An old community public facility operation and maintenance work order intelligent scheduling method

Abstract

The application belongs to the technical field of operation and maintenance work order scheduling, and particularly relates to an old community public facility operation and maintenance work order intelligent scheduling method, which comprises the following steps: receiving repair work order data and preprocessing the data into standardized work order records; after grading, storing ordinary and planned grade work orders into a dynamic aggregation buffer queue; constructing a facility degradation correlation graph, combining a topological path distance, a spatial distance, an arrival time difference and a fault type consistency to calculate a correlation score and extract a candidate work order cluster; searching for a minimum common upstream influence node, combining a spatial distance to divide the candidate work order cluster into a systematic homologous fault cluster or a regional concentrated operation cluster; dynamically freezing the candidate work order cluster, and generating a work order package according to the cluster type to execute differentiated dispatching. The application can accurately identify homologous faults and preferentially repair root causes, reduce the fault recurrence rate and improve operation and maintenance efficiency.

Description

Technical Field

[0001] This invention relates to the field of operation and maintenance work order scheduling technology. More specifically, this invention relates to an intelligent scheduling method for operation and maintenance work orders of public facilities in old residential communities. Background Technology

[0002] With the continuous advancement of urbanization, a large number of old residential communities are gradually entering a period of high incidence of aging public facilities. The public power supply lines and water supply and drainage networks in these communities are often intricate and the physical connections between upstream and downstream are extremely close. The aging of facilities often exhibits obvious characteristics of propagation along the pipelines, which makes daily facility operation and maintenance face severe challenges.

[0003] Currently, for the scheduling of maintenance work orders for public facilities in old residential areas, existing technologies mainly rely on manual scheduling experience or simple clustering algorithms based on geographical location. These methods usually package and aggregate multiple repair work orders that are spatially close and assign them to the same maintenance personnel. The core purpose is to improve the efficiency of a single shift by reducing the commuting distance of maintenance personnel.

[0004] However, this scheduling method, which relies solely on spatial distance, ignores the propagation characteristics of aging facilities along physical pipelines in older residential areas. This leads to the incorrect merging of spatially adjacent but topologically unrelated work orders, increasing unnecessary troubleshooting time; while work orders with similar upstream root causes but located further apart are often overlooked and separated. Maintenance personnel can only passively address isolated surface-level faults, unable to accurately locate and repair upstream systemic root causes. Ultimately, this results in repeated fault recurrences on the same pipeline loop, significantly increasing the property management's daily maintenance costs, leading to a continuous decline in resident satisfaction, low overall operational efficiency, and difficulty in guaranteeing service timeliness. Summary of the Invention

[0005] To address the technical problem that existing scheduling methods ignore the propagation characteristics of facility aging along pipelines, leading to inaccurate repair of upstream root causes, recurring faults, and low maintenance efficiency, this invention provides an intelligent scheduling method for maintenance work orders of public facilities in old residential communities, comprising:

[0006] Receive repair work order data, preprocess the repair work order data, and obtain standardized work order records containing the spatial location of the faulty facility and the type of fault.

[0007] Standardized work order records are classified according to fault type, and the classified ordinary level work orders and planned level work orders are stored in a dynamic aggregation buffer queue.

[0008] Based on the physical topology connections of community public facilities, a facility degradation association graph is constructed. For any two work orders in the dynamic aggregation buffer queue, the association score between the two work orders is determined according to the topological path distance between the facilities corresponding to the two work orders in the facility degradation association graph, the spatial distance between the facilities corresponding to the two work orders, the arrival time difference between the two work orders, and the consistency of the fault type of the two work orders. Candidate work order clusters are extracted based on the association scores of each work order pair in the dynamic aggregation buffer queue.

[0009] Search for the minimum common upstream impact node of each work order corresponding to the facility within the candidate work order cluster in the facility degradation association graph. Based on the minimum common upstream impact node and the spatial distance between each work order corresponding to the facility within the candidate work order cluster, divide the candidate work order cluster into a systemic common source fault cluster or a regional concentrated operation cluster.

[0010] Candidate work order clusters are frozen based on preset dynamic freezing conditions, and corresponding work order packages are generated according to the types of systemic common-source fault clusters and regional centralized operation clusters to execute differentiated work order dispatch.

[0011] Preferably, the step of classifying standardized work order records according to fault type and storing the classified ordinary-level work orders and planned-level work orders into a dynamic aggregation buffer queue includes:

[0012] Based on the fault type, the scope of impact, and the safety risks to residents, the standardized work order records are divided into emergency level work orders, ordinary level work orders, and planned level work orders.

[0013] Store the ordinary level work orders and the planned level work orders into a dynamic aggregation buffer queue;

[0014] Calculate the latest response time limit for each work order stored in the dynamic aggregation buffer queue; calculate the remaining waiting time for each work order in the dynamic aggregation buffer queue based on the latest response time limit and the current system time; when the remaining waiting time is less than or equal to the preset critical time threshold, forcibly remove the corresponding work order from the dynamic aggregation buffer queue and execute the dispatch.

[0015] Preferably, the construction of the facility degradation association graph based on the physical topology connection relationship of community public facilities includes:

[0016] The public facilities of old residential areas are used as nodes to form a node set, and the physical connections between facilities are used as directed edges to form a directed edge set. The basic topology of the facility degradation association graph is constructed, and the directed edges point from upstream nodes to downstream nodes.

[0017] The aging transmission intensity of each directed edge is calculated based on the service life, historical maintenance count, and number of downstream branches directly connected to the upstream node of the upstream and downstream nodes.

[0018] Based on the number of times downstream nodes experience the same type of failure within a preset time period after an upstream node fails in historical records, calculate the common cause failure probability of each directed edge.

[0019] The aging conduction intensity, common cause failure probability, and physical subordination distance between adjacent nodes along the actual physical pipeline or route are used as directed edge attributes of the facility degradation correlation graph.

[0020] Preferably, the association score satisfies the expression:

[0021] ;

[0022] In the formula, Indicates work order and The correlation score between them; Indicates work order Corresponding facilities and work orders The topological hop count of the directed shortest path in the facility degradation association graph for the corresponding facility; Indicates work order Corresponding facilities and work orders The corresponding facility's Euclidean distance within the community's spatial grid; Indicates work order and The absolute value of the difference in arrival times; This indicates a function to indicate the consistency of fault types, when a work order... and When the fault type codes belong to the same major category, The value is 1, otherwise The value is 0.5; Represents the topological attenuation coefficient. Indicates the spatial attenuation coefficient. This represents the time decay coefficient.

[0023] Preferably, the step of extracting candidate work order clusters based on the association scores of each work order pair in the dynamic aggregation buffer queue includes:

[0024] Work order pairs in the dynamic aggregation buffer queue whose association score is greater than the preset association score threshold are identified as having candidate aggregation relationships.

[0025] The disjoint-set data structure algorithm is used to connect work orders with candidate aggregation relationships and extract an initial candidate work order cluster.

[0026] Verify whether the fault types of each work order in the initial candidate work order cluster belong to the same fault category; if the initial candidate work order cluster contains work orders of different fault categories and there is no matching full-category repair personnel, then split the initial candidate work order cluster into independent work orders.

[0027] The initial candidate work order cluster that has not been split is identified as a candidate work order cluster.

[0028] Preferably, the step of dividing the candidate work order cluster into a systemic common-source fault cluster or a regional concentrated operation cluster based on the minimum common upstream impact node and the spatial distance between the facilities corresponding to each work order within the candidate work order cluster includes:

[0029] The minimum aging propagation strength of a work order is the minimum value of the aging propagation strength of all directed edges on the directed shortest path from the facility corresponding to the work order within the candidate work order cluster to the least common upstream affected node.

