A Smart Heating Multi-System Integrated Management and Control Method and Platform Based on Data Platform

By establishing a spatial mapping relationship between users and heat exchange stations in the centralized heating system, and adopting a 'batch + real-time' architecture to access the data platform for standardized processing, the problem of data silos has been solved, cross-system data association and rapid fault location have been achieved, and the decision-making efficiency and business process collaboration of heating companies have been improved.

CN123117435APending Publication Date: 2026-11-17HOHHOT KELIN THERMOELECTRICITY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611271346.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-20
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

The production scheduling, customer service work order, and billing management systems of centralized heating companies operate independently, resulting in severe data silos. This makes it impossible to achieve cross-system data correlation analysis, leading to low decision-making efficiency, fragmented business processes, and poor system scalability.

Method used

A spatial attribution mapping relationship between users and heat exchange stations is constructed. A hybrid synchronous architecture of "batch + real-time" is adopted to connect heterogeneous data from production scheduling and customer service work order systems to the data platform for standardized processing. When a customer service work order is created, the heat exchange station is automatically associated and located, and operating parameters are extracted for deviation comparison and rule matching to generate a fault analysis report.

Benefits of technology

It has achieved data format unification across multiple systems, improved fault location efficiency, reduced manual cross-system query time, enhanced decision-making efficiency and business process collaboration, and strengthened system scalability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117435A_ABST
    Figure CN123117435A_ABST
Patent Text Reader

Abstract

This invention discloses a smart heating multi-system integrated management and control method and platform based on a data middle platform, belonging to the field of digital and intelligent management and control technology for centralized heating systems. By constructing a spatial affiliation mapping relationship between users and heat exchange stations, a "batch + real-time" hybrid synchronization architecture is adopted to integrate heterogeneous data from the production scheduling system and customer service work order system into the data middle platform for standardized processing. Using the creation event of a customer service work order as a trigger, the associated heat exchange station is automatically located, and the current operating parameters and historical reference operating parameters for the complaint period are extracted for deviation comparison and rule matching. A fault analysis report with confidence score is output and pushed to the business terminal. This invention solves the data silo problem caused by the independent operation of multiple business systems, realizing a complete closed loop from data access, event triggering, deviation judgment to report push, significantly improving the efficiency and accuracy of fault location in heating quality complaint scenarios.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of digital and intelligent management and control technology for centralized heating systems, and in particular to a method, platform, and equipment for the integrated management and control of multiple intelligent heating systems based on a data middleware platform. Background Technology

[0002] As urbanization continues to advance in my country, the complexity of operation and management of centralized heating enterprises is increasing. Currently, large and medium-sized centralized heating enterprises typically deploy three core business systems: Supervisory Control and Data Acquisition (SCADA) system, customer service work order system, and billing management system.

[0003] The production scheduling system collects operating parameters such as temperature, pressure, and flow rate from heat exchange stations via PLCs. It uses a time-series database to store high-frequency operating data and a relational database to store site files and equipment ledgers. However, the data only serves the scheduling operation itself and is not accessible to other systems. The customer service work order system independently manages user repair requests and complaints, using a relational database to ensure transaction consistency. However, it lacks a connection channel with production operation data, so customer service personnel cannot know the heating operation status of the corresponding area when receiving a complaint. The billing management system independently manages heating area, heat consumption metering, and cost accounting, using a relational database to ensure financial transaction consistency. However, it also operates in isolation from other systems.

[0004] The three systems operate independently, with data stored in their respective databases, forming a closed data loop. This leads to the following systemic drawbacks: First, severe data silos prevent cross-system data correlation analysis. When faced with heating quality complaints, managers must log into each of the three systems separately to query information, making it impossible to complete a comprehensive diagnosis from a single view. Second, decision-making efficiency is low, as different systems use different statistical standards, making it difficult to form a unified operational analysis view. Third, business processes are fragmented; customer service cannot automatically link abnormal operational data from the production system after receiving a repair request, resulting in delays in fault location and dispatch. Fourth, system scalability is poor; adding new business systems requires developing customized interfaces for each existing system, leading to high development and maintenance costs.

[0005] To address the aforementioned issues, traditional data warehouse solutions aggregate data in batches, primarily for post-event analysis, and cannot support real-time cross-system data association and online business collaboration. Point-to-point interface solutions experience exponential growth in interfaces as the number of systems increases, and fail to resolve the core contradiction of inconsistent data standards. Therefore, there is an urgent need for a new method that can achieve real-time aggregation, association, and collaborative management of multi-source heterogeneous data without modifying the internal logic of existing systems. Summary of the Invention

[0006] The present invention aims to at least partially solve one of the technical problems in the related art.

[0007] To address this, this invention proposes a smart heating multi-system integrated management and control method based on a data platform. By constructing a spatial attribution mapping relationship between users and heat exchange stations, and employing a "batch + real-time" hybrid synchronization architecture, heterogeneous data from the production scheduling system and customer service work order system are integrated into the data platform and standardized. When a customer service work order is created, the system automatically associates and locates the heat exchange station using the user code as the entry point, extracts the operating parameters for the corresponding time window, performs deviation comparison and rule matching, and finally outputs a fault analysis report with confidence scores, which is then pushed to the business terminal, realizing "complaint" functionality. position diagnosis A data-driven closed-loop service for "order dispatch".

