Park management method based on digital twinning

Abstract

The invention discloses a park management method based on digital twinning, particularly relates to the field of park operation, and is used for solving the problem that a multi-source event is difficult to stably chain under the conditions of compensation delay and evidence gap and causes misalignment of disposal triggering. Comprising the following steps of: registering an event in a pool to generate an event number and writing the event number into a platform arrival moment; screening and mapping an anchor point event to a park object identifier and a process stage identifier to generate a management token; checking a stage sequence and an evidence item requirement set according to an event chain rule table in event chain construction and generating a conflict point list; the delay portrait library generates a credible mark when a source is generated and forms a reverse sequence violation spectrum; the digital twin object relation graph forms an influence domain span mark; the closed table is returned to obtain a disposal convergence guarantee level; the cost-sensitive decision tree outputs a disposal template identifier, and the disposal template table generates a management action instruction and a trigger voucher and records an instruction acceptance identifier; and returning, comparing and freezing the exception management token, and updating the event chain rule table and the delay portrait library.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of park operation, more particularly, the present application relates to a park management method based on digital twinning. BACKGROUND

[0002] In the park management method based on digital twinning, the park often hopes to make security, access, energy consumption, equipment operation and other work more timely, so the information generated by the access control, video, sensor, equipment control and work order platform is converged to the digital twinning platform. The platform maps the space structure, asset distribution and running state of the park to the virtual model, and tries to align the events reported by each system according to the occurrence place and occurrence time, so that the management personnel can view the situation and review the process in the same interface, and complete the linkage disposal and single closed loop under the driving of rules.

[0003] But the existing method usually only pursues the synchronization, and it is easy to ignore that the event time of each business system is not the same: some are based on collection time, some are based on gateway reporting, and some are based on platform storage. In the alignment process, the event sequence may be rewritten in the case of delay compensation and correction; the platform uses this modified timeline to judge the trigger condition, which may trigger the condition that should not be triggered, and the condition that should be triggered but is not triggered, and then the linkage control and single process are deviated, and the disposal records before and after are also difficult to be self-consistent. More troublesome is that such deviation is often hidden in the data fusion link, and it is difficult for the on-site personnel to directly see where the problem is, and the management party can only downgrade the automation to manual confirmation in order to avoid risks, which makes it difficult for digital twinning to truly play the role of the nerve center of park management.

[0004] In order to solve the above problems, a technical scheme is provided. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a park management method based on digital twinning, which comprises: event pool registration, generating an event number and writing a platform arrival time; anchor event screening, mapping park object identifier and process phase identifier to generate management token; event chain construction, checking phase sequence and evidence item requirement set according to event chain rule table and generating conflict point list; delay profile library, generating source time credible mark and forming reverse sequence violation spectrum; digital twinning object relationship diagram, forming influence domain span mark; receipt closure table, obtaining disposal convergence guarantee level; cost-sensitive decision tree, outputting disposal template identifier; disposal template table, generating management action instruction and trigger credential and recording instruction acceptance identifier; and abnormal management token freezing and updating event chain rule table and delay profile library in the receipt comparison, so as to solve the problems in the background art.

[0006] To achieve the above purpose, the present application provides the following technical scheme: S1: receive the event type and source time sent by the park business system, write into the event pool, generate event number, record event type and source time access channel identifier, complete event pool registration; S2: filter anchor events from the event pool, parse park object identifier and process phase identifier according to anchor events, generate management token, and write management token and token back to the event pool; S3: put the event carrying the management token into the event chain, check the event sequence and evidence integrity according to the event chain rule table, output the disposal template identifier, record the event chain state and conflict point list; S4: generate management action instruction and trigger credential according to disposal template identifier, write trigger credential into event pool, send management action instruction to execution system according to disposal template identifier, and record instruction acceptance identifier; S5: receive execution receipt to form a receipt event and join the event chain, compare the receipt event and the trigger credential, freeze the management token and generate a review work order in case of exception, and update the event chain rule table and the delayed portrait library according to the review conclusion.

[0007] Further, the access layer receives event type and source time, event type is verified by event type dictionary and written into event pool, source time is converted into standard timestamp and written into source time missing marker; the access layer extracts the access channel identifier and gets the channel sequence number, calculates the time slice sequence number according to the time slice length and updates the channel time slice counter to get the intra-slice sequence number, generates the event number by using the channel sequence number segment, time slice sequence number segment and intra-slice sequence number segment; the event pool writes event number, event type, source time, source time missing marker, access channel identifier, channel sequence number and platform arrival time.

[0008] Further, the anchor event table returns the anchor allowed event type set and the forbidden access channel identifier set with event type as the entry, the event pool reads event type and access channel identifier, and when the event type belongs to the anchor allowed event type set and the access channel identifier does not belong to the forbidden access channel identifier set, writes the event number into the anchor candidate queue, and the anchor candidate queue keeps the event number order consistent according to the platform arrival time.

[0009] Further, the object mapping table returns the park object identifier with object clue as the entry, the process mapping table returns the process phase identifier and phase order threshold with event type as the entry, the anchor candidate queue reads the event number corresponding to the object clue one by one and gets the park object identifier, generates the token primary key according to the process phase identifier and combines the event number and the platform arrival time to generate the management token version number to get the management token, and the token index table writes the valid management token with the park object identifier and suppresses the phase rollback according to the phase order threshold.

[0010] Further, the management token is retrieved as a merge key from the event pool to write event records into an event chain, the event chain is sorted by platform arrival time and event number, the event chain rule table maps the process phase identifier to a phase sequence number and gives a set of evidence item requirements, a conflict point list is generated according to the descending position of the phase sequence number sequence, an evidence gap marker is generated according to the set of evidence item requirements to check the event upload content, and the conflict point list records the event number pair and the process phase identifier pair.

[0011] Further, the delay image library gives delay segment boundaries according to event type and access channel identifier, source time and platform arrival time difference are used to generate source time credibility markers and construct reverse order violation spectrum, the evidence gap marker source time credibility marker is used to judge the time semantic consistency level according to the set of phase boundaries allowed to be rearranged by the key phase set, the digital twin object relationship graph is expanded in two rounds to generate an impact domain object set and judge the impact domain span marker, the disposition convergence guarantee level is obtained by checking the selected management action type set in the reply closure table and the event pool configuration library, and the time semantic consistency level and the disposition convergence guarantee level are used to output a disposition template identifier by the cost-sensitive decision tree and write it into the event chain state.

[0012] Further, the disposition template table writes the disposition template identifier, retrieves the management action type, the execution end identifier, the rule identifier, and writes them into the event chain state, the management action instruction payload is generated by concatenating the fixed-length encoded order of the park object identifier management token management action type disposition template identifier, the payload check code is obtained by performing a digest operation on the payload input string, and the payload check code is written into the management action instruction header and delivered with the payload to the delivery address corresponding to the execution end identifier.