[0030] In response to the existence of the least common upstream impact node and the fact that the fault type of each work order in the candidate work order cluster matches the aging propagation characteristics of the least common upstream impact node, and the average value of the minimum aging propagation intensity of all work orders in the candidate work order cluster is greater than or equal to the preset propagation intensity threshold, the candidate work order cluster is marked as a systemic homogeneous fault cluster.

[0031] In response to the fact that the candidate work order cluster is not marked as a systemic common source fault cluster and the Euclidean distance between each pair of facilities corresponding to each work order in the preset spatial grid is less than the spatial distance threshold, the candidate work order cluster is marked as a regional concentrated operation cluster.

[0032] In response to work orders that are not marked as systemic common-source fault clusters and where the Euclidean distance between the corresponding facilities within a candidate work order cluster is greater than or equal to the spatial distance threshold, work orders whose Euclidean distance exceeds the spatial distance threshold will be stripped into independent work orders, and the remaining work orders will be retained as regional centralized operation clusters.

[0033] Preferably, freezing the candidate work order cluster based on preset dynamic freezing conditions includes:

[0034] When any one of the conditions 1, 2 or 3 is met, the candidate work order cluster is dynamically frozen and the work order is output.

[0035] The first condition is: the number of work orders contained in the candidate work order cluster is greater than or equal to the preset work order number threshold, and the arrival rate of new work orders with an association score greater than the preset association score threshold within the preset time window is less than the preset rate threshold.

[0036] The second condition is: the minimum common upstream impact node remains stable within a preset number of consecutive scans, and the arrival rate of new work orders is less than the rate threshold.

[0037] The third condition is: the expected marginal aggregation revenue is less than the marginal aggregation revenue threshold, and the number of work orders contained in the candidate work order cluster is greater than or equal to the minimum number of work orders threshold.

[0038] Preferably, the marginal aggregation benefit satisfies the expression:

[0039] ;

[0040] In the formula, This indicates the expected marginal aggregation return; This indicates the number of work orders currently contained in the candidate work order cluster; This represents the average common cause failure probability among work order pairs within a candidate work order cluster. The method for obtaining the common cause failure probability among work order pairs within a candidate work order cluster is as follows: any two work orders in the candidate work order cluster are combined into a work order pair, the directed shortest path of the facility corresponding to the work order pair in the facility degradation association graph is obtained, and the minimum common cause failure probability of all directed edges on the directed shortest path is taken as the common cause failure probability of the work order pair. This represents the base value of the time saved per round trip due to the merger.

[0041] Preferably, the step of generating corresponding work order packages based on the types of systemic fault clusters and regional centralized operation clusters to execute differentiated work order dispatch includes:

[0042] For a cluster of systemic common faults, a root cause repair work order package is generated. The repair path of the root cause repair work order package is set to prioritize the repair of the least common upstream affected node and verify the downstream fault point in the order from upstream to downstream according to the topological path.

[0043] For regional centralized operation clusters, batch maintenance work order packages are generated. Starting from the current location of the maintenance personnel, the locations of the facilities corresponding to each work order within the regional centralized operation cluster are used as the waypoints. The shortest path is solved by the nearest neighbor insertion method and then dispatched.

[0044] Preferably, after executing differentiated order dispatch, the process further includes:

[0045] For the facilities corresponding to the work orders that have been repaired, monitor whether the facilities will report the same fault again within the set fault recurrence monitoring time window;

[0046] In response to a recurrence of a similar fault report, the recurrence event is included in the historical maintenance count and total fault count of the corresponding facility node, the historical statistics in the facility degradation correlation graph are updated, and the aging conduction intensity and common cause failure probability are recalculated.

[0047] The beneficial effects of this invention are as follows: By constructing a facility degradation correlation graph and combining it with multi-dimensional features to calculate correlation scores, candidate work order clusters are extracted. This allows work orders that are spatially distant but physically topologically similar to be accurately aggregated, avoiding the omission of similar faults. Furthermore, by searching for the minimum common upstream impact node, candidate work order clusters are divided into systemic similar fault clusters or regional concentrated operation clusters, accurately identifying systemic faults sharing the same upstream root cause. This achieves a shift from single spatial clustering to integrated fault propagation mechanism discrimination. Subsequently, candidate work order clusters are frozen based on dynamic freezing conditions, effectively balancing aggregation waiting and response time, avoiding premature freezing that misses subsequent related faults or late freezing that delays dispatch. Finally, differentiated dispatch is performed according to cluster type. For systemic similar fault clusters, upstream root causes are prioritized for repair and downstream issues are reviewed, eliminating the source of the fault and resolving the problem of repeated fault recurrence on the same pipeline loop. For regional concentrated operation clusters, optimal travel paths are planned, significantly reducing the ineffective commuting time of maintenance personnel. This comprehensively improves the overall operation and maintenance efficiency of public facilities in old residential areas and enhances resident satisfaction. Attached Figure Description

[0048] Figure 1 This is a flowchart illustrating an intelligent scheduling method for operation and maintenance work orders of public facilities in old residential communities according to the present invention. Detailed Implementation

[0049] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0050] The operation and maintenance scheduling system for old residential communities is deployed on the local server of the community property service center. It interacts with each repair reporting terminal through a distributed soft bus to realize the intelligent scheduling method for operation and maintenance work orders of public facilities in old residential communities provided in this embodiment of the invention.

[0051] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0052] This invention discloses an intelligent scheduling method for operation and maintenance work orders of public facilities in old residential communities, referring to... Figure 1 This includes steps S1-S5:

[0053] S1: Receive repair work order data and perform spatial mapping preprocessing to obtain standardized work order records containing the spatial location of the faulty facility and the fault type.

[0054] The system receives work order data submitted from three sources: repair cards submitted by residents via mobile phones, data entered by property management front desk staff, and automatic alarms triggered by IoT sensing devices. Each work order data includes a unique identifier for the faulty facility, a fault type code, a fault description text, a repair timestamp, and the source channel for the repair.

[0055] For each work order data, perform the following preprocessing operations:

[0056] Faulty facility location and spatial mapping: Based on the unique identifier of the faulty facility carried in the work order, the complete installation location information of the facility is retrieved from the facility database. The facility database pre-stores archive records of all public facilities within the community. Each record includes the facility's unique identifier, facility type code, brand and model, installation date, building code, unit code, floor code, precise coordinates in the community's spatial grid, historical repair count, and the number of repeated repair requests in the past 12 months. If the resident or property management staff does not specify a specific facility identifier when reporting a repair or entering data, fuzzy matching is performed using the repair location description text to retrieve the closest archive record from the facility database to determine the unique identifier of the faulty facility corresponding to the work order. Specifically, this is done by extracting keywords related to building, unit, and floor from the description text for hierarchical matching and calculating the text edit distance to select the archive record with the highest similarity. For example, if the description of the repair location is "Building 5, Unit 2, 3rd floor corridor light", the system will parse the building as Building 5, the unit as Unit 2, and the floor as 3rd floor. Then, it will search for archive records of facilities with the facility type of corridor light in that location and fill the unique facility identifier in the archive record back into the work order.