[0008] Another objective of this invention is to propose a smart heating multi-system integrated management and control platform based on a data middleware platform.

[0009] The third objective of this invention is to provide an electronic device.

[0010] To achieve the above objectives, this invention proposes a smart heating multi-system integrated management and control method based on a data platform, comprising: Establish a spatial attribution mapping relationship between users and heat exchange stations; Heterogeneous data from the production scheduling system and customer service work order system are integrated into the data platform and standardized to obtain standardized data. When a customer service work order creation event is captured, the mapping relationship is queried according to the user code in the work order to determine the associated heat exchange station. The heat exchange station is used as a spatial limit and the preset time window is extended based on the work order creation time. The current operating parameters and historical reference operating parameters of the heat exchange station are extracted from the standardized data. The extracted current operating parameters are compared with the historical reference operating parameters to obtain deviation indicators. The deviation indicators are then input into a preset fault determination rule set for matching, and the confidence scores corresponding to each candidate fault cause are output. Based on the output confidence scores, the fault causes that meet the preset conditions are selected, a fault analysis report associated with the corresponding work order is generated, and the fault analysis report is pushed to the business terminal.

[0011] An embodiment of the present invention provides a smart heating multi-system integrated management and control method based on a data middle platform, which may also have the following additional technical features: In one embodiment of the present invention, constructing a spatial affiliation mapping relationship between users and heat exchange stations includes: obtaining heat exchange station profile data from a production scheduling system and user profile data from a customer service work order system; establishing a spatial affiliation mapping table containing a hierarchical affiliation relationship of "user-community-heat exchange station" using the heat exchange station code in the production scheduling system as a first primary key and the user code in the customer service work order system as a second primary key; and storing the spatial affiliation mapping table in the cache layer of the data platform.

[0012] In one embodiment of the present invention, heterogeneous data from a production scheduling system and a customer service work order system are connected to a data platform, including: streaming the time-series operation data generated by the production scheduling system through a real-time message channel to the data platform; incrementally connecting the transaction change data generated by the customer service work order system to the data platform through a database log parsing channel; and writing the connected time-series operation data and the transaction change data into the source data layer of the data platform, retaining the original data format.

[0013] In one embodiment of the present invention, the standardization process for heterogeneous data includes: aligning the timestamps of different data sources according to a preset uniform time granularity, wherein the alignment formula is:

[0014] The division process uses integer division, the original time is in Unix timestamp format, and the aligned time is in the same unit as the original time, generating a time-series data sequence with a unified time base. According to a preset encoding mapping dictionary, the heterogeneous encoding fields of each system are converted into a unified standard encoding. The data after time alignment and encoding conversion is stored in the detailed data layer of the data platform.

[0015] In one embodiment of the present invention, the standardization process for heterogeneous data further includes: filling in short-term missing data using a linear interpolation method, wherein the filling formula is:

[0016] in, The interpolated value for the missing point. and These are the nearest valid values ​​before and after the missing point, respectively. The timestamp of the missing point. and These are the timestamps corresponding to the most recent valid values ​​before and after; for data sequences that are continuously missing for more than a preset duration, they are marked as missing and will not participate in the subsequent deviation index calculation.

[0017] In one embodiment of the present invention, a preset time window is extended based on the work order creation time to extract the current operating parameters corresponding to the heat exchange station from the standardized data. This includes: using the creation time of the target work order as the base time, extending forward by a first preset time and backward by a second preset time to form a sliding time window; using the associated heat exchange station code as a spatial filtering condition, extracting the operating value sequence of each parameter type of the heat exchange station within the sliding time window from the standardized data as the current operating parameters.

[0018] In one embodiment of the present invention, extracting the historical synchronous reference operating parameters corresponding to the heat exchange station from the standardized data includes: determining the historical synchronous time interval based on the creation time of the target work order; wherein the historical synchronous time interval is a continuous period of time within the previous 7 natural days that is the same hour as the creation time, and the duration of the continuous period is the same as the duration of the preset time window; using the associated heat exchange station code as a spatial filtering condition, extracting the operating value sequence of each parameter type of the heat exchange station within the historical synchronous time interval from the standardized data, and calculating the arithmetic mean of the same parameter type within the historical synchronous time interval as the historical synchronous reference operating parameter.

[0019] In one embodiment of the present invention, the deviation index is input into a preset fault determination rule set for matching, and a confidence score corresponding to each candidate fault cause is output. A fault analysis report is generated based on each confidence score, including: matching the deviation index with multiple fault determination rules in the fault determination rule set level by level, each fault determination rule corresponding to a candidate fault cause and having a preset weight coefficient; accumulating the weight coefficients corresponding to each matched fault determination rule, calculating the comprehensive confidence score of each candidate fault cause; filtering out candidate fault causes whose comprehensive confidence score is higher than a preset threshold, and generating a fault analysis report containing a fault type label, a confidence percentage value, and a handling suggestion field.