[0013] Further, the event chain state generates an event number sequence encoding string in the order of the sorting order and concatenates it with the rule identifier, the disposition template identifier, and the management token version number to form a credential input string, the credential check code is obtained by performing a digest operation on the credential input string, the trigger credential record is written into the event pool and set to read-only attribute, the instruction request record is written into the event pool, the event pool locates the instruction request record according to the credential check code, and writes the instruction acceptance identifier after checking that the payload check code is consistent.

[0014] Further, the access layer receives the execution end identifier, the management action type, the management token, the instruction acceptance identifier, the payload check code, the credential check code, the reply event type, and the platform arrival time, and writes them into the event pool to generate a reply event, the event pool locates the instruction request record according to the management token, the instruction acceptance identifier, and the execution end identifier, generates a reply event de-duplication key according to the management token, the instruction acceptance identifier, the reply event type, the execution end identifier, the payload check code, and the credential check code, and writes it into the event pool, and the reply event de-duplication key is consistent when the reply event record is merged.

[0015] Further, the event pool colludes with the return event and the execution end identifier management action type, the processing template identifier, the load verification code, the voucher verification code of the instruction request record, and colludes with the voucher verification code, the processing template identifier, and the management token version number of the trigger voucher record to complete consistency verification, and writes a freeze flag and generates a review work order when verification fails or positioning fails, and updates the evidence item requirement set and the allowed rearranged stage boundary set of the event chain rule table and updates the delay profiling library delay segmentation boundary.

[0016] The technical effects and advantages of the park management method based on digital twinning of the present application are as follows: 1. Event entry registration adopts access channel identifier source time slice channel incremental sequence to generate event number and write platform arrival time, event pool retains event type and source time at the same time, event number keeps sequence consistent within the same access channel, keeps unique traceable across access channels, delay profiling library can depict channel reporting form with source time and platform arrival time, anchor point event table and object mapping table hit stably under unified field caliber, reducing ambiguity and drift in event collection and traceability process.

[0017] 2. Anchor point event screening and management token generation solidifies event type access channel identifier park object identifier process stage identifier into management token, and locks valid management token through token index table, event chain construction takes management token as merging key to avoid cross-object mixed chain, event chain rule table checks stage order and evidence item requirement set and generates conflict point list, delay profiling library generates source time credible marker and constructs reverse order violation spectrum, digital twinning object relationship diagram gives influence domain span marker, return closure table and event pool configuration library registration table get treatment convergence guarantee level, cost-sensitive decision tree outputs treatment template identifier, treatment selection takes into account timeline credibility and treatment controllable range, reducing the risk of false triggering and difficult convergence treatment.

[0018] 3. Management action instruction and trigger voucher generation parses the treatment template identifier into the management action type and the execution end identifier, the trigger voucher record event number sequence rule identifier treatment template identifier management token version number and sets the read-only attribute, the instruction request record binds the load verification code and the voucher verification code and writes the instruction acceptance identifier, the return comparison executes the instruction request record with the management token instruction acceptance identifier execution end identifier positioning and completes consistency verification, the exception writes a freeze flag and generates a review work order, and the review conclusion updates the event chain rule table and the delay profiling library, so that the treatment closed loop has a verifiable evidence chain and a sustainable self-correction ability. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 The flowchart of the park management method based on digital twinning of the present application. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0021] Embodiment 1 Figure 1 A park management method based on digital twinning is given, comprising: S1: receiving an event type and a source time sent by a park business system, writing into an event pool, generating an event number, recording the event type, the source time and an access channel identifier, and completing event pool registration.

[0022] S2: screening an anchor event from the event pool, analyzing a park object identifier and a process stage identifier according to the anchor event, generating a management token, and writing the management token and the event number back to the event pool.

[0023] S3: classifying events carrying the management token into an event chain, checking the sequence of the events and the completeness of the evidence according to an event chain rule table, outputting a disposal template identifier, and recording an event chain state and a conflict point list.

[0024] S4: generating a management action instruction and a trigger credential according to the disposal template identifier, writing the trigger credential into the event pool, sending the management action instruction to a designated execution system according to the disposal template identifier, and recording an instruction acceptance identifier.

[0025] S5: receiving an execution receipt to form a receipt event and join the event chain, comparing the receipt event and the trigger credential, freezing the management token and generating a review work order in an abnormal case, and updating an event chain rule table and a delayed image library according to a review conclusion.

[0026] Park digital twinning management aggregates scattered events into an event pool. The event pool undertakes two responsibilities, one of which is to ensure that each event has a unique identifier and stable classification, and the other of which is to retain the original appearance of the source time, facilitating subsequent judgment around time semantic consistency. The source time has missing, late reporting and delay, and the event number cannot be stably generated according to the channel sequence once. Anchor event identification and event chain merging may drift. However, too many event pool fields will increase access and maintenance costs, and step S1 needs to complete traceable and orderable pool registration with a small number of fields.

[0027] The scattered events here refer to management-related events generated by different business domains, different devices and different applications in the daily operation process of the park, for example, which can typically include: Access control and personnel access events: card swiping records, visitor appointment verification results, forced opening of access control alarm, offline alarm of access control.

[0028] Security and video linkage events: perimeter intrusion alarm, key area loitering alarm, fire and smoke identification alarm, video equipment offline alarm.

[0029] Fire and emergency events: smoke sensing triggered alarm, fire hydrant pressure abnormality alarm, fire control host state change, emergency broadcast linkage feedback.

[0030] Energy consumption and equipment maintenance events: power distribution circuit abnormality alarm, air conditioning unit failure alarm, elevator trapped person alarm, pump room water level abnormality alarm, work order creation and completion events.

[0031] Environment and space service events: temperature and humidity abnormality alarm, air quality overrun alarm, parking lot full alarm, charging pile failure alarm.

[0032] Business process events: patrol task start event, patrol task completion event, alarm confirmation event, dispatch acceptance event, disposal receipt event.

[0033] S101. Event type and source time receiving verification.

[0034] Park events come from access control alarm, energy consumption and maintenance, and other business domains. Inconsistent event type criteria will cause unstable anchor event table hits. Before entering the pool, the event type is compared with the event type dictionary to facilitate unified classification. The access layer receives the event type and source time. When the event type dictionary is hit, the event type field is written to the event pool. When the event type dictionary is not hit, the original value of the event type field is written, and the event type to be confirmed flag is written at the same time. The source time is converted to a standard timestamp format according to a unified time unit. The conversion process only aligns the format and does not compensate for time. When the source time is missing or the format is incorrect, the source time missing flag is written, and the source time field remains empty and is not replaced. The event pool record keeps the original value of the event type and the verification state. When the anchor event table is retrieved according to the event type field, it can stably distinguish between confirmed types and types to be confirmed.

[0035] S102. Access channel identification extraction and channel number determination.

