjustment mechanism ensures that the screening process for key event sequences can promptly reflect the evolving trends of equipment status, providing more reliable data support for fault prediction.

[0131] Step S4 continuously monitors the real-time event stream and matches the core migration path, enabling real-time monitoring of building equipment status and fault warnings. Specifically, based on the weighted directed graph generated in step S3, step S4 determines whether the current event sequence deviates from the normal evolution path, updates the state attributes of the building information model, and triggers an alarm signal. Simultaneously, the confidence interval threshold for the correlation strength calculation in step S2 is adjusted based on the degree of deviation, ensuring that the model can adapt to changes in the equipment operating environment. Step S4 provides dynamic and efficient technical support for the intelligent operation and maintenance of building equipment, and has high practical value.

[0132] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0133] It should be noted that the system of the present invention can be deployed on the device itself to realize embedded applications, and can also be run on a PC or other terminal with a user interface, thereby meeting a variety of hardware environments and usage requirements.

[0134] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

[0135] It should be noted that, in this document, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article or device. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device that includes the element.

[0136] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A method for matching building equipment with a building information model, characterized in that: Including steps: S1. Collect real-time status data of building equipment through the device control bus, encode each data point into a three-dimensional vector containing a timestamp, event type, and value interval, and construct a spatiotemporal matrix based on the time axis, event type axis, and value axis for storage; S2. Calculate the transfer entropy and temporal closeness of adjacent events based on the spatiotemporal matrix, fuse the two through asymmetric cross entropy to generate a correlation strength value, and select continuous event sequences whose strength values exceed the confidence interval threshold as key sequences; S3. Map the key sequence into a weighted directed graph, with event types as nodes and association strength values as edge weights. Apply the maximum flow-minimum cut theorem to extract the core migration path, convert the path into state migration rules, and bind them to the dynamic rule attributes of the BIM entity. S4. Continuously monitor the real-time event stream and match the current event sequence with the core migration path in the weighted directed graph. If the accumulated path weight exceeds the judgment standard value and deviates from the historical path topology, update the BIM model status attributes, trigger an alarm, and adjust the confidence interval threshold based on the deviation.

2. A method for matching construction equipment and building information models according to claim 1, characterized in that: Step S1 includes the following contents: Real-time status data of building equipment is collected through the device control bus, and each data point is encoded as a three-dimensional vector containing a timestamp, event type, and numerical interval. A space-time matrix is constructed according to the time axis, event type axis, and numerical axis for storage. Data collection includes operating parameters, fault alarms, maintenance records, and discrete states. The timestamp records the moment of data collection, the event type is represented by a predefined code, and the numerical interval discretizes the continuous parameters according to the equipment specifications. The space-time matrix uses sparse storage technology to record the number of event occurrences and processes new data through an incremental update strategy.

3. A method for matching construction equipment and building information models according to claim 2, characterized in that: Step S2 includes the following contents: The transfer entropy and temporal closeness of adjacent events are calculated based on the space-time matrix. The transfer entropy and temporal closeness are fused through asymmetric cross entropy to generate the correlation strength value. The adjacent event pairs whose correlation strength values exceed the confidence interval threshold are screened, and the screened adjacent event pairs are connected to form a continuous event sequence. The corresponding continuous event sequence is defined as a key sequence.

4. A method for matching construction equipment and building information models according to claim 3, characterized in that: Step S2 also includes the following: By counting the frequencies of adjacent event pairs, a joint probability distribution is constructed. The joint probability of the preceding and subsequent events occurring simultaneously, the conditional probability of the subsequent event under given conditions of the preceding event, and the marginal probability of the subsequent event are calculated. These probability values are used to evaluate the predictive ability of the preceding event on the occurrence of the subsequent event. The specific method is to take the logarithm of the ratio of the joint probability to the conditional probability and the marginal probability, and then perform a weighted summation on all event pairs to finally obtain the transfer entropy value.

5. The method for matching construction equipment and building information model according to claim 3, characterized in that: Step S2 also includes the following: Determine the time periods in which the predecessor and successor events occur, then calculate the difference between the two time periods to obtain the time interval between them. Then, take the inverse of the corresponding time interval to generate a closeness index that increases as the time interval decreases, and finally obtain the time series closeness.

6. A method for matching construction equipment and building information models according to claim 3, characterized in that: Step S3 includes the following contents: Map the selected key event sequences into a weighted directed graph, with event types as nodes and the correlation strength values of adjacent event pairs as edge weights; The maximum flow-minimum cut theorem is applied to extract the core migration path in the weighted directed graph. The source point is set as the initial event type and the sink point is set as the target event type. The main information flow from the source point to the sink point is calculated and the minimum cut set is identified. The connection path is traced from the minimum cut set to extract the core migration path.

7. A method for matching construction equipment and building information models according to claim 6, characterized in that: Step S3 also includes the following: The core migration path is converted into state transition rules, the event type sequence in the path is decomposed into adjacent event pairs to generate migration rules, and the migration rules and association strength values are bound to the dynamic rule attributes of the building information model entity.

8. A method for matching construction equipment and building information models according to claim 7, characterized in that: Step S4 includes the following contents: The real-time status data of building equipment is continuously monitored through the device control bus, and the monitored data is encoded into data units containing timestamps, event types and numerical intervals. The event stream is constructed in the order of timestamps. A time window is set from the event stream to extract the current event sequence, and the corresponding path is matched in the weighted directed graph. The sum of the association strength values between adjacent event types in the path is calculated as the weight accumulation value.

9. A method for matching construction equipment and building information models according to claim 8, characterized in that: Step S3 also Includes the following: A judgment standard value is set based on the historical weight cumulative distribution of the core migration path. When the weight cumulative value does not exceed the judgment standard value, it is determined that the current event sequence deviates from the core migration path, the building information model status attribute is updated as abnormal, and an alarm is triggered.

10. A method for matching construction equipment and building information models according to claim 9, characterized in that: Step S3 also includes the following: The confidence interval threshold of the correlation strength calculation is adjusted according to the degree of deviation to ensure that the model adapts to changes in the operating status of building equipment.

Images