[0020] To achieve the above objectives, this application also proposes a smart heating multi-system integrated management and control platform based on a data middleware platform, comprising: The spatial mapping module is used to construct the spatial affiliation mapping relationship between users and heat exchange stations; The data access module is used to connect heterogeneous data from the production scheduling system and the customer service work order system to the data platform and perform standardized processing to obtain standardized data. The parameter extraction module is used to query the mapping relationship based on the user code in the work order when a customer service work order creation event is captured to determine the associated heat exchange station, and to extract the current operating parameters and historical reference operating parameters of the heat exchange station from the standardized data by using the heat exchange station as a spatial limitation and the work order creation time as a benchmark to expand a preset time window. The fault determination module is used to compare the extracted current operating parameters with the historical reference operating parameters to obtain a deviation index, and input the deviation index into a preset fault determination rule set for matching, and output the confidence score corresponding to each candidate fault cause. The linkage output module is used to filter out the fault causes that meet the preset conditions based on the confidence scores of each output, generate a fault analysis report associated with the corresponding work order, and push the fault analysis report to the business terminal.

[0021] This invention discloses a smart heating multi-system integrated management and control method and platform based on a data middle platform. By constructing a spatial attribution mapping relationship between users and heat exchange stations and caching it in the data middle platform, it automatically reverse-locates the heat exchange station using the user code as the entry point when a customer service work order is created. This achieves millisecond-level querying of the three-level spatial topology of "user-community-heat exchange station," solving the problem of the lack of spatial association channels between traditional customer service systems and production systems. Furthermore, through a "batch + real-time" hybrid synchronization architecture, it integrates heterogeneous data from the production scheduling system and the customer service work order system into the data middle platform and performs standardized processing, without modifying the internal logic of existing business systems. Under this premise, it achieves unified aggregation of heterogeneous data sources and unified time benchmarks, solving the integration obstacles of different data formats, inconsistent time granularity, and inconsistent coding standards among multiple systems. By expanding the preset time window based on the work order creation time and extracting the current operating parameters of the heat exchange station and the historical reference operating parameters for the same period for deviation comparison and rule matching, the data association is transformed from "manual table lookup after the fact" to "automatic triggering during the event". This enables rapid automatic location of fault causes and quantitative assessment of confidence in heating quality complaint scenarios, fundamentally solving the problems of cross-system data silos, low decision-making efficiency, and fragmented business processes in existing technologies.

[0022] To achieve the above objectives, a third aspect of this application provides an electronic device, including a processor and a memory; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement a smart heating multi-system integrated management and control method based on a data platform as described in the first aspect embodiment.

[0023] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0024] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating a smart heating multi-system integrated management and control method based on a data middle platform according to an embodiment of the present invention. Figure 2 This is an overall system architecture diagram of a smart heating multi-system integrated management and control platform based on a data middleware according to an embodiment of the present invention; Figure 3 This is a flowchart of the entire data processing workflow of the data platform in an embodiment of the present invention; Figure 4 This is a structural diagram of the functional modules of the integrated management and control platform according to an embodiment of the present invention; Figure 5 This is the core entity relationship diagram of the multidimensional data association graph in this embodiment of the invention; Figure 6 This is a flowchart illustrating the linkage between customer service work orders and production data in an embodiment of the present invention. Detailed Implementation

[0025] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0026] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0027] The following description, with reference to the accompanying drawings, describes a smart heating multi-system integrated management and control method, platform, and electronic equipment based on a data middleware platform, according to an embodiment of the present invention.

[0028] Example 1 like Figure 1 As shown, the present invention provides a smart heating multi-system integrated management and control method based on a data middle platform, comprising the following steps: Step S1, Constructing the spatial affiliation mapping relationship: Obtain heat exchange station data from the production scheduling system and user profile data from the customer service work order system.

[0029] In one embodiment of the present invention, from a physical perspective, the operation and scheduling of a centralized heating system uses heat exchange stations as the basic management unit, customer service uses users as the basic service unit, and billing management also uses users as the basic accounting unit. A spatial affiliation chain naturally exists among the three: "users belong to the community, and the community is heated by the heat exchange station." However, because the three business systems are built independently, the same heat exchange station may be identified by a "station code" in the production system, but by a "station name" or "station address" in the customer service system. The same user may have different user codes in the customer service system and the billing system. This inconsistency in the coding system is the root cause of the inability to link data across systems.

[0030] Based on the above mechanism, this invention first establishes a unified cross-system coding and identification system, using the heat exchange station code in the production scheduling system as the first global primary key and the user code shared by the customer service work order system and the billing management system as the second global primary key. Specifically, it obtains the file information of each heat exchange station in the production scheduling system, such as its code, name, address, and heating area, and obtains the file information of each user in the customer service work order system, such as their code, name, contact number, house address, and residential area. Using the heat exchange station code as the first primary key and the user code as the second primary key, a system is constructed that includes "user..." residential area A spatial attribution mapping table is constructed to establish a hierarchical relationship between heat exchange stations. Each row contains three levels of mapping information: user code, the code or name of the user's residential area, and the code of the heat exchange station supplying heat to that residential area. The constructed spatial attribution mapping table is stored in the Redis cache layer of the data platform to support millisecond-level fast queries for the user's assigned heat exchange station when a customer service ticket is created.