[0036] The delay characteristics of the same event type are significantly different in different access links, and the delay profile library needs to depict the channel behavior in channel granularity. The access channel identifier can avoid mixing different channels when entering the pool. The access layer extracts the access channel identifier from the access link meta information, and the access channel identifier is matched with the channel dictionary. When the channel dictionary is hit, the channel sequence number is obtained. When the channel dictionary is not hit, the channel to be confirmed mark is written and the temporary channel sequence number is allocated. The temporary channel sequence number is mapped back after the channel dictionary is perfected. The event pool writes the access channel identifier and the channel sequence number. The access channel identifier is used to trace the link source, and the channel sequence number is used for event number generation and channel time slice segment counter positioning. After the channel granularity is clear, the delay profile library can form a stable group.

[0037] S103. Time slice segment sequence number calculation and channel time slice segment counter update.

[0038] The channel time slice segment counter is used to ensure that the same channel sequence number has strictly increasing and non-repeating segment sequence numbers within the same time slice segment, avoiding event number reuse caused by multi-thread concurrent writing. The time slice segment sequence number is obtained by performing integer division on the source time standard timestamp and the time slice length. When the source time missing mark exists, the platform arrival time is used instead of the source time to participate in the integer division. The source time missing mark is retained. The channel time slice segment counter uses the channel sequence number and the time slice segment sequence number as a joint index. The event pool side uses conditional write-back when performing read-modify-write process on the joint index record. The old value of the segment sequence number is read first, and then the new value of the segment sequence number is calculated as the old value plus one. The write-back request is submitted with the write-back condition that the joint index record still retains the old value of the segment sequence number. The event pool only updates the segment sequence number to the new value when the write-back condition is met and returns a success mark. When the write-back condition is not met, a failure mark is returned and the old value is required to be read again. The access layer repeatedly reads and performs conditional write-back until a success mark appears when a failure mark appears. The new value of the segment sequence number returned by the success mark directly participates in event number generation, and the segment sequence number in a concurrent scenario is constrained by the event pool conditional write-back to be a totally ordered increasing sequence.

[0039] S104. Event number generation and event pool write solidification.

[0040] The event number is used to uniquely identify events in the event pool and support sequential comparison. The event number is encoded in a fixed-length segment and introduces an event number version number to ensure version resolution. The event number is obtained by concatenating the event number version number, channel sequence number segment, time slice segment sequence number segment, and intra-segment sequence number segment in a fixed order. The channel sequence number segment converts the channel sequence number to a decimal character sequence and left-pads zeros to the version registration length. The time slice segment sequence number segment converts the time slice segment sequence number to a decimal character sequence and left-pads zeros to the version registration length. The intra-segment sequence number segment converts the intra-segment sequence number to a decimal character sequence and left-pads zeros to the version registration length. The event pool configuration library registration table registers the channel sequence number segment length, time slice segment sequence number segment length, intra-segment sequence number segment length, and total event number length for each event number version number. Event number resolution selects the corresponding segment length definition based on the event number version number to ensure that event numbers of different versions can be restored to the channel sequence number, time slice segment sequence number, and intra-segment sequence number triple. The extended version marker is written by the channel time slice counter when the intra-segment sequence number reaches the upper limit of the version registration length and triggers the event number version number switching. After switching, the event pool configuration library registration table adds segment length definitions and takes effect, and the old version event number remains unchanged. Event pool writing is uniquely constrained by event numbers to ensure non-repetition. When there is a unique constraint conflict, the intra-segment sequence number is reacquired and the event number is regenerated before writing again until success. Fixed-length zero padding ensures that the string order under the same event number version number is consistent with the numerical order of the channel sequence number segment, time slice segment sequence number segment, and intra-segment sequence number segment.

[0041] After step S1 is completed, the event pool forms a complete original event record, which includes the event number, event type, source time, source time missing marker, access channel identifier, channel sequence number, and platform arrival time. The event type field supports the anchor event table to filter anchor events by event type. The channel sequence number and platform arrival time support the delay profile library to depict delay behavior by channel. The event number supports the event chain to maintain the same order in the channel after merging according to the management token. The source time missing marker avoids misidentifying the arrival time as the source time and introducing time semantic bias.

[0042] The event entry pool registration stage writes the event number, event type, source time, access channel identifier, and platform arrival time into the event pool. The event pool already has the minimum elements required for anchor event filtering. The park management process needs to merge the same event chain from the starting event. Unstable starting event identification can mix different park objects into the same event chain, and too strict starting event identification can miss the real starting point. However, the starting event still needs to fall into the park object identifier and process stage identifier. Step S2 carries out deterministic processing around the anchor event table object mapping table process mapping table token index table.

[0043] S201. Anchor event filtering determination.

[0044] The anchor event table takes the event type as the entry, and the anchor event table also gives the allowed access channel identifier set and the forbidden access channel identifier set, the allowed set is used to admit real-time channels, and the forbidden set is used to isolate the historical supplement channel to avoid the historical supplement event being regarded as a new starting point. The event pool reads the event type corresponding to the event number and the access channel identifier, first searches the anchor event table whether the event type belongs to the anchor allowed event type, and then searches whether the access channel identifier falls into the forbidden access channel identifier set, and when the two searches are satisfied at the same time, the event number is written into the anchor candidate queue. The anchor candidate queue means that it is specially used to store the event number sequence that may be used as the starting point of the management process, and the writing action keeps the time sequence consistent when the event pool platform arrives, reduces the drift of the chain caused by the starting point out of order from the source, and the event number in the anchor candidate queue can be stably reproduced when the subsequent analysis of the event number is performed.

[0045] S202. Park object identification analysis and object mapping table mapping.

[0046] The park object identification needs to fall into a specific access point device space unit, and the event chain merging can be carried out around the same management object, and the lack of park object identification will mix the events of multiple objects together. The event pool reads the object clue corresponding to each event number in the anchor candidate queue, the object clue means the identification fragment that can uniquely point to the management object in the event upload content, and the source priority order is fixed as asset number point number space number, and the priority order is used to suppress ambiguity when the same event carries multiple clues.

[0047] The object mapping table saves the unique mapping of the object clue to the park object identification, writes the park object identification and the object clue hit marker when the query hits, and writes the object clue unmapped marker and transfers the event number into the to-be-verified queue when the query misses. The to-be-verified queue means a set of event numbers whose object clues need to be completed manually or by rules, and after the park object identification hits, the token index table aggregates the anchor event according to the park object identification, and the anchor event will not drift across objects.

[0048] S203. Flow stage identification determination and management token generation.

[0049] The flow stage identifier is used to express the position of the treatment flow, the absence of the flow stage identifier will classify events of different treatment stages into the same stage, and the treatment template identifier judgment will deviate from the real treatment semantics. The flow mapping table returns the flow stage identifier with the event type as the entry, and returns the stage order caliber corresponding to the flow stage identifier. The stage order caliber means the order of the flow stage identifier in the standard treatment flow. The management token generation adopts fixed sequence field splicing. First, the park object identifier is standardized coded. The standardized coding means that the park object identifier is mapped into a fixed-length object coding string. The mapping relationship is maintained and kept one-to-one by the object mapping table. Then, the flow stage identifier is standardized coded. The standardized coding means that the flow stage identifier is mapped into a fixed-length stage coding string. The mapping relationship is maintained and kept one-to-one by the flow mapping table. Then, the object coding string and the stage coding string are spliced in a predetermined order to form a token primary key. The management token version number adopts the combination coding of the event number of the anchor event and the platform arrival time. The combination coding means that the event number and the platform arrival time are respectively converted into fixed-length coding strings, and then spliced in a predetermined order to form a version coding string. The management token is obtained by splicing the token primary key and the version coding string. The event pool writes the park object identifier, the flow stage identifier, the token primary key, the management token, the management token version number, and the management token. After the token field construction rule is fixed, the cross-module parsing remains consistent.