[0057] The method for establishing the spatial grid is as follows: taking the southwest corner of the community planning map as the origin of the coordinate system, and taking the due east direction as the coordinate system... The positive direction of the axis and the true north direction are The entire community is divided into 2-meter sections along the positive axis. A uniform grid of 2 meters is used. Each facility is mapped to a corresponding grid cell based on its actual installation location's GPS coordinates or as-built drawing coordinates, and its grid coordinates are recorded. Considering that the spacing between units in residential buildings is typically between 6 and 8 meters, The precision of meters is sufficient to distinguish the location differences between different units in the same building. At this precision, system memory usage and retrieval time are within a reasonable range, and the computational overhead is moderate. Therefore, in this embodiment, the community is divided into... Rice A uniform grid of meters, i.e., setting the grid resolution to 10 ... In other embodiments, implementers can adaptively adjust the grid resolution based on the actual area of ​​the community, building density, and server computing performance.

[0058] Furthermore, the preprocessed work order information is assembled into standardized work order records. Each work order record includes a work order number, a unique facility identifier, a facility type code, a fault type code, a building code, a unit code, a floor code, spatial grid coordinates, a repair timestamp, and a fault description text. The work order records are then stored in the system's scheduling queue, awaiting entry into step S2 for hierarchical and distributed processing.

[0059] At the same time, archival data of all public facilities in old residential areas, as-built drawings of power supply lines, as-built drawings of water supply networks, and building floor plans are collected to provide basic data for the subsequent construction of facility degradation correlation maps.

[0060] For older residential areas lacking blueprints, engineers manually mark the physical relationships between facilities through on-site surveys. These on-site markings are categorized according to the professional systems to which the facilities belong, and specifically include:

[0061] (1) For power supply facilities, mark the topological relationship of their upstream power supply nodes. For example, the upstream node of the corridor light is the floor distribution box, and the upstream node of the floor distribution box is the unit main distribution box.

[0062] (2) For water supply and drainage facilities, mark the topological relationship of their upstream water supply nodes. For example, the upstream node of a faucet is a branch valve, the upstream node of a branch valve is a main water supply pipe, and the upstream node of a main water supply pipe is an inlet pipe.

[0063] (3) For all the facilities with physical pipeline or line connections mentioned above, mark the pipeline or line directions between adjacent nodes, which will be used to calculate the physical subordination distance between adjacent nodes when constructing the facility degradation association diagram later.

[0064] After the on-site survey and labeling are completed, the data is entered into the system in the form of a directed edge list for use in the subsequent construction of the facility degradation correlation map.

[0065] S2: Classify work orders according to fault type, store ordinary level work orders and planned level work orders into a dynamic aggregation buffer queue, continuously monitor the remaining waiting time of work orders in the queue, and forcibly retrieve and execute the dispatch when the remaining waiting time is less than or equal to the critical time threshold.

[0066] It should be noted that a repair work order indicates a facility malfunction, but the impact and urgency of different malfunctions vary. For example, elevator outages and burst water pipes require immediate response and cannot be delayed by waiting for aggregation. While malfunctions such as single-floor hallway lights not working or slow access control in apartment buildings also affect residents' lives, there is a certain tolerable response window. Therefore, this invention first categorizes work orders into emergency, normal, and planned levels based on the type of malfunction, the scope of impact, and the safety risks to residents. Emergency work orders are dispatched immediately, while normal and planned work orders enter a dynamic aggregation buffer queue so that multiple work orders can be merged into a single work order package for subsequent aggregation and dispatch. Simultaneously, each work order entering the dynamic aggregation buffer queue is assigned a latest response deadline. This latest response deadline acts as a mandatory constraint to ensure that aggregation waiting does not drag on indefinitely. Once the latest response deadline is approached, a work order is forcibly dispatched. Without exceeding service time limits, non-urgent and mergeable work orders are given a short-term dynamic buffer.

[0067] Specifically, work orders are extracted from the system's queue of pending scheduling. For each work order, the emergency level of the work order is retrieved from the predefined emergency rule base according to the work order's fault type code and fault impact range, thereby performing a graded and diversion judgment on the work order.

[0068] The judgment logic of the emergency rule base is as follows:

[0069] (1) If the fault type of the work order belongs to any of the following categories: elevator shutdown, water supply main pipe burst, power supply main line outage, fire protection facility failure, gas leak alarm, or large-scale failure of night lighting, it shall be directly marked as an emergency level. Among them, elevator shutdown affects personal safety, water supply main pipe burst affects water supply in a large area, power supply main line outage affects power supply to multiple buildings, fire protection facility failure affects fire safety, gas leak alarm affects life and property safety, and large-scale failure of night lighting affects the safety of night passage in multiple buildings. The criterion for judging large-scale failure of night lighting is that the same fault causes the lighting facilities of three or more buildings to fail at the same time.

[0070] (2) If the fault type of the work order does not belong to the emergency category such as elevator shutdown, water supply main pipe burst, power supply main line power outage, fire protection facility failure, gas leak alarm, or large-scale failure of night lighting, but the scope of the fault affects more than this building, such as the same power supply circuit causing two adjacent buildings to lose power at the same time, or the main road water supply network of the community breaking, causing multiple units to stop water supply, it is also marked as an emergency level.

[0071] (3) If the fault type of the work order belongs to the category that affects the basic life of residents but is controllable, such as the single-floor corridor light not working, the unit access control malfunctioning, or the single household water pressure insufficient, it is marked as ordinary level;

[0072] (4) If the fault type of the work order belongs to the non-emergency facility optimization, such as insufficient lighting in public areas but not affecting traffic safety, or damage to green facilities, it is marked as a planned level.

[0073] All work orders marked as urgent are not entered into the dynamic aggregation buffer queue. The system immediately triggers an independent dispatch process, selecting the most skilled and currently least overloaded maintenance personnel from among the on-duty maintenance staff to directly dispatch the work order, ensuring immediate response to emergency faults. All work orders marked as normal or planned are entered into the dynamic aggregation buffer queue.

[0074] For work orders marked as normal or planned, when they enter the dynamic aggregation buffer queue, the system calculates the latest response time for each work order, satisfying the expression:

[0075] ;

[0076] In the formula, Indicates the latest response time for a work order, in absolute timestamps in seconds; This indicates the time when the work order arrived in the system, in absolute timestamps in seconds. This indicates the baseline response time for the facility type to which the work order belongs, in minutes. Used to convert minutes to seconds. This latest response time limit serves as a hard constraint for the subsequent dynamic aggregation window, setting an absolute safety boundary for each work order and ensuring that the waiting time for any work order must not exceed its corresponding baseline response time limit. This ensures that subsequent aggregation decisions do not compromise service timeliness.

[0077] In this embodiment, for work orders marked as "normal," the baseline response time for corridor lighting is... Pick Minutes, the baseline response time for access control systems Pick Minutes, the benchmark response time for water supply and drainage Pick Minutes, the baseline response time for public facilities Pick Minutes; for work orders marked as planned, the baseline response time for all facility types. All Minutes. It should be noted that the above are the baseline response times for standard-level work orders. The settings are based on industry standards and the community's service commitment agreement. Specifically, water supply and drainage faults have the shortest baseline response time due to the risk of water leakage and neighborhood disputes; corridor lighting faults have a moderate baseline response time due to nighttime safety concerns; access control faults have a slightly longer response time due to convenience of entry and exit without directly endangering safety; and public facilities faults have the least direct impact on residents' lives and therefore the longest response time. In other embodiments, implementers can adaptively adjust the above baseline response time values ​​based on the actual community size, historical work order data statistics, or maintenance team configuration.