[0031] Step S2, Multi-source heterogeneous data access and standardization: Heterogeneous data from the production scheduling system and customer service work order system are integrated into the data platform and processed in a standardized manner.

[0032] In one embodiment of the present invention, from the perspective of data characteristics, the data of the production scheduling system is high-frequency time-series data (continuously generated at the second or minute level), the data of the customer service work order system is transactional data (generated only when work orders are created and updated), and the data of the billing management system is batch operation data (summarized daily or monthly). These three types of data differ significantly in terms of generation frequency, data structure, storage medium, and timeliness requirements. The production scheduling system typically uses time-series databases such as TDengine to store operational data (high-frequency writing, querying by time range) and MySQL to store configuration data such as site files and equipment ledgers. The customer service work order system uses relational databases such as Oracle to store work order transaction data, requiring support for complex multi-table join queries and transaction consistency. The billing management system also uses relational databases such as Oracle, with the highest requirements for data consistency and security. If a unified access method is adopted, either real-time performance will be sacrificed (using T+1 batch processing for all data) or the system burden will increase (using real-time synchronization for all data will put pressure on the transaction database).

[0033] Based on the aforementioned differences in data characteristics, this invention employs a hybrid synchronization architecture of "batch + real-time" to achieve differentiated access. Specifically: For the time-series operational data generated by the production scheduling system (data from measuring points such as temperature, pressure, flow, and heat at heat exchange stations, collected every 5 minutes), the data is streamed to the data platform via a real-time Kafka message channel. Data is pushed as soon as it is generated, with end-to-end latency controlled within 10 seconds. As a distributed message queue, Kafka can buffer high-frequency data streams, preventing the production system database from being directly exposed to external access pressure.

[0034] For transaction-related data changes (work order creation, status transitions, processing records, etc.) generated by the customer service ticket system, the database redo log (binlog) is monitored using Canal or Flink CDC components. Data change events are captured incrementally and synchronized to the data platform, with latency controlled within 5 minutes. This method is non-intrusive to the source database, does not add extra query load, and is suitable for near real-time synchronization scenarios of transactional databases.

[0035] For the operational accounting data (accounts receivable, actual revenue, outstanding fees, invoices, etc.) generated by the billing management system, the DataX offline batch synchronization tool is used to perform T+1 batch synchronization during the daily business off-peak period (such as 2:00 am), which ensures data integrity while minimizing the performance impact on the source database.

[0036] The accessed time-series running data and transaction change data are written into the source data layer (ODS layer) of the data platform, and the original data format is preserved in the form of Hive external tables to achieve the traceability and rerun capability of the original data.

[0037] In one embodiment of the present invention, after the above three types of data are accessed, they need to undergo standardization processing before entering the subsequent analysis process. The standardization processing includes three sub-steps: timestamp alignment, encoding unification, and missing value handling.

[0038] S21, Timestamp Alignment. Timestamp precision varies across different data sources: SCADA systems may be accurate to the second or even millisecond, customer service ticket systems to the minute, and billing systems to the date. Furthermore, the servers hosting these systems may have different system time zones. To eliminate time base differences, all timestamps are first converted to Beijing time (UTC+8), and then aligned by rounding down to the nearest 5-minute interval. The alignment formula is:

[0039] The division uses integer division. The original time, `original_time`, is in Unix timestamp format (in seconds). The denominator 300 corresponds to 5 minutes (300 seconds). After alignment, the time has the same unit as the original time. The essence of this integer division operation is: divide the original second-level timestamp by 300, discard the remainder (round down), and then multiply by 300, thereby merging any second-level timestamp to its corresponding 5-minute hour. For example, if the original timestamp is 1640995200 (corresponding to 2022-01-01 00:01:30), dividing by 300 gives 5469984.0 (rounded down to 5469984), multiplying by 300 gives 1640995200 (corresponding to 2022-01-01 00:00:00); similarly, if the original timestamp is 1640995320 (corresponding to 2022-01-01 00:02:00), dividing by 300 gives 5469984.4 (integer division yields 5469984), multiplying by 300 gives 1640995200, which is also rounded down to 00:00:00. This formula is applicable to original timestamps of arbitrary precision, generating time-series data sequences with a unified time base. For data with an original frequency higher than 5 minutes (such as SCADA second-level data), time window aggregation is used: the average value within the window is used for state measurements such as temperature and pressure, and the last value within the window is used for cumulative measurements such as flow rate and heat. For data with an original frequency lower than 5 minutes, forward filling is used, filling the subsequent empty spaces with the most recent valid value until the next valid value arrives.