[0050] S204. Token index table writing and valid management token determination.

[0051] The token index table needs to quickly give the valid management token corresponding to the park object identifier, and the event chain construction stage can classify events into the event chain with the management token as the key. The token index table saves the valid management token and the valid flow stage identifier with the park object identifier as the primary key, and saves the valid management token version number and the valid platform arrival time. The determination logic first reads the valid flow stage identifier and the valid platform arrival time corresponding to the park object identifier in the token index table, and then reads the flow stage identifier and the platform arrival time corresponding to the newly generated management token. The flow stage comparison adopts the stage order caliber provided by the flow mapping table. The flow stage later in the stage order caliber is not allowed to be covered by the flow stage earlier. The coverage marker is triggered when the coverage marker is true. The coverage marker means that the emergency restart scenario permission is marked by the anchor event table. The platform arrival time comparison is used to handle the update within the same flow stage. The platform arrival time update later writes the token index table and covers the original record. The synchronous record coverage source event number is covered. The stage rollback situation writes the new management token into the token history table and retains the original valid management token. After the token index table is stable, the same park object identifier always points to a valid management token, the stage rollback is suppressed, and the supplementary anchor will not damage the ongoing treatment link.

[0052] After step S2, the anchor event record in the event pool has the park object identifier, the process phase identifier, the management token, the token version number, and the token index table. The token index table gives the valid management token and keeps the phase order consistent. The anchor candidate queue solidifies the start point event screening criteria into a joint decision of event type and access channel identifier. The object mapping table maps the start point event to the park object identifier. The process mapping table maps the start point event to the process phase identifier. The management token combines the object and the phase into a stable chain key. The token index table suppresses phase rollback and isolates the anchor point. The event chain merging will not produce cross-object and cross-phase mixed chains due to time semantic drift.

[0053] The park object identifier, the process phase identifier, and the management token have been written into the event pool. The event chain construction phase has a clear merging key. The park management action trigger depends on the sequence of events and evidence integrity. The sequence is affected by the reporting and queuing. The evidence integrity is affected by the missing fields. Once the disposition template identifier is triggered on an untrusted timeline, the linkage and dispatching may deviate from the review. However, event chain construction cannot stop at simple sorting. It needs to include the time semantic credibility and disposition convergence ability in the judgment. Step S3 carries out detailed operations around the event chain rule table and the image library digital twin object relationship graph return closure table.

[0054] S301. Event chain merging and stable sorting.

[0055] The management token carries the park object identifier and the process phase identifier. Event chain merging needs to lock the same disposition semantic range with the management token to avoid cross-object and cross-phase mixed chains. The token index table reads the valid management token corresponding to the park object identifier. The event pool retrieves the event records with the same management token. The event records are written into the event chain. The event chain sorting arranges the platform arrival time from early to late. The event records with the same platform arrival time are arranged from small to large according to the event number. The event chain state writes the event number sequence, the event type sequence, and the process phase identifier sequence. After stable sorting, the conflict point positioning is carried out around the same sequence. The review time line is consistent, and the merging result is stable and reproducible.

[0056] S302. Phase order check and evidence gap marking.

[0057] The event chain rule table maps the process phase identifier to the phase sequence number and lists the corresponding evidence item requirement set of each process phase identifier, which is used to constrain the integrity of the trigger basis. The event chain generates a phase sequence number sequence according to the sorted process phase identifier sequence, and records a conflict point when the phase sequence number decreases in adjacent positions. The conflict point is written into the conflict point list, and the conflict point list records the event number pair and the process phase identifier pair involved in the conflict. The event pool reads the content uploaded on each event record, and checks the field existence and format legality item by item according to the evidence item requirement set. When there is a missing or illegal format, a evidence gap marker is written to indicate a gap state. When all are satisfied, a evidence gap marker is written to indicate a complete state. The conflict point list and the evidence gap marker form a same caliber checkable record, and the disposal template identifier judgment no longer depends on subjective inference.

[0058] S303. Source time credibility marker calculation and reverse sequence violation spectrum construction.

[0059] The source time credibility marker is used to explain the cause of the conflict between the event chain ordering and the source time sequence, and to avoid misjudging the supplement and queuing report as real reverse sequence. The event chain traverses each event record according to the stable sorting result, and calculates the difference value between the source time and the platform arrival time point item by item. The difference value is the platform arrival time point minus the source time. When the source time missing marker exists, the difference value state is recorded as unknown source time and does not participate in the delay segment boundary comparison. The delay profile library saves the delay segment boundary with event type and access channel identifier as joint key. The delay segment boundary divides the difference value into three types of intervals: real-time reporting segment, queuing reporting segment and supplement segment. The difference value and the boundary are compared segment by segment to obtain the delay segment marker. The monotonicity judgment is limited within the same access channel identifier. The adjacent source times are compared according to the event chain ordering sequence. When the source time backtracks, the delay segment marker is corrected to the supplement segment. When the source time is unknown, the source time unknown marker is not corrected. The source time credibility marker takes the value domain as real-time reporting segment, queuing reporting segment, supplement segment and unknown source time, and is directly determined by the delay segment marker and the source time unknown marker. The reverse sequence violation spectrum adopts an entry structure for storage. Each entry writes the previous event number, the subsequent event number, the previous process phase identifier, the subsequent process phase identifier, the previous source time credibility marker, the subsequent source time credibility marker, the previous evidence gap marker, the subsequent evidence gap marker, and the key stage set hit marker. The entry generation rule is to generate an entry when the phase sequence number backtracks and the backtrack boundary does not belong to the allowed rearrangement phase boundary set. The entry structure ensures that the reverse sequence cause can be calculated item by item according to the field.

[0060] S304. Time semantic consistency level adjudication rule.

[0061] The time semantic consistency level arbitration is used to converge the combination of reverse violation spectrum and evidence gap markers into a stable gear, avoiding different trigger interpretations of the same scene in different replays. The arbitration first scans the reverse violation spectrum entries. When the reverse violation spectrum entry hits the key stage set and the previous sequence evidence gap marker or the subsequent sequence evidence gap marker of the entry is in the gap state, the time semantic consistency level is directly written as low consistency. When low consistency is not triggered, there are entries in the reverse violation spectrum and the previous sequence source time credible marker and the subsequent sequence source time credible marker are both in the supplement segment, and the previous sequence evidence gap marker and the subsequent sequence evidence gap marker are both in the complete state, the time semantic consistency level is written as medium consistency. When the reverse violation spectrum is empty, or the reverse violation spectrum entries all fall into the stage boundary set allowed to rearrange, and the evidence gap markers on both sides of the entry are in the complete state and the source time credible marker does not exist source time unknown, the time semantic consistency level is written as high consistency. The arbitration adopts the most strict priority criterion, the low consistency criterion is prior to the medium consistency criterion, the medium consistency criterion is prior to the high consistency criterion, and the allowed rearrangement boundary does not cover the disallowed reverse entry.