[0078] After entering the dynamic aggregation buffer queue, the queue is sorted according to the arrival time of the work orders, and the remaining waiting time of each work order in the queue is continuously monitored. The remaining waiting time satisfies the expression:

[0079] ;

[0080] In the formula, This indicates the remaining waiting time for the work order, in minutes. Indicates the latest response time for a work order, in absolute timestamps in seconds; This represents the current system time, in absolute timestamp format (seconds). Used to convert seconds to minutes.

[0081] When the remaining waiting time of a work order is less than or equal to a preset critical time threshold, the system marks the work order as critical, forcibly removes it from the dynamic aggregation buffer queue, and initiates the dispatch process to ensure that the work order is processed within the response time limit. The critical time threshold is the necessary time margin reserved for the system to execute the dispatch process and for maintenance personnel to accept the order. In this embodiment, the critical time threshold is 15 minutes for ordinary level work orders and [missing information] for planned level work orders. In other embodiments, the implementer can adaptively adjust the critical time threshold based on the actual size of the community, the statistical results of historical work order data, or the configuration of the maintenance team.

[0082] S3: Construct a facility degradation association graph based on the physical topology connection relationship of community public facilities. For any two work orders in the dynamic aggregation buffer queue, calculate the association score by combining topological path distance, spatial distance, arrival time difference and fault type consistency, and extract candidate work order clusters accordingly.

[0083] It should be noted that the aging of facilities in old residential areas is not uniformly distributed, but rather exhibits a clear propagation characteristic along power supply circuits and water supply pipelines. The aging of upstream nodes accelerates the failure of downstream nodes, and multiple downstream facilities sharing the same upstream node often tend to fail synchronously. Therefore, this invention incorporates the unique aging propagation attributes of old residential areas to construct a facility degradation correlation graph. Each directed edge in the facility degradation correlation graph points from an upstream node to a downstream node, and the edge attributes are jointly determined by physical distance, pipeline age, and degree of sharing. The probability of common-cause failure is calculated driven by historical repeated repair reports and maintenance recurrence rate data. The facility degradation correlation graph enables the system to infer from multiple repaired facilities whether they share the same upstream root cause, providing physical mechanism support for common-source fault identification. Simultaneously, each work order in the dynamic aggregation buffer queue corresponds to a different node in the facility degradation correlation graph. If clustering is simply performed based on spatial distance, work orders that are spatially distant but topologically belong to the same upstream node will be missed. Therefore, this invention further filters out potentially related work order pairs through multi-dimensional comprehensive measurement to form a candidate work order cluster, providing basic data for subsequent same-source fault identification and dynamic window control.

[0084] Specifically, based on the as-built drawings of the power supply lines, water supply networks, and building plans of the old residential area, the basic topology of the facility degradation correlation diagram is constructed:

[0085] ;

[0086] In the formula, Represents the correlation diagram of facility degradation; This represents a set of nodes, where each node corresponds to a public facility. Node attributes include the facility's unique identifier, facility type, installation date, design life, historical maintenance history, and the number of repeated maintenance requests in the past 12 months. This represents a set of directed edges, where each directed edge points from an upstream node to a downstream node, reflecting the propagation direction of aging effects. Power supply-related edges represent physical connections between power supply facilities, with the direction from the distribution box to the floor distribution box and then to the corridor lights or switches. Water supply-related edges represent physical connections between water supply and drainage facilities, with the direction from the water inlet pipe to the main water supply pipe, then to the branch valve, and finally to the faucet or fire hydrant. When two facilities have both power supply lines and water supply pipelines physically connected, corresponding power supply-related edges and water supply-related edges are established separately. Each directed edge carries three key attributes: aging conduction strength, common cause failure probability, and physical dependency distance. It should be noted that in addition to power supply and water supply / drainage facilities, public facilities in older residential areas also include access control systems, elevators, and fire protection facilities. Since the degradation associations of these facilities are relatively independent and usually do not have multi-level upstream and downstream conduction relationships, they are supplemented in the facility degradation association graph as isolated nodes or a single-layer star topology. This embodiment uses power supply and water supply and drainage facilities as core examples for detailed explanation, as they cover the vast majority of recurring repair scenarios in old residential areas and have a clear upstream and downstream transmission relationship in physical topology.

[0087] The aging conductivity satisfies the expression:

[0088] ;

[0089] In the formula, Indicates from the upstream node to downstream nodes The aging conductivity, with a value range of [value missing]. In the facility degradation correlation graph, the upstream node With the downstream node There are directed edges directly connecting them; Indicates upstream node The number of years of service, in years; Indicates downstream node The service life of the upstream node, in years. With downstream nodes When all are newly built facilities, that is To avoid calculation errors caused by a denominator of 0, the following steps are taken: The value is directly assigned to 0; Indicates upstream node The number of repairs in the past 12 months; Indicates downstream node The number of maintenance visits in the past 12 months, when the upstream node and downstream nodes If no repairs have been made in the past 12 months, that is To avoid calculation errors caused by a denominator of 0, the following steps are taken: The value is directly assigned to 0; Indicates upstream node The total number of directly connected downstream branches, i.e., upstream nodes The number of directly connected downstream facility branches, when an extreme anomaly occurs... To avoid calculation errors caused by a denominator of 0, the following steps are taken: The value of is directly assigned to 0. This is the service weighting coefficient; This is the maintenance weighting coefficient; This represents the branch weight coefficient.

[0090] In the formula, This reflects the relative proportion of downstream nodes in the total service life of upstream and downstream nodes. When this ratio... A smaller ratio indicates that the downstream nodes have a relatively short service life and a strong ability to resist aging, while the aging effect of upstream nodes is less transmitted to downstream nodes; when this ratio is lower... When the value increases, it indicates that the downstream node has been in service for a long time, its natural aging degree is high, and its ability to resist external aging effects is reduced. Therefore, the aging state of the upstream node is more likely to induce or accelerate the failure of the downstream node. This reflects the relative maintenance intensity, when the ratio An increase in the ratio indicates that the maintenance frequency of downstream nodes accounts for a relatively high proportion of the total maintenance frequency across the upstream and downstream nodes. Since the high-frequency maintenance of downstream nodes is often caused by collateral damage resulting from continuously poor operating conditions of their source, the upstream node, the ratio is therefore significant. The increase in indicates that the upstream node degradation has a more significant driving effect on the downstream node failure. This reflects the degree to which the aging transmission of upstream nodes is distributed to downstream branches. The more downstream branches an upstream node connects to, the larger the physical area affected by an upstream node failure, and the more downstream branches its aging impact needs to be distributed. Therefore, the aging transmission weight distributed to a single downstream branch is lower. In this embodiment, considering that in the scenario of old residential areas, the service life of facilities is the most critical factor determining aging transmission, followed by maintenance frequency, while the impact of upstream shared infrastructure is relatively small, the following is considered. , as well as The values ​​were set to 0.4, 0.3, and 0.3 respectively. In other embodiments, implementers can adjust the weighting coefficients based on the actual size of the community, historical work order data statistics, or maintenance team configuration. , as well as Adjust settings adaptively.