[0040] S22, Unified Coding. Different business systems use different coding conventions for the same entity. For example, a heat exchange station might be coded as "ZH-037" in the production system, but in the customer service system it might be recorded as "Station 037" or simply use the station name. User request types in the customer service system are categorized as "Repair Request," "Complaint," and "Inquiry," coded as "R," "C," and "I" respectively, while work order statistics reports expect to use the full Chinese name. To eliminate coding differences, this invention uses a pre-set coding mapping dictionary to convert the heterogeneous coding fields from various systems into a unified standard coding system. The standard coding system includes: heat exchange station coding uses "HZ" + a 3-digit serial number (e.g., HZ001). (HZ999) Measuring point codes adopt the format "medium_location_parameter_serial number" (e.g., WATER_SUPPLY_TEMP_01). Request type codes are uniformly mapped to four standard labels: "Repair Report," "Complaint," "Inquiry," and "Suggestion." Payment status codes are uniformly mapped to four standard statuses: "Paid," "Out of Balance," "Partially Paid," and "Exempt from Payment." Community codes adopt the format "6-digit administrative division code + 4-digit sequence number." All coding mapping relationships are uniformly maintained at the data governance layer. Adding or changing codes only requires updating the mapping dictionary; no modifications to individual business systems are needed.

[0041] S23, Missing Value Handling. Due to sensor malfunctions, communication interruptions, or system anomalies, missing values ​​are unavoidable in the raw data. This invention employs a tiered processing strategy for different types of missing data. For short-term missing data (continuous missing values ​​not exceeding 30 minutes), linear interpolation is used to fill in the missing values. The filling formula is:

[0042] in, The interpolated value for the missing point. and These are the nearest valid values ​​before and after the missing point, respectively. The timestamp of the missing point. and These are the timestamps corresponding to the most recent valid values ​​before and after the given points. The geometric meaning of this linear interpolation is that, given two points... and Connect them with straight lines, missing points corresponding ordinate The linear proportions along this straight line are based on the physical principle that the changes in the operating parameters of the heating system over a short period of time can be approximated as linear gradual changes. For data missing for 30 minutes to 2 hours, historical values ​​from the same period are used for imputation, with the arithmetic mean of the corresponding values ​​at the same time over the previous 7 calendar days used as the imputation value. For data missing for 2 hours to 24 hours, historical values ​​from the same period are used for imputation with a trend correction term. For data sequences missing for more than 24 hours, they are marked as missing and are not included in subsequent deviation index calculations to ensure the reliability of subsequent analyses.

[0043] After time alignment, encoding conversion, and missing data filling are completed, the data is stored in the detailed data layer (DWD layer) of the data platform. This layer of data provides a standardized data foundation with a unified format, unified time reference, and unified encoding system for all subsequent analyses.

[0044] Step S3, Extraction of Customer Service Event Triggering and Spatiotemporal Correlation Parameters: When a customer service work order creation event is captured, the mapping relationship is queried based on the user code in the work order to determine the associated heat exchange station. The heat exchange station is used as a spatial limit, and the preset time window is extended based on the work order creation time. The current operating parameters and historical reference operating parameters of the heat exchange station are extracted from the standardized data.

[0045] In one embodiment of the invention, step S2 establishes a standardized data foundation, but the standardized data itself is still a statically stored historical record. This is to achieve "customer service complaints". The real-time linkage goal of "automatic fault location" requires the precise extraction of spatiotemporal data fragments related to the current complaint from massive standardized data triggered by the dynamic event of work order creation.

[0046] Specifically, this invention sets up a data change listener in the customer service ticket system to monitor the creation of new tickets in real time. When a user calls the customer service hotline or submits a repair / complaint through online channels, and the customer service staff enters and saves the ticket information, the system captures the ticket creation event and obtains two key pieces of information carried by the ticket: the target user code and the time the target ticket was created.

[0047] First, using the target user's code as the query key, a search is performed in the space affiliation mapping table built and cached in Redis in step S1 to match and obtain the code or name of the cell to which the user belongs. Then, based on the "cell", The mapping relationship of "heat exchange station" determines the code of the heat exchange station responsible for supplying heat to the community. This process uses the hash table structure of Redis to achieve a fast query with O(1) time complexity. From user code input to heat exchange station code output, the entire query chain is controlled within milliseconds. The heat exchange station code obtained from this query is used as a spatial filtering condition to limit the scope of all subsequent parameter extractions—only heat exchange station data directly related to the heating area where the complaining user is located is extracted, rather than a full data scan.

[0048] Secondly, using the creation time of the target work order as the baseline time, a first preset time period (preferably 1 hour) is extended forward and a second preset time period (preferably 1 hour) is extended backward to form a 2-hour sliding time window centered on the complaint time. The selection of this window length is based on the physical propagation characteristics of heating faults: when the heat exchange station experiences insufficient heating, changes in the user's indoor temperature are usually perceived and trigger complaints within 30 minutes to 2 hours. Therefore, a 1-hour window before and after the complaint time can completely cover the entire time interval from "fault occurrence or deterioration" to "user perception and complaint".

[0049] Then, using the associated heat exchange station code as a spatial constraint and the sliding time window as a temporal constraint, the operating value sequence of each parameter type of the heat exchange station within the 2-hour window is extracted from the DWD layer standardized data generated in step S2, and used as the current operating parameters. The parameter types include, but are not limited to: primary network supply water temperature, primary network return water temperature, secondary network supply water temperature, secondary network return water temperature, primary network supply water pressure, primary network return water pressure, secondary network supply water pressure, secondary network return water pressure, supply and return water pressure difference, instantaneous flow rate, and accumulated heat.