[0062] S305. Treatment convergence assurance level calculation and treatment template identification output.

[0063] The treatment convergence assurance level is used to measure the controllability of risk diffusion after the management action is triggered, and the treatment convergence assurance level is jointly determined by the influence domain diffusion range and the receipt closing ability. The digital twin object relationship graph takes the park object identifier as the node and the object relationship edge as the edge. The object relationship edge at least includes two types of control dependency edges and shared resource edges. The control dependency edge represents the control propagation path, and the edge direction points from the control initiating object to the controlled object. The shared resource edge represents the resource coupling path, and the edge connects two shared resource objects. The digital twin object relationship graph is stored in the form of an adjacency list. The adjacency list returns the object relationship edge set with the edge type marker as the key. The adjacency list update criterion comes from the digital twin park object account and the resource binding account. The account change triggers the adjacency list to update synchronously and record the update timestamp, ensuring the reproducibility of the influence domain expansion. The influence domain expansion includes the park object identifier ontology. The first round of expansion obtains a set of directly controlled objects along the control dependency edge. The second round of expansion obtains a set of resource coupled objects along the shared resource edge starting from the set of directly controlled objects. The expansion process de-duplicates the park object identifier and stores it in a fixed traversal order. The influence domain object set is formed by merging the park object identifier ontology, the set of directly controlled objects, and the set of resource coupled objects.

[0064] The influence domain span mark is decided according to the partition identifier set covered by the influence domain object set, and the partition identifier set is obtained through the park object identifier to partition identifier mapping table, and the park object identifier to partition identifier mapping table is maintained by the digital twin park space model and is one-to-one correspondence. When the influence domain object set only contains the park object identifier ontology, write to the single-point device domain; when the influence domain object set covers a single partition identifier and the resource coupling object set is empty, write to the local space domain; when the influence domain object set covers multiple partition identifiers or the resource coupling object set is not empty, write to the cross-zone sharing domain, which is used to reflect the cross-domain linkage risk caused by shared resources. The standby management action type set is returned by the event chain rule table according to the process phase identifier and the event type, the reply closing table lists the execution reply event type set and the read reply event type set according to the management action type, and the event pool configuration library registration table records the reply event type and registers the event format definition.

[0065] The reply closing check is executed for each standby management action type set, and the execution reply event type set and the read reply event type set are read for a single management action type, and it is checked whether each reply event type in the set exists in the event pool configuration library registration table, and after the existence check, it is checked whether the event format definition contains the management token field and the park object identifier field, and the read reply event type set is checked whether the event format definition contains the load check code field or the voucher check code field. When the execution reply event type set and the read reply event type set both satisfy the existence check and the format check, the record is closed complete, only the execution reply event type set satisfies, the record is closed partial, and both sets do not satisfy, the record is closed missing. The treatment convergence guarantee level is obtained by mapping the influence domain span mark and the closing check result, the cross-zone sharing domain superimposes the closed partial or closed missing record low guarantee, the local space domain superimposes the closed complete record high guarantee, and the remaining combination record medium guarantee.

[0066] The cost-sensitive decision tree takes the time semantic consistency level and the treatment convergence guarantee level as input, and outputs the treatment template identifier. The decision tree training sample comes from the historical event chain, and the sample label comes from the management action final result type. The cost setting records the false trigger under the low guarantee condition as the high cost, and records the delay trigger under the high consistency and high guarantee condition as the second high cost. The reasoning stage obtains the treatment template identifier according to the tree path and writes it into the event chain state, and writes the time semantic consistency level and the treatment convergence guarantee level. The treatment template identifier output restricts risk diffusion and retains treatment efficiency, and the event chain state record has a checkable basis.

[0067] The cost-sensitive decision tree takes time semantic consistency level and treatment convergence guarantee level as input features, takes treatment template identification as output category, and the training sample source is limited to the event chain records that have completed the reply consistency verification and the review conclusion is passed in the historical event chain. The sample label adopts the treatment template identification of the management action final result type reverse deduction. The sample is stored in the database first. The time semantic consistency level and the treatment convergence guarantee level are only allowed to take the pre-defined level. The records exceeding the level are entered into the verification queue to avoid introducing uninterpretable value domain into the training set. During training, the cost matrix is used to express the misjudgment risk. The row of the cost matrix represents the true treatment template identification, the list represents the predicted treatment template identification, and the matrix element represents the business loss level caused by judging the true category as the predicted category. The matrix construction determines the relative priority according to the park treatment scene, defines the risk of mis-triggering under low guarantee conditions as the highest loss level, defines the risk of delayed triggering under high consistency and high guarantee conditions as the second highest loss level, and defines the rest confusion as the general loss level. The node splitting evaluation does not use pure classification purity index, but takes the expected loss reduction as the target. For each candidate split, the optimal prediction category of the left and right child nodes under the cost matrix constraint is calculated respectively. The optimal prediction category takes the category that minimizes the child node expected loss. The difference between the sum of the expected losses of the left and right child nodes and the expected loss of the parent node is taken as the splitting benefit. The splitting condition with the maximum splitting benefit is selected. The splitting condition comes from the level combination rule of the time semantic consistency level and the treatment convergence guarantee level, such as splitting according to the threshold of a single level or splitting according to the cross combination of two levels, to ensure that the tree path has clear business meaning.

[0068] The parameter setting is determined around interpretability and stability. The upper limit constraint of tree depth is used to avoid overfitting. The minimum sample size constraint of leaf node is used to avoid being dominated by a small number of abnormal links. The minimum splitting benefit threshold is used to avoid producing weak splitting. The pruning adopts the loss complexity criterion based on the validation set. The validation set is obtained by time division from the training samples to ensure that the validation set represents the latest park operation mode. The training process adopts the parameter combination traversal method to complete the parameter tuning. The evaluation index adopts the joint evaluation of the expected loss of the validation set and the occurrence frequency of the key misjudgment. The expected loss is obtained by accumulating the cost matrix sample by sample. The key misjudgment refers to the situation of being judged as a strong action template under low guarantee conditions and the situation of being judged as an observation template under high consistency and high guarantee conditions. Iterative optimization takes the review conclusion and the reply consistency verification result as the incremental sample source. The newly added passing samples are integrated into the training set at a fixed period. At the same time, the sliding training set is formed by retaining the samples within the recent window to avoid the long-term domination of the model by old scenes. The monitoring index adopts the online expected loss and the key misjudgment frequency. When the monitoring index continuously deteriorates, retraining is triggered and parameter combination traversal is re-executed. The cost matrix is adjusted synchronously when a new risk pattern appears in the review conclusion to ensure that the treatment template identification output always obeys the priority constraint of park treatment safety and convergence.