[0091] The common cause failure probability satisfies the expression:

[0092] ;

[0093] In the formula, Indicates upstream node With downstream nodes The probability of common cause failure between them, with a value range of (0,1); Indicates upstream node Total number of failures; Represents the upstream node in the historical record Downstream nodes after failure During the preset time period The number of similar faults occurring within a given timeframe is considered. Furthermore, repair requests for multiple fault points on the same circuit typically occur within a short period of 3 days, with the correlation significantly weakening after 72 hours. Therefore, this embodiment will... Set as In other embodiments, the implementer can set the preset time period based on the aging characteristics of community facilities, the distribution of historical fault reporting time intervals, or the physical length of the pipeline network. Adjust settings adaptively; Represents the Laplace smoothing coefficient. is the smoothing coefficient of the denominator, and In this embodiment The value is 1. The value is 2, which is the Laplace smoothing coefficient. Smoothing coefficient of denominator The introduction of this method ensures that the probability estimate is neither 0 nor 1 when historical data is sparse, preserving uncertainty and avoiding overconfidence estimation based on sparse data. This invention calculates the common-cause failure probability by statistically analyzing the frequency of downstream node failures following upstream node failures in historical data. This probability measures the empirical probability of fault propagation between upstream and downstream nodes, reflecting the degree of correlation between upstream node failures and downstream node failures.

[0094] The physical subordinate distance is the actual path length between upstream and downstream nodes along the actual physical pipeline or route, in meters. It is obtained by on-site measurement of the actual pipeline laying length between adjacent nodes, or by measurement and calculation based on the pipeline or route between adjacent nodes and the pipeline scale in the community as-built drawings.

[0095] Because the node attributes in the facility degradation correlation graph change dynamically over time, this invention establishes a dynamic update mechanism to maintain the accuracy of the graph structure. Specifically, it automatically reads the latest service life and maintenance frequency of all facilities in the community each month, substitutes the latest data into the calculation formula for aging conduction intensity, and recalculates and updates the aging conduction intensity of each directed edge in the facility degradation correlation graph. Furthermore, if the community completes pipeline or power line renovations, it supports manual input of new topology structures, and the system synchronously adds or deletes graph nodes and directed edges in the facility degradation correlation graph to ensure that the facility degradation correlation graph remains consistent with the actual physical topology of the community.

[0096] After completing the construction of the facility degradation correlation diagram and setting the dynamic update mechanism, the work orders in the dynamic aggregation buffer queue will be listed as a set. A candidate work order cluster generation task is triggered once with a preset aggregation scan cycle. Considering that 10 minutes is sufficient to accumulate new work orders, and that work order response will not be delayed due to excessively long scan intervals, this embodiment sets the aggregation scan cycle to 10 minutes. In other embodiments, implementers can adaptively adjust the aggregation scan cycle based on the actual cell size, historical work order data statistics, or maintenance team configuration.

[0097] For any two work orders in the dynamic aggregation buffer queue and Calculate the correlation score based on the facility degradation correlation diagram:

[0098] ;

[0099] In the formula, Indicates work order and The correlation score between them, with a range of values ​​being: ; Indicates work order Corresponding facilities and work orders Corresponding facilities in the facility degradation correlation diagram The topological hop count of the directed shortest path in the work order, i.e., the number of directed edges traversed, is calculated using Dijkstra's algorithm in this embodiment. Corresponding facility node to work order Obtain the directed shortest path hop count of the corresponding facility node If the work order The corresponding facility node cannot reach the work order. The corresponding facility nodes, Take infinity; Indicates work order Corresponding facilities and work orders The Euclidean distance of the corresponding facilities in the community spatial grid satisfies The unit is meters, among which, For work orders The coordinates of the corresponding facilities in the community's spatial grid. For work orders The coordinates of the corresponding facilities in the community's spatial grid; Indicates work order and The absolute value of the difference in arrival times satisfies The unit is minutes, of which Indicates work order The time of arrival at the system Indicates work order The time of arrival at the system; This indicates a function to indicate the consistency of fault types, when a work order... and When the fault type codes belong to the same major category, The value is 1, for example, a work order. and When the fault type codes are both lighting or both water supply and drainage ,otherwise The value is set to 0.5, which is to preserve the slight possibility of aggregation of different major categories of faults under extreme spatial proximity conditions; This represents the topology attenuation coefficient, expressed in units of per hop. This represents the spatial attenuation coefficient, expressed in units of per meter. This represents the time decay coefficient, measured in minutes. It is used to eliminate differences in the numerical units and magnitudes of topological distance, spatial distance, and time difference, enabling normalized decay calculations for multi-dimensional features. In this embodiment, it will... , as well as The values ​​are set to 0.05 per jump, 0.01 per meter, and 0.005 per minute, respectively. In other embodiments, the implementers can adaptively adjust the values ​​of the above attenuation coefficients according to the actual size of the community, the statistical results of historical work order data, or the configuration of the maintenance team.

[0100] In the formula, This represents the topological distance factor, reflecting the decaying effect of the topological path on the association score. The shorter the directed shortest path distance between two facilities in the facility degradation association graph, the closer they are on the aging propagation path, and the higher the probability of common source failure. Each hop means that for every 20 hops increase in the topology hop count, the topology distance factor decreases by approximately 63%, which characterizes that the shorter the topology distance, the higher the probability of a common source fault. This represents the spatial distance factor, reflecting the attenuating effect of physical spatial distance on the association score. The closer the physical distance between two facilities, the lower the passage cost for aggregation processing. Each meter means that for every 100-meter increase in spatial distance, the spatial distance factor decreases by approximately 63%, which characterizes that the greater the spatial distance, the higher the cost of aggregation processing. This is a device type factor used to ensure that faults of different categories are not easily aggregated even if they are spatially adjacent, because their maintenance skills and material requirements are different. This represents the time decay factor, reflecting the attenuation effect of the repair time difference on the correlation score. Work orders with closer repair times are more likely to be the result of the same systemic fault being exposed successively. Each minute means that for every 200-minute increase in time difference, the time decay factor decreases by approximately 63%. This value ensures that when the time difference reaches 200 minutes, the time decay factor decreases to approximately 63%, thus limiting the effective aggregation time range to the typical time window of successive exposure of faults from the same source. This invention comprehensively measures the correlation strength between two work orders through the product of four factors. Mismatch in any dimension will significantly lower the correlation score. Only work orders that simultaneously meet the four conditions of topological proximity, spatial proximity, consistent fault type, and temporal proximity can obtain a high correlation score.

[0101] In response to Determine the work order and There are candidate aggregation relationships, where, The threshold for association score is set to 0.4. In this embodiment, the threshold is set to 0.4 to balance the sufficiency of work order aggregation with the timeliness of response. This threshold is the system default value and can be manually adjusted by property management personnel according to the size of the community and the configuration of the maintenance team. It can also be iteratively optimized based on historical work order big data.

[0102] Furthermore, a disjoint-set data structure algorithm is used to extract connected subgraphs, and work orders with direct or indirect candidate aggregation relationships are connected to form an initial candidate work order cluster. For the generated initial candidate work order clusters, verify whether the fault category within the cluster is unique. If it contains multiple fault categories such as electrical and water supply / drainage, then search for on-duty personnel with full-category maintenance qualifications. If no matching all-around personnel are found, automatically split the initial candidate work order cluster into independent work orders, prohibiting the merging and dispatch of work orders across major categories. The initial candidate work order clusters that are not split are then determined as candidate work order clusters.

[0103] S4: Search for the minimum common upstream impact node of each facility corresponding to each work order in the facility degradation association graph. Based on the minimum common upstream impact node and the spatial distance between facilities, divide the candidate work order cluster into a systemic common source fault cluster or a regional concentrated operation cluster. Freeze the candidate work order cluster based on the preset dynamic freezing conditions.