[0050] Meanwhile, to determine whether the current operating status is abnormal, it is necessary to refer to the "normal operating parameter levels of the heat exchange station". This invention uses the creation time of the target work order as a benchmark to determine the historical contemporaneous time interval. This historical contemporaneous time interval is a continuous period of time within the previous 7 natural days that coincides with the creation time, and the duration of this continuous period is the same as the duration of the preset time window (i.e., also 2 hours). For example, if the work order is created on January 15, 2026 at 14:30, then the historical contemporaneous time interval is the period from 14:00 to 16:00 each day from January 8 to 14, 2026 (the previous 7 days). Using the associated heat exchange station code as a spatial filtering condition, the operating value sequence of each parameter type of the heat exchange station within the historical contemporaneous time interval (a total of 7 days) is extracted from the DWD layer standardized data. (2 hours / day = 14 hours of operating data), and calculate the arithmetic mean of the same parameter type within the same historical time interval, as the reference operating parameter for the same historical period.

[0051] Through the above steps, when each customer service work order is created, the system automatically outputs a set of four-tuple data: "complaint time + complaint area + current operating parameters + historical reference operating parameters", providing complete input for subsequent fault diagnosis.

[0052] Step S4, Deviation Calculation and Fault Determination: The extracted current operating parameters are compared with the historical reference operating parameters to obtain a deviation index. The deviation index is then input into a preset fault determination rule set for matching, and the confidence score corresponding to each candidate fault cause is output.

[0053] In one embodiment of the present invention, the current operating parameters extracted in step S3 and the historical reference operating parameters from the same period constitute a comparison between "current situation" and "benchmark". The essence of a heating system malfunction is that the operating parameters deviate from the normal range or normal pattern. However, isolated deviations of a single parameter are insufficient to determine the type of malfunction. For example, "low water supply temperature" may stem from multiple causes such as insufficient heat source output, abnormal opening of primary network valves, secondary network circulating water pump malfunction, or blockage in the user's indoor system. Therefore, the present invention employs a method of comprehensively matching multi-dimensional deviation indicators with a fault determination rule set to achieve automatic fault identification.

[0054] Specifically, for each parameter type, the values ​​at each moment in the current operating parameter sequence are compared item by item with the historical reference operating parameters (the arithmetic mean of the same parameter type for the same period over the previous 7 days), and the real-time deviation rate of that parameter at the target creation moment is calculated. The formula for calculating the deviation rate is:

[0055] This deviation rate is a percentage of relative deviation; a positive value indicates the parameter is above the normal level, and a negative value indicates the parameter is below the normal level. For example, if the secondary network water supply temperature at the time of the complaint is 42.3℃, and the arithmetic mean of this parameter for the same period over the previous 7 days is 51.6℃, then the deviation rate = (42.3 - 51.6) / 51.6 100% -18.0% indicates that this parameter is 18% lower than the historical normal level.

[0056] The calculated real-time deviation rates are matched step-by-step against a multi-dimensional fault determination rule set pre-installed in the rule engine. This fault determination rule set is pre-configured by heating industry experts based on their industry operational experience. Each fault determination rule contains three elements: trigger conditions (the deviation rate of one or more parameters meets a specific threshold combination), the corresponding candidate fault cause, and the rule's weight coefficient. Exemplary rules include, but are not limited to: Rule 1: If "secondary network water supply temperature deviation rate" "And the temperature difference between supply and return water" If "C" is selected, the candidate fault cause "insufficient heating at the heat exchange station" is matched, with a weighting coefficient of 0.30. Rule 2: If "primary network return water temperature deviation rate" "And the deviation rate of primary network water supply pressure" If the value is '", then the candidate fault cause "abnormal opening of primary network valve" is matched, with a weighting coefficient of 0.25. Rule 3: If "the deviation rate of the pressure difference between the secondary network supply and return water" "And the deviation rate of secondary network water supply pressure" If the value is '", then the candidate fault cause "secondary network circulating water pump failure" is matched, with a weighting coefficient of 0.25. Rule 4: If "primary network water supply temperature deviation rate" "And if "the primary network water supply pressure is normal", then the candidate fault cause "insufficient heat source output" is matched, with a weighting coefficient of 0.20.

[0057] During the matching process, there may be situations where the same set of deviation indicators simultaneously matches multiple rules. In such cases, it is necessary to comprehensively evaluate the confidence level of each candidate fault cause. This invention uses a weighted cumulative method to calculate the comprehensive confidence score of each candidate fault cause: the weight coefficients of all rules matching the candidate fault cause are summed to obtain the comprehensive confidence score of the fault cause, presented as a percentage. The higher the confidence score, the more independent rules support the fault cause, and the greater its probability of occurrence.

[0058] For example, if the deviation index at the time of the complaint hits both Rule 1 (weight 0.30) and Rule 2 (weight 0.25), and both Rule 1 and Rule 2 point to the candidate fault cause "insufficient heating at the heat exchange station" (in actual rule configuration, different rules can point to the same candidate cause or different candidate causes), then the overall confidence score of this candidate cause = 0.30 + 0.25 = 0.55, that is, a confidence level of 55%.