[0069] After the event chain construction and the disposition template identification determination are completed, the event chain state includes the event number sequence, the event type sequence, the process phase identification sequence, the evidence gap marker, the conflict point list, the source time credibility marker, the reverse sequence violation spectrum, the time semantic consistency level, the disposition convergence guarantee level, and the disposition template identification. The conflict point list locates the phase sequence problem to the event number pair, the source time credibility marker fixes the delay and the supplementary report form into a stable position, the reverse sequence violation spectrum associates the conflict with the evidence gap, the time semantic consistency level converges the timeline credibility degree into an executable determination, the disposition convergence guarantee level converges the influence domain diffusion and the feedback closing ability into a controllable determination, the disposition template identification outputs a clear path, and the review explanation has a consistent standard.

[0070] After the event chain state has been written into the disposition template identification, the conflict point list, the time semantic consistency level, and the disposition convergence guarantee level, and the event pool has been written into the park object identification, the process phase identification, the management token, and the management token version number, the park management action needs to be fixed into a trigger credential before entering the execution, and at the same time, the execution request needs to be fixed into a management action instruction. The lack of consistent coding standards of the trigger credential and the management action instruction will cause the feedback to be unable to be bound, and the review to be unable to restore the trigger basis. However, the word and operation standards in the aforementioned step four need to be revised. After the revision, step four develops around the event pool writing rule of the disposition template table execution end.

[0071] S401. Disposition template analysis and action element information locking.

[0072] The disposition template identification only represents the disposition policy code, and the execution semantics need to fall into the management action type and the execution end identification, so as to avoid different semantics of the same disposition template identification in different access ends. The disposition template table takes the disposition template identification as the retrieval key, returns the management action type and the execution end identification, and at the same time returns the rule identification. The rule identification is taken from the rule record that has been hit in the event chain rule table. The return result of the disposition template table is written into the event chain state, and the fields include the management action type, the execution end identification, and the rule identification. The disposition template identification and the management action type and the execution end identification form a one-to-one correspondence relationship, the delivery path remains consistent, and the review basis remains consistent.

[0073] S402. Management action instruction payload generation and payload check code generation.

[0074] The management action instruction needs to carry the park object identification, the management token, the management action type, and the disposition template identification. The field absence will cause the execution end to be unable to locate the object or to be unable to bind the management token. The management action instruction payload is generated in a fixed-length coding concatenation manner. The fixed-length coding rule is fixed as the park object identification coding segment, the management token coding segment, the management action type coding segment, and the disposition template identification coding segment. The length of each coding segment is registered by the disposition template table, and the left zero padding and the right truncation rule are registered by the disposition template table.

[0075] The payload check code is generated using a digest operation. This digest operation employs a preset secure hash mapping to compress an input string of arbitrary length into a fixed-length output string. The output strings are identical when the input strings are identical, and differing characters in the input strings result in different output strings. The payload check code input string is obtained by concatenating the park object identifier code segment, management token code segment, management action type code segment, and disposal template identifier code segment according to the coding order registered in the disposal template table. Each code segment uses a structure of segment length marker plus segment content. The segment length marker uses a unified character set in decimal representation, and the segment content uses a unified character encoding with escaped delimiters. Null values ​​are represented by a zero segment length marker, thus eliminating ambiguity caused by direct concatenation. The payload check code is written into the header of the management action instruction. After receiving the management action instruction, the execution end reconstructs the payload check code input string according to the same coding order and segment length marker rules, and recalculates the payload check code. If the recalculated result is inconsistent with the header payload check code, a consistency failure flag is written, and the application is rejected. After the event pool establishes a payload check code index for the instruction request record, it performs a second verification by comparing the consistency of four fields—park object identifier, management token, management action type, and disposal template identifier—during the binding process. Summary collisions are eliminated during the second verification, and the binding relationship between the instruction request record and the trigger voucher record remains reproducible.

[0076] S403. Trigger voucher generation and voucher verification code generation.

[0077] The trigger credential solidifies the triggering basis from the event chain state to the event pool. The triggering basis needs to have a recalculated encoding standard to support post-event restoration. The trigger credential fields are fixed as event number sequence, rule identifier, disposal template identifier, and management token version number. The event number sequence is arranged according to the sorting order in the event chain state. Each event number is converted into an event number encoding segment according to a fixed-length rule. The event number encoding segment uses a segment length marker plus segment content structure, and the segment content retains the original character sequence of the event number without rearrangement. The rule identifier, disposal template identifier, and management token version number are converted into corresponding encoding segments according to the fixed-length encoding rules registered in the disposal template table, and then appended to the end of the event number sequence encoding string in a fixed order to form the credential input string. The credential verification code uses the same digest operation as the payload verification code to generate the credential input string and write it to the trigger credential record. During verification, the trigger credential record fields are read from the event pool, and the credential input string is reconstructed according to the same encoding rules to recalculate the credential verification code. If the recalculation is inconsistent, it is directly determined that the trigger credential content has changed. The triggering basis remains stable and verifiable in the event pool.

[0078] S404. Trigger the binding of credential and instruction request record writing.

[0079] The trigger credential needs to establish a locatable binding relationship in the event pool, and the lack of binding relationship will cause the reply event to only find the management token but not the trigger basis. After the event pool writes the trigger credential record, the event pool writes the instruction request record, and the instruction request record field is fixed as the park object identifier management token management action type execution end identifier disposal template identifier payload check code credential check code. After the instruction request record is written, the read-only attribute is set. The credential check code and the payload check code form a binding anchor point together. The instruction request record and the trigger credential record under the same management token maintain a one-to-one correspondence, and the query path is stable.

[0080] S405. Management action instruction delivery and instruction acceptance identification write-back.

[0081] The management action instruction delivery needs to ensure that the execution end acceptance reply can be bound back to the instruction request record, avoiding the acceptance write falling into the wrong record. The execution end access table saves the delivery address and communication constraints corresponding to the execution end identifier. The management action instruction retrieves the delivery address according to the execution end identifier, and the delivery content includes the management action instruction payload, the payload check code, and the credential check code. When the execution end returns the instruction acceptance identifier, the event pool locates the instruction request record according to the credential check code, and then compares the consistency of the payload check code in the instruction request record with the returned payload check code, and the consistency of the disposal template identifier in the instruction request record with the returned disposal template identifier. If they are consistent, write the instruction acceptance identifier; if they are not consistent, write the acceptance refusal mark and retain the returned content. After the instruction acceptance identifier is written into the path, the reply binding path is clear, and the wrong binding probability is reduced.

[0082] After the management action instruction and the trigger credential are generated, the event pool simultaneously exists the trigger credential record and the instruction request record. The trigger credential record carries the event number sequence rule identifier, the disposal template identifier, the management token version number, the credential check code, and the instruction request record carries the park object identifier, the management token, the management action type, the execution end identifier, the payload check code, and the instruction acceptance identifier. The disposal template table maps the disposal template identifier to the management action type and the execution end identifier. The abstract operation compresses the input string into a fixed length check code. The event pool binds the acceptance result to the instruction request record through the write-back comparison. The disposal basis and the execution semantics maintain consistent caliber, and the review and trace have stable grasp.