[0104] It should be noted that after candidate work order clusters are generated, systemic faults are often not reported all at once, but rather gradually emerge over a window of several hours. Freezing too early will miss subsequent related faults, while freezing too late will delay the response time of already reported work orders. Therefore, this invention first performs common-source fault discrimination on candidate work order clusters, tracing back the minimum common upstream impact node of the facilities corresponding to these work orders in the facility degradation association graph, and classifying candidate clusters into systemic common-source fault clusters or regional concentrated operation clusters based on the matching degree between the fault type and the aging propagation characteristics of the upstream nodes. Through the above processing, the traditional single spatial distance clustering is transformed into discriminative clustering that incorporates the constraint of fault propagation mechanism. Subsequently, a dynamic aggregation window control mechanism is further introduced, which comprehensively considers factors such as the arrival rate of new work orders, the stability of root cause nodes, the service time limit of the earliest work order, and the aggregation benefits and waiting costs to calculate the optimal freezing time of the current candidate work order cluster in real time. When the rate of new work orders arriving slows down, the root cause node remains unchanged after multiple consecutive updates, the remaining time limit of the earliest work order is insufficient to support continued waiting, or the marginal aggregation benefit is lower than the waiting cost, the system immediately freezes the candidate work order cluster and enters the work order dispatch process.

[0105] Specifically, for each candidate work order cluster Further perform source-based fault discrimination: extract candidate work order clusters. All work orders and their corresponding facilities are shown in the facility degradation correlation diagram. The set of nodes in A breadth-first search-based shortest path algorithm is used in the facility degradation correlation graph. Search node set The least common upstream affected node The least common upstream influential node For: In the facility degradation correlation diagram In this context, a set of nodes can be reached directly or indirectly along one or more directed edges. The node with the deepest topological level among the common ancestor nodes of all nodes.

[0106] The candidate work order cluster is marked as a systemic common-source fault cluster when the following three conditions are met:

[0107] (1) Minimum common upstream influence node exist.

[0108] (2) Fault types and minimum common upstream impact nodes of each work order within the candidate work order cluster The aging transmission characteristics are matched. Specifically, based on the least common upstream influence node. The professional system type determines the corresponding fault type matching rule: if the least common upstream affected node For power supply nodes such as distribution boxes and floor switches, at least 80% of the fault types in the candidate work order clusters must be electrical faults; if the minimum common upstream impact node... For water supply nodes such as main water supply pipes and branch valves, at least 80% of the fault types in the candidate work order clusters should be water supply faults. For facility nodes such as access control systems, elevators, and fire protection facilities, which are modeled as isolated nodes or single-layer star topologies in the facility degradation association diagram, since their degradation associations are relatively independent and usually do not have multi-level upstream and downstream transmission relationships, the minimum common upstream influence node should be considered. When the fault belongs to this type of independent facility node, the systemic common-source fault determination is not performed, and the subsequent Euclidean distance determination logic is directly entered. In this embodiment, the fault type matching ratio is set to 80%, which ensures the strictness of the determination while allowing a small number of exceptions caused by secondary faults. The fault type matching ratio is the system default calibration value, which can be manually customized by property management personnel according to the size of the community and the configuration of the maintenance team, and can also be iteratively and adaptively optimized based on historical work order big data.

[0109] (3) The aging conduction intensity of the path from the facility corresponding to each work order in the candidate work order cluster to the suspected root cause node meets the preset condition. Specifically, the directed shortest path from the facility corresponding to each work order in the candidate work order cluster to the suspected root cause node in the facility degradation association graph is calculated, and the minimum value of the aging conduction intensity of all directed edges on the directed shortest path is obtained as the minimum aging conduction intensity corresponding to the work order. If the average value of the minimum aging conduction intensity corresponding to all work orders in the candidate work order cluster is greater than or equal to the preset conduction intensity threshold, then the condition is determined to be met. Considering that in old residential areas, the service life of upstream main facilities is usually longer than that of downstream branch facilities, and the number of downstream branches connected to upstream nodes is large, the actual calculated value of the aging conduction intensity with significant aging conduction driving effect is usually concentrated between 0.25 and 0.45; while when the upstream facilities are relatively new or there is no downstream damage, the calculated value is usually lower than 0.2. Therefore, in this embodiment, the conduction strength threshold is set to 0.2 as the critical point to distinguish between weak conduction and significant conduction, so as to ensure that a systemic common source fault is only determined when the aging of the upstream node has a significant conduction driving effect on the downstream node.

[0110] If all three conditions above are met, the candidate work unit cluster will be marked as a systemic common-origin fault cluster, and the least common upstream affected node will be... Mark the suspected root cause node as a candidate work order cluster; otherwise, obtain the pairwise Euclidean distance between the facilities corresponding to each work order in the spatial grid within the candidate work order cluster. If the pairwise Euclidean distance between the facilities corresponding to each work order in the spatial grid within the candidate work order cluster is less than the preset spatial distance threshold, then mark the candidate work order cluster as a regional centralized operation cluster. If there is an Euclidean distance between the facilities corresponding to work orders within the candidate work order cluster that is greater than or equal to the preset spatial distance threshold, then perform a secondary connected subgraph split on the candidate work order cluster based on the spatial distance threshold, and separate the work orders that cause the Euclidean distance to exceed the preset spatial distance threshold into independent work orders, and retain the subset of work orders in the cluster whose pairwise Euclidean distance is less than the preset spatial distance threshold as regional centralized operation clusters. Considering that the distance between buildings in old residential areas is usually between 20 and 40 meters, a 30-meter threshold can cover the spatial concentration of work orders between adjacent buildings or different units in the same building, while avoiding the erroneous merging of work orders that are far apart. Therefore, this embodiment sets the spatial distance threshold to 30 meters. In other embodiments, implementers can adaptively adjust the spatial distance threshold according to the actual size of the residential area, the distribution of building spacing, or the statistical results of historical work order data.

[0111] After the candidate work order clusters are marked, dynamic aggregation window control is further executed. For each candidate work order cluster... In each aggregation scan cycle, the following dynamic decisions are performed: monitoring the arrival rate of new work orders, determining the stability of root cause nodes, and evaluating marginal aggregation benefits.

[0112] New work order arrival rate monitoring includes: calculating the arrival rate of candidate work orders in the dynamic aggregation buffer queue. The arrival rate of new work orders with candidate aggregation relationships within a preset time window satisfies the expression:

[0113] ;

[0114] In the formula, This indicates the rate at which new work orders arrive, measured in orders per minute. This indicates a preset time window; in this embodiment, the time window is set to 30 minutes. In other embodiments, implementers can adjust the preset time window based on the frequency characteristics of community repair requests or the historical work order time distribution. Adjust settings adaptively. This indicates newly arrived work orders within a preset time window that are related to the candidate work order cluster. The number of work orders whose association score is greater than a preset association score threshold. The association score between a newly arrived work order and a candidate work order cluster is the maximum value of the association scores between the newly arrived work order and all work orders within that candidate work order cluster. This is calculated over two consecutive aggregation scan cycles. When all rates are less than the preset rate threshold, it indicates that the arrival of related work orders has slowed down. In this embodiment, the preset rate threshold is set to 0.033 orders / minute, meaning less than one new order is added within 30 minutes. In other embodiments, implementers can adaptively adjust the preset rate threshold according to the size of the community or the load of the maintenance team.

[0115] The stability of the root cause node is determined by recording the candidate work order cluster. Suspected root cause nodes in the aggregated scan cycle within the most recent preset number of scans The update status. In this embodiment, the preset number of scans is set to 3, i.e., within the last 30 minutes. If the preset number of scans is exceeded consecutively... If no changes are observed, the root cause node is considered to be stable. In other embodiments, the implementer can adaptively adjust the preset number of scans based on the convergence speed of the root cause node in the historical work order data.