[0059] Step S5, Fault Analysis Report Generation and Business Linkage Output: Based on the output confidence scores, the fault causes that meet the preset conditions are selected, a fault analysis report associated with the corresponding work order is generated, and the fault analysis report is pushed to the business terminal.

[0060] In one embodiment of the present invention, the candidate fault causes and their confidence scores output in step S4 constitute a candidate list for fault diagnosis. To select the most credible fault cause from the candidate list and output actionable remedial suggestions, the present invention further performs a screening and report generation process.

[0061] Specifically, the confidence scores of each candidate fault cause are sorted from high to low, and candidate fault causes with a comprehensive confidence score higher than a preset threshold (preferably 0.6, i.e., 60%) are selected. If there are multiple candidate causes higher than the threshold, all of them are included in the report; if all candidate causes are lower than the threshold, a prompt message of "no clear matching fault" is generated, suggesting manual intervention for troubleshooting.

[0062] A structured fault analysis report is generated based on the screening results. The report must include at least the following fields: fault type label (e.g., "insufficient heating at the heat exchange station," "abnormal primary network valve opening," "secondary network circulating water pump failure," etc.), confidence percentage value, and corresponding standardized handling suggestions (e.g., "suggest checking primary network valve opening," "suggest checking circulating water pump operating status," "suggest checking heat source output parameters," etc.). If the report contains multiple candidate fault causes, they are displayed in descending order of confidence level.

[0063] The generated fault analysis report is automatically linked to the additional information field of the target work order via data write-back, allowing customer service personnel to directly view the fault diagnosis conclusion on the work order details page without switching systems. Simultaneously, the report is pushed to the customer service processing terminal (for reference when communicating with users) and the dispatch monitoring screen (for remote confirmation by operation dispatchers), triggering a tiered handling process. If the system determines it to be a heat exchange station equipment fault (confidence level higher than the threshold and fault type: equipment abnormality), an alarm is simultaneously pushed to the dispatch center screen, allowing dispatchers to remotely adjust operating parameters or dispatch maintenance personnel to the site based on the suggested measures in the report.

[0064] After the target work order is processed, the actual fault attribution data (i.e., the actual fault cause confirmed manually) entered by customer service or maintenance personnel is obtained. This actual fault attribution data is then sent back as positive and negative samples to the sample database of the data platform. Periodically (e.g., weekly), based on the returned sample data, the matching parameters or weight coefficients corresponding to the fault causes in the fault determination rule set are updated. For example, if the positive correlation between the hit rate and the manual confirmation rate of a certain rule decreases, its weight coefficient is reduced; conversely, if a rule consistently has a high accuracy rate, its weight coefficient is increased. Through this closed-loop iterative mechanism, the fault determination rule set is continuously optimized in actual operation, constantly improving the accuracy of automatic fault location.

[0065] The method of this invention constructs a spatial attribution mapping relationship between users and heat exchange stations and caches it in a data platform. It employs a "batch + real-time" hybrid synchronous architecture to integrate heterogeneous data from the production scheduling system and customer service work order system into the data platform and perform standardized processing. Using the customer service work order creation event as a trigger, it automatically locates the associated heat exchange station and extracts current operating parameters and historical reference operating parameters within a preset time window before and after the complaint time. Deviation comparison and rule matching are then performed, and a fault analysis report with confidence score is output and pushed to the business terminal. This effectively solves the problems of existing technologies where fault location relies on manual cross-system queries and suffers from slow response due to the independent operation of multiple business systems and the inability to communicate with each other. It achieves an integrated solution for the entire process from spatial mapping construction, heterogeneous data aggregation and standardization, event trigger parameter extraction, deviation judgment to report generation and push, significantly improving the efficiency and accuracy of fault location in heating quality complaint scenarios. Specifically, the initial fault location by customer service is reduced from 15 minutes in the traditional method to 30 seconds, improving the overall processing efficiency by approximately 70%. This enhances the adaptability and engineering practical value of the method in the multi-system heterogeneous data environment of heating enterprises.

[0066] Example 2 This embodiment takes a smart heating platform construction project of a centralized heating company as an example to further explain the specific implementation process of the method of the present invention.

[0067] Project Overview: This company is responsible for the centralized heating supply in the main urban area of ​​a northern city. It currently has 82 heat exchange stations, covering a heating area of ​​approximately 12 million square meters and serving about 150,000 households. The company has deployed three major business systems: a production scheduling SCADA system, a customer service work order system, and a billing management system. However, these three systems operate independently, and data cannot be shared. Management faces severe data silos, low decision-making efficiency, and fragmented business processes, necessitating the construction of a unified data platform and integrated management and control platform using the method described in this invention.

[0068] Step 1: Deployment of data middle platform infrastructure.

[0069] The hardware and software environment for the data middle platform was built using a technical solution that combines relational databases and big data platforms. The production scheduling SCADA system uses the TDengine time-series database to store operational data (82 heat exchange stations). Each station has approximately 30 measurement points, with a data collection frequency of 5 minutes per measurement. MySQL is used to store site files, equipment ledgers, and other configuration data. The customer service system uses Oracle 19c (2-node RAC cluster) to store work order data. The billing system uses Oracle 19c (2-node RAC cluster) to store operational data.