[0083] The park management action has formed a trigger credential record and an instruction request record, the event chain state has been written into the handling template identification conflict point list time semantics consistency level handling convergence guarantee level. The on-site feedback can only rely on the reply event, and the reply event has three common disturbances: delay, repeated submission and missing fields. If the disturbance handling is not in place, the corresponding relationship between the trigger credential record and the instruction request record will be scattered. However, the handling link needs to keep the checkable evidence. Step S5 takes the reply event as the entrance, freezes the positioning verification, renews the update string into a consistent caliber, and ensures that the handling result can be traced back to the trigger basis.

[0084] S501. Receipt and pool registration of reply event.

[0085] The reply event deduplication key is used to identify repeated submission of the reply event, so as to avoid pollution of the event chain state by repeated reply. The reply event deduplication key input string is obtained by concatenating a management token coding segment, an instruction acceptance identification coding segment, a reply event type coding segment, an execution end identification coding segment, a payload verification code coding segment, and a credential verification code coding segment in a preset order. Each coding segment adopts a segment length marker and segment content structure, the segment content adopts uniform character coding and performs escaping on the separator character, and the null value adopts a zero segment length marker, so as to ensure one-to-one correspondence of the field combination and concatenation result. The reply event deduplication key performs digest operation on the input string to obtain and write into the event pool. Before writing the reply event record into the event pool, the event pool queries whether there is an existing same-key reply event record by using the reply event deduplication key. If there is, the reply event record is marked as repeated submission, and the event number sequence is stopped from being appended to the event chain state. If not, the pool registration is completed and participates in the reply verification. The deduplication rule can be reproduced in the same input string reconstruction mode in the review.

[0086] S502. Positioning instruction request record of reply event.

[0087] The instruction request record carries the park object identification management token management action type execution end identification handling template identification payload verification code credential verification code instruction acceptance identification. Unstable positioning rules will lead to incorrect instruction request record writing of the reply. The event pool uses the management token instruction acceptance identification execution end identification as a joint matching key to retrieve the instruction request record. After retrieval hit, the park object identification management action type handling template identification payload verification code credential verification code in the instruction request record are read. If the retrieval does not hit, the event pool writes the reply isolated marker into the reply event record, and generates a review work order. The review work order writes the management token instruction acceptance identification execution end identification reply event type platform arrival time credential verification code. After the joint matching key is fixed, the reply event landing point is unique, the wrong binding probability is significantly reduced, and the review positioning field is complete.

[0088] S503. Consistency verification and late arrival determination.

[0089] The handling link needs to meet content consistency and sequence consistency at the same time. The content consistency is used to confirm that the reply event corresponds to the same trigger credential record and the same instruction request record. The sequence consistency is used to confirm that the reply event does not destroy the process stage criterion after entering the event chain. The event pool reads the instruction request record and the trigger credential record. The trigger credential record contains the event number sequence rule identifier, the handling template identifier, the management token version number and the credential verification code. The event pool compares the reply event and the instruction request record item by item in terms of the execution end identifier, the management action type, the handling template identifier, the load verification code and the credential verification code. When the comparison is inconsistent, a consistency failure mark is written and the difference field list is recorded. The event pool compares the reply event and the trigger credential record item by item in terms of the credential verification code, the handling template identifier and the management token version number. When the comparison is inconsistent, a consistency failure mark is written and the difference field list is recorded. The event pool maps the reply event type to the process stage identifier and the stage sequence number. The mapping rule comes from the event chain rule table. The event chain state reads the last confirmed stage sequence number. When the reply event stage sequence number is less than the confirmed stage sequence number, a reverse sequence failure mark is written and the conflict is written into the conflict point list. The event pool compares the platform arrival time and the instruction request record writing time in terms of the front and back relationship. When the platform arrival time is obviously later than the writing time and the reply event deduplication key appears for the first time, a late mark is written. The late mark is used to distinguish the reissue reply and the real-time reply. The verification record interpretation path is clearer.

[0090] S504. Abnormality handling and freeze flag writing.

[0091] When any of the consistency failure mark, the reverse sequence failure mark and the reply isolation mark appears, the management token continues to advance, which will spread the error handling to linkage and dispatching. The event pool writes the freeze flag into the management token state record. The freeze reason category is written according to the failure type. The wrong binding class corresponds to the consistency failure. The missing class corresponds to the reply isolation. The sequence class corresponds to the reverse sequence failure. The event pool generates a review work order and writes it into the conflict point list, the trigger credential digest, the instruction request record digest and the reply event digest. The trigger credential digest contains the event number sequence rule identifier, the handling template identifier, the management token version number and the credential verification code. The instruction request record digest contains the park object identifier, the management action type, the execution end identifier, the load verification code and the instruction acceptance identifier. The reply event digest contains the reply event type, the platform arrival time, the late mark, the consistency failure mark and the reverse sequence failure mark. After the freeze flag and the review work order are stored in the database, the abnormality chain is isolated. The review entry field is concentrated and consistent. The review workload is controllable.

[0092] S505. Review conclusion storage and rule sedimentation update.

[0093] The review conclusion needs to be deposited as an executable update record of the event chain rule table and the delay image library to avoid the same type of abnormality from repeating. When the review conclusion confirms that the evidence item is missing, causing a misjudgment, the event chain rule table updates the evidence item requirement set under the corresponding process phase identifier entry, adds the necessary fields indicated by the review conclusion to the evidence item requirement set, and registers the field format legality criterion. The evidence gap marker trigger condition is updated synchronously, and the update record is written with a version number and an effective time, and the historical version is retained for review. When there is a rearrangement allowed situation in the review conclusion confirmation phase boundary, the event chain rule table updates the phase boundary set that allows rearrangement, and the phase boundary set records the ordered combination of the previous process phase identifier and the subsequent process phase identifier and registers the effective version number. The matching rule is based on the phase sequence number criterion. When the access channel of the review conclusion confirmation appears stable delay form changes, the delay image library updates the delay segment boundary. The update process is to collect the event records passed by the review conclusion and calculate the difference between the source time and the platform arrival time. After sorting the difference from small to large, the boundary position is determined according to the delay image library segment configuration. The segment configuration records the target coverage ratio of each delay segment in the sorted sequence. The boundary position takes the sequence position difference corresponding to the target coverage ratio and writes it into the delay segment boundary. The new version of the delay segment boundary is written and the old version is retained. After writing, the same batch of event records is played back to calculate the source time credibility marker and the time semantics consistency level. If the playback result is inconsistent with the abnormal type given by the review conclusion, the old version of the delay segment boundary is rolled back and a review work order is generated. The version switching has a traceable criterion.