[0116] Marginal aggregation benefit assessment is based on calculating the impact of waiting for the next related work order to enter the candidate work order cluster. Potential marginal aggregation benefits Satisfies the expression:

[0117] ;

[0118] In the formula, This represents the expected marginal aggregate return, expressed in minutes. Represents candidate work order clusters The number of work orders currently included. This represents the average common cause failure probability among work order pairs within a candidate work order cluster. The method for obtaining the common cause failure probability among work order pairs within a candidate work order cluster is as follows: any two work orders in the candidate work order cluster are combined into a work order pair, the shortest directed path of the facility corresponding to the work order pair in the facility degradation association graph is obtained, and the minimum common cause failure probability of all edges on the shortest directed path is taken as the common cause failure probability of the work order pair. This represents the base value of the single round-trip travel time saved due to the merger, with a preset value of 30 minutes. This value can be set according to actual conditions or estimated based on the average round-trip commuting time of the community. In other embodiments, implementers can adjust the value based on the actual community size, historical work order data statistics, or maintenance team configuration. Adaptive adjustment settings are implemented. When candidate work order clusters... Number of work orders currently included Greater than or equal to the preset work order quantity threshold and the expected marginal aggregation revenue When the time threshold is less than the preset marginal aggregation revenue threshold, it indicates that the aggregation revenue has decreased significantly, and the marginal value of continuing to wait has decreased. In this embodiment, the preset work order quantity threshold is set to 5, and the preset marginal aggregation revenue threshold is set to 10 minutes. In other embodiments, the implementer can adaptively adjust the above-mentioned work order quantity threshold and marginal aggregation revenue threshold according to the distribution density of community facilities or the commuting cost of maintenance personnel.

[0119] Before making a freeze decision, the system first iterates through the remaining waiting time of all work orders within the candidate work order cluster. Work orders that have reached the critical time threshold are separated and directly generated into independent work order packages for dispatch. The remaining work orders that have not reached the critical time threshold are retained as new sub-clusters and continue to participate in subsequent cycle aggregation scans. Then, the system makes a freeze decision. When any of the following conditions are met, the system immediately freezes the candidate work order cluster. And output the dispatch order:

[0120] (1) The number of work orders in the candidate work order cluster is greater than or equal to the preset work order number threshold, and the arrival rate of new work orders is less than the preset rate threshold.

[0121] (2) Suspected root cause nodes The number of consecutive preset scans remains stable, and the arrival rate of new work orders is less than the preset rate threshold.

[0122] (3) The marginal aggregation revenue is less than the preset marginal aggregation revenue threshold, and the number of work orders in the candidate work order cluster is greater than or equal to the preset minimum work order number threshold. In this embodiment, the preset minimum work order number threshold is set to 3. In other embodiments, the implementer can adaptively adjust the preset minimum work order number threshold according to the cell size.

[0123] Once frozen, the candidate work order cluster enters the differentiated dispatch process in step S5. After freezing, the candidate work order cluster will no longer accept new work order merging, but isolated work orders in the dynamic aggregation buffer queue can still participate in the aggregation determination of subsequent scan cycles.

[0124] S5: Generate corresponding work order packages based on the types of systemic fault clusters and regional centralized operation clusters to execute differentiated work order dispatch.

[0125] Differentiated work order dispatching is implemented based on the type of regional centralized work clusters and systemic common fault clusters to achieve intelligent scheduling of maintenance work orders for public facilities in old residential areas:

[0126] For a cluster of systemic faults originating from the same source, the system generates a root cause repair work order package, the contents of which include suspected root cause nodes. The system provides the location and troubleshooting guidelines, a list of affected downstream facilities and their verification sequence, and a description of the expected repair scope. The root cause repair work order package's repair path is set to prioritize the repair of upstream root cause nodes and then verify downstream fault points. The system instructs repair personnel to proceed to the suspected root cause node based on the root cause repair work order package. The system investigates and repairs the affected downstream facilities, and after root cause repair, instructs maintenance personnel to verify, one by one, whether the original fault points in the affected downstream facility list have been restored to normal, following the topological path from upstream to downstream. When planning the route of this topological path, the system uses the physical subordination distance between adjacent nodes in the facility degradation correlation graph as the travel cost, and employs the nearest neighbor insertion method to solve for the optimal investigation and verification path, closely aligning with the physical routing of actual pipeline inspections. If the downstream fault point is still not restored after the maintenance personnel complete the root cause repair, the system reprocesses that downstream fault point as an independent work order.

[0127] For regional centralized operation clusters, the system generates batch maintenance work order packages. These packages include the optimal travel route and maintenance sequence, the fault type for each fault point, and a list of required tools and materials. The batch maintenance work order package starts from the current location of the maintenance personnel, uses the locations of each fault point within the regional centralized operation cluster as waypoints, calculates the spatial travel cost using the Euclidean distance of the corresponding facility in the community's spatial grid, solves for the shortest travel path using the nearest neighbor insertion method, and considers the skill requirements of each fault point for personnel matching.

[0128] After the generated work order package is pushed to the maintenance personnel's terminal, the system synchronously updates the status of each original work order to "merged and pending." After maintenance is completed, the system enters the maintenance effectiveness monitoring phase. For specific fault points within a regional centralized work cluster, the system monitors the recurrence monitoring time window for that facility after maintenance. Whether the same fault report recurs within the specified time window This is used to set the observation period for the effect after repair. In this embodiment, the fault recurrence monitoring time window is used. The time window for monitoring fault recurrence is set to 30 days. In other embodiments, implementers may adjust the monitoring time window based on facility type, maintenance process requirements, or historical recurrence cycles. The system adaptively adjusts the values. If a recurrence occurs, the system determines that the fault was not eradicated during the maintenance. The recurrence event is automatically included in the historical maintenance count and total fault count of the node where the facility is located to update the historical statistics of the corresponding node in the facility degradation correlation graph, and the aging propagation strength and common cause failure probability of the directed edges involving the node are recalculated. If no recurrence occurs, the current probability estimate is maintained. At the same time, the system records the actual effect of each aggregation decision, including the actual completion time after aggregation, the number of work orders resolved in a single attendance, and the recurrence status. It periodically outputs aggregation performance reports to provide data support for the adaptive adjustment of the decision threshold.

[0129] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

[0130] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A method for intelligent scheduling of maintenance work orders for public facilities in old residential communities, characterized in that, include: Receive repair work order data, preprocess the repair work order data, and obtain standardized work order records containing the spatial location of the faulty facility and the type of fault. Standardized work order records are classified according to fault type, and the classified ordinary level work orders and planned level work orders are stored in a dynamic aggregation buffer queue. The process involves constructing a facility degradation correlation graph based on the physical topology connections of community public facilities. This includes: using the public facilities of the old community as nodes to form a node set, and using the physical connections between facilities as directed edges to form a directed edge set, thus constructing the basic topology of the facility degradation correlation graph. The directed edges point from upstream nodes to downstream nodes. The aging conduction intensity of each directed edge is calculated based on the service life of the upstream and downstream nodes, the number of historical maintenance visits, and the number of downstream branches directly connected to the upstream node. The common-cause failure probability of each directed edge is calculated based on the number of times the downstream node experiences a similar failure within a preset time period after an upstream node failure, according to historical records. The aging conduction intensity, common-cause failure probability, and the physical subordination distance between adjacent nodes along actual physical pipelines or lines are used as the directed edge attributes of the facility degradation correlation graph. For any two work orders in the dynamic aggregation buffer queue, the association score between the two work orders is determined based on the topological path distance between the facilities corresponding to the two work orders in the facility degradation association graph, the spatial distance between the facilities corresponding to the two work orders, the arrival time difference between the two work orders, and the consistency of the fault type of the two work orders; and candidate work order clusters are extracted based on the association scores of each work order pair in the dynamic aggregation buffer queue. Search for the minimum common upstream impact node of each work order corresponding to the facility within the candidate work order cluster in the facility degradation association graph. Based on the minimum common upstream impact node and the spatial distance between each work order corresponding to the facility within the candidate work order cluster, divide the candidate work order cluster into a systemic common source fault cluster or a regional concentrated operation cluster. Candidate work order clusters are frozen based on preset dynamic freezing conditions, and corresponding work order packages are generated according to the types of systemic common-source fault clusters and regional centralized operation clusters to execute differentiated work order dispatch.