[0094] After the reply comparison and review deposition are completed, the event pool forms a stable corresponding relationship between the reply event trigger and the voucher record instruction request record, and the verification result marker freezes the flag review work order to form a complete traceable chain. The event chain rule table solidifies the phase boundary set that allows rearrangement of the phase sequence criterion evidence item requirement set into executable rules, and the delay image library solidifies the access channel delay segment boundary into an updateable image. The park management action is continuous from triggering to reply, and the digital twin park management process has a traceable landing form.

[0095] Specifically, the above is only a preferred embodiment of the present application and is not used to limit the present application.

[0096] In the description of the present specification, the description referring to the terms "one embodiment", "an example", "a specific example" and the like means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0097] The preferred embodiments of the application disclosed above are only to facilitate the elucidation of the application. The preferred embodiments do not describe all the details of the application and limit the application to the specific embodiments. Obviously, many modifications and variations can be made in light of the teachings above. The description is chosen and described in order to provide the best illustration of the application and its practical application to those skilled in the art and to enable those skilled in the art to best utilize the application. The application is limited only by the claims and their full scope and equivalents.

Claims

1. A park management method based on digital twins, characterized in that, Including the following steps: S1: Receive the event type and source time from the park's business system, write them into the event pool, generate an event number, record the event type and source time access channel identifier, and complete the event entry registration into the pool; S2: Filter anchor events from the event pool, parse the park object identifier and process stage identifier based on the anchor events, generate management tokens, associate the management tokens with the event numbers, and write the tokens back to the event pool; S3: Include events carrying management tokens into the event chain, check the sequence of events and the completeness of evidence according to the event chain rule table, output the handling template identifier, and record the event chain status and conflict point list; S4: Generate management action instructions and trigger credentials based on the disposal template identifier. Write the trigger credentials into the event pool. Send the management action instructions to the execution system specified by the disposal template identifier and record the instruction acceptance identifier. S5: Receive execution receipts, generate receipt events, and add them to the event chain. Compare receipt events with trigger credentials. In case of an anomaly, freeze the management token and generate a review work order. Update the event chain rule table and delay profile library with the review conclusion.

2. The park management method based on digital twins according to claim 1, characterized in that, Step S1 includes: The access layer receives the event type and source time. The event type is validated using an event type dictionary and written to the event pool. The source time is converted to a standard timestamp and a source time missing flag is written to it. The access layer extracts the access channel identifier and obtains the channel sequence number. It calculates the time segment sequence number based on the time segment length and updates the channel time segment counter to obtain the segment sequence number. The event number is generated using the channel sequence number segment, time segment sequence number segment, and segment sequence number segment. The event pool writes the event number, event type, source time, source time missing flag, access channel identifier, channel sequence number, and platform arrival time.

3. The park management method based on digital twins according to claim 2, characterized in that, Step S2 includes: The anchor event table returns the set of allowed event types and the set of prohibited access channel identifiers for the anchor point, with the event type as the entry point. The event pool reads the event type and the access channel identifier. When the event type belongs to the set of allowed event types for the anchor point and the access channel identifier does not belong to the set of prohibited access channel identifiers, the event number is written into the anchor candidate queue. The anchor candidate queue maintains the same order of event numbers according to the arrival time of the platform.

4. The park management method based on digital twins according to claim 3, characterized in that, Step S2 also includes: The object mapping table returns the park object identifier by using the object clue as the entry point, and the process mapping table returns the process stage identifier and stage sequence caliber by using the event type as the entry point. The anchor point candidate queue reads the object clue corresponding to the event number one by one and obtains the park object identifier. The token primary key is generated based on the process stage identifier, and the management token version number is generated by combining the event number and the platform arrival time to obtain the management token. The token index table writes the valid management token with the park object identifier and suppresses stage backtracking according to the stage sequence caliber.

5. A park management method based on digital twins according to claim 4, characterized in that, Step S3 includes: The management token is used as a merge key to retrieve event records from the event pool and write them into the event chain. The event chain is sorted by the platform arrival time and event number. The event chain rule table maps the process stage identifier to the stage sequence number and provides a set of evidence requirements. A conflict point list is generated based on the descending position of the stage sequence number. The evidence gap mark is generated by verifying the event submission content based on the set of evidence requirements. The conflict point list records the event number pair and the process stage identifier pair.

6. A park management method based on digital twins according to claim 5, characterized in that, Step S3 also includes: The delay profile library provides delay segment boundaries based on event type and access channel identifier. It generates source time credibility markers and constructs reverse violation spectrum based on the difference between source time and platform arrival time. It adjudicates the time semantic consistency level based on the evidence gap markers of the stage boundary set that allows rearrangement of key stage set, the source time credibility markers, and the digital twin object relationship graph is expanded in two rounds to generate the set of influence domain objects and adjudicate the influence domain span markers. The receipt closure table and the event pool configuration library registration table are checked against the set of candidate management action types to obtain the handling convergence guarantee level. The cost-sensitive decision tree outputs the handling template identifier based on the time semantic consistency level and the handling convergence guarantee level and writes it into the event chain status.

7. A park management method based on digital twins according to claim 6, characterized in that, Step S4 includes: The disposal template table is written with disposal template identifier, retrieval management action type, execution terminal identifier, rule identifier and event chain status. The management action instruction payload is generated by concatenating fixed-length codes of park object identifier, management token, management action type and disposal template identifier. The payload check code is obtained by performing a digest operation on the payload input string. The payload check code is written into the management action instruction header and delivered to the delivery address corresponding to the execution terminal identifier along with the payload.

8. A park management method based on digital twins according to claim 7, characterized in that, Step S4 also includes: The event chain status generates an event number sequence code string in sorted order and concatenates it with the rule identifier, handling template identifier, management token, and version number to form a voucher input string. The voucher verification code is obtained by performing a digest operation on the voucher input string, triggering the voucher record to be written to the event pool and set to read-only attribute. The instruction request record is written to the event pool. The event pool locates the instruction request record according to the voucher verification code, verifies that the payload verification code is consistent, and writes the instruction acceptance identifier.

9. A park management method based on digital twins according to claim 8, characterized in that, Step S5 includes: The access layer receives the execution terminal identifier, management action type, management token, instruction acceptance identifier, payload check code, credential check code, receipt event type, and platform arrival time, and writes them into the event pool to generate a receipt event. The event pool generates a deduplication key for the receipt event based on the management token, instruction acceptance identifier, execution terminal identifier, and location instruction request record, as well as the management token, instruction acceptance identifier, receipt event type, execution terminal identifier, payload check code, and credential check code, and writes it into the event pool. If the deduplication keys of the receipt events are consistent, the receipt event records are merged.

10. A park management method based on digital twins according to claim 9, characterized in that, Step S5 also includes: The event pool compares the execution end identifier, management action type, disposal template identifier, load check code, and voucher check code of the receipt event and instruction request record, and compares the voucher check code, disposal template identifier, and management token version number of the receipt event and trigger voucher record to complete consistency verification. If the verification fails or the location fails, a freeze flag is written and a review work order is generated. The review conclusion updates the evidence item requirement set and the stage boundary set that can be rearranged in the event chain rule table, and updates the delay segment boundary in the delay profile library.

Images