2. The intelligent scheduling method for operation and maintenance work orders of public facilities in old residential areas according to claim 1, characterized in that, The step of classifying standardized work order records according to fault type and storing the classified ordinary-level work orders and planned-level work orders into a dynamic aggregation buffer queue includes: Based on the fault type, the scope of impact, and the safety risks to residents, the standardized work order records are divided into emergency level work orders, ordinary level work orders, and planned level work orders. Store the ordinary level work orders and the planned level work orders into a dynamic aggregation buffer queue; Calculate the latest response time limit for each work order stored in the dynamic aggregation buffer queue; calculate the remaining waiting time for each work order in the dynamic aggregation buffer queue based on the latest response time limit and the current system time; when the remaining waiting time is less than or equal to the preset critical time threshold, forcibly remove the corresponding work order from the dynamic aggregation buffer queue and execute the dispatch.

3. The intelligent scheduling method for operation and maintenance work orders of public facilities in old residential areas according to claim 1, characterized in that, The association score satisfies the expression: ； In the formula, Indicates work order and The correlation score between them; Indicates work order Corresponding facilities and work orders The topological hop count of the directed shortest path in the facility degradation association graph for the corresponding facility; Indicates work order Corresponding facilities and work orders The corresponding facility's Euclidean distance within the community's spatial grid; Indicates work order and The absolute value of the difference in arrival times; This indicates a function to indicate the consistency of fault types, when a work order... and When the fault type codes belong to the same major category, The value is 1, otherwise The value is 0.5; Represents the topological attenuation coefficient. Indicates the spatial attenuation coefficient. This represents the time decay coefficient.

4. The intelligent scheduling method for operation and maintenance work orders of public facilities in old residential areas according to claim 3, characterized in that, The step of extracting candidate work order clusters based on the association scores of each work order pair in the dynamic aggregation buffer queue includes: Work order pairs in the dynamic aggregation buffer queue whose association score is greater than the preset association score threshold are identified as having candidate aggregation relationships. The disjoint-set data structure algorithm is used to connect work orders with candidate aggregation relationships and extract an initial candidate work order cluster. Verify whether the fault types of each work order in the initial candidate work order cluster belong to the same fault category; if the initial candidate work order cluster contains work orders of different fault categories and there is no matching full-category repair personnel, then split the initial candidate work order cluster into independent work orders. The initial candidate work order cluster that has not been split is identified as a candidate work order cluster.

5. The intelligent scheduling method for operation and maintenance work orders of public facilities in old residential communities according to claim 1, characterized in that, The process of classifying candidate work order clusters into systemic common-source fault clusters or regional concentrated operation clusters based on the minimum common upstream impact node and the spatial distance between the facilities corresponding to each work order within the candidate work order cluster includes: The minimum aging propagation strength of a work order is the minimum value of the aging propagation strength of all directed edges on the directed shortest path from the facility corresponding to the work order within the candidate work order cluster to the least common upstream affected node. In response to the existence of the least common upstream impact node and the fact that the fault type of each work order in the candidate work order cluster matches the aging propagation characteristics of the least common upstream impact node, and the average value of the minimum aging propagation intensity of all work orders in the candidate work order cluster is greater than or equal to the preset propagation intensity threshold, the candidate work order cluster is marked as a systemic homogeneous fault cluster. In response to the fact that the candidate work order cluster is not marked as a systemic common source fault cluster and the Euclidean distance between each pair of facilities corresponding to each work order in the preset spatial grid is less than the spatial distance threshold, the candidate work order cluster is marked as a regional concentrated operation cluster. In response to work orders that are not marked as systemic common-source fault clusters and where the Euclidean distance between the corresponding facilities within a candidate work order cluster is greater than or equal to the spatial distance threshold, work orders whose Euclidean distance exceeds the spatial distance threshold will be stripped into independent work orders, and the remaining work orders will be retained as regional centralized operation clusters.

6. The intelligent scheduling method for operation and maintenance work orders of public facilities in old residential areas according to claim 4, characterized in that, The freezing of candidate work order clusters based on preset dynamic freezing conditions includes: When any one of the conditions 1, 2 or 3 is met, the candidate work order cluster is dynamically frozen and the work order is output. The first condition is: the number of work orders contained in the candidate work order cluster is greater than or equal to the preset work order number threshold, and the arrival rate of new work orders with an association score greater than the preset association score threshold within the preset time window is less than the preset rate threshold. The second condition is: the minimum common upstream impact node remains stable within a preset number of consecutive scans, and the arrival rate of new work orders is less than the rate threshold. The third condition is: the expected marginal aggregation revenue is less than the marginal aggregation revenue threshold, and the number of work orders contained in the candidate work order cluster is greater than or equal to the minimum number of work orders threshold.

7. The intelligent scheduling method for operation and maintenance work orders of public facilities in old residential areas according to claim 6, characterized in that, The marginal aggregation benefit satisfies the expression: ； In the formula, This indicates the expected marginal aggregation return; This indicates the number of work orders currently contained in the candidate work order cluster; This represents the average common cause failure probability among work order pairs within a candidate work order cluster. The method for obtaining the common cause failure probability among work order pairs within a candidate work order cluster is as follows: any two work orders in the candidate work order cluster are combined into a work order pair. The directed shortest path of the facility corresponding to the work order pair in the facility degradation association graph is obtained. The minimum common cause failure probability of all directed edges on the directed shortest path is taken as the common cause failure probability of the work order pair. This represents the base value of the time saved per round trip due to the merger.

8. The intelligent scheduling method for operation and maintenance work orders of public facilities in old residential areas according to claim 1, characterized in that, The step of generating corresponding work order packages based on the types of systemic fault clusters and regional centralized operation clusters to execute differentiated work order dispatch includes: For a cluster of systemic common faults, a root cause repair work order package is generated. The repair path of the root cause repair work order package is set to prioritize the repair of the least common upstream affected node and verify the downstream fault point in the order from upstream to downstream according to the topological path. For regional centralized operation clusters, batch maintenance work order packages are generated. Starting from the current location of the maintenance personnel, the locations of the facilities corresponding to each work order within the regional centralized operation cluster are used as the waypoints. The shortest path is solved by the nearest neighbor insertion method and then dispatched.

9. The intelligent scheduling method for operation and maintenance work orders of public facilities in old residential communities according to claim 1, characterized in that, After implementing differentiated order dispatch, the following is also included: For the facilities corresponding to the work orders that have been repaired, monitor whether the facilities will report the same fault again within the set fault recurrence monitoring time window; In response to a recurrence of a similar fault report, the recurrence event is included in the historical maintenance count and total fault count of the corresponding facility node, the historical statistics in the facility degradation correlation graph are updated, and the aging conduction intensity and common cause failure probability are recalculated.

Images