[0070] The big data platform uses Hadoop (CDH / HDP 3.1.5, 5-node cluster), including HDFS distributed file storage and YARN resource scheduling; real-time computing uses Apache Flink 1.17 (3 nodes); offline computing uses Apache Spark 3.4 (based on YARN); message queues use Apache Kafka 3.5 (3-node Broker cluster); analytical databases use Apache Doris 2.0 (3-node FE+BE cluster); graph databases use Neo4j 5.x to store relational graphs; and caching uses Redis 7.0 (3-node Sentinel cluster) to cache space mapping tables and hot data.

[0071] Step 2: Constructing spatial affiliation mapping relationships.

[0072] Retrieve the heat exchange station files (82 records: station code, station name, address, heating area) from the production scheduling MySQL database and the user files (approximately 150,000 records: user code, name, contact number, community name, house address) from the customer service Oracle database.

[0073] The heat exchange station is coded "HZ001" in the production scheduling system. Using "HZ082" as the primary key and the user code (globally unique, in the format "USER" + 10 digits) from the customer service system as the secondary primary key, a structure is built containing "user"... residential area Spatial attribution mapping table for the three-level attribution relationship of a heat exchange station. Taking a specific user as an example: User code USER1002345678 The residential community is named "Green Garden Community". The heat exchange station is coded HZ037 and stored as a single mapping record. This mapping table contains approximately 150,000 records, occupying about 15MB of Redis storage space, with a query response time of [missing information]. 2ms.

[0074] Step 3: Configure multi-source heterogeneous data access.

[0075] Configure data synchronization tasks for the databases of the three major business systems respectively: (i) Production operation data synchronization: Configure a Kafka Producer on the PLC data acquisition terminal of the heat exchange station. As soon as data is generated, it is pushed to the "scada_realtime" topic of the Kafka cluster. Flink consumes the data in this topic and writes it to the Hive ODS layer, with end-to-end latency. 10 seconds. Approximately 708,000 new operational data entries are added daily (82 stations). 30 measuring points 24 hours 12 times / hour = 708,480 messages / day).

[0076] (ii) Customer service ticket data synchronization: Deploy the Canal component on the customer service Oracle database, monitor its redo logs, capture INSERT and UPDATE events of the ticket table in real time, process them through Flink CDC and write them to the Hive ODS layer with a delay. 5 minutes. This method has no performance impact on the Oracle source database.

[0077] (III) Billing data synchronization: Configure a DataX scheduled task to perform batch synchronization at 2:00 AM every day, export the previous day's billing transaction data from Oracle and write it to the Hive ODS layer, using the T+1 batch synchronization method.

[0078] After the three types of data are synchronized to the ODS layer, the total number of new data entries per day is about 4.65 million, which is compressed and stored at about 270MB / day. In a heating season (180 days), there are about 840 million data entries, which are stored at about 50GB.

[0079] Step 4: Data cleaning and standardization.

[0080] Perform standardization processing on the raw data aggregated to the ODS layer: Step 4.1, Timestamp Alignment. Convert all data timestamps to Beijing time (UTC+8) and round down to the nearest 5-minute interval. For example, a temperature value timestamp recorded by the SCADA system as 2026-01-15 14:03:27 is rounded down to 2026-01-15 14:00:00; a work order creation timestamp recorded by the customer service system as 2026-01-15 14:02:00 is also aligned to 2026-01-15 14:00:00. This alignment operation unifies the time coordinates of data from different systems with different time precisions onto a 5-minute time grid, providing a time reference for cross-system correlation analysis.

[0081] Step 4.2, Outlier Filtering. Based on the "Technical Specification for Energy-Saving Retrofitting of Heating Systems" (GB / T 50893-2013) and industry operating experience, reasonable value range thresholds are set for each parameter: primary network supply water temperature 40-130℃, primary network return water temperature 30-80℃, secondary network supply water temperature 35-85℃, secondary network return water temperature 25-60℃, primary network supply water pressure 0.4-1.6MPa, primary network return water pressure 0.2-1.0MPa, secondary network supply water pressure 0.2-0.8MPa, secondary network return water pressure 0.1-0.6MPa, supply and return water pressure difference 0.02-0.3MPa, and user indoor temperature 5-35℃. Data points exceeding the upper and lower limits are marked as out of range (quality=2) and entered into the abnormal data pool for manual review.

[0082] Step 4.3, missing value imputation. For short-term missing data (continuous missing values) For missing values ​​(within 30 minutes), linear interpolation is used to fill in the gaps. For example, if the secondary network water supply temperature at a station is 51.5℃ at 14:00 and 52.3℃ at 14:30, then the five missing values ​​at 14:05, 14:10, 14:15, 14:20, and 14:25 are interpolated linearly as follows: 14:05 is 51.63℃, 14:10 is 51.77℃, 14:15 is 51.90℃, 14:20 is 52.03℃, and 14:25 is 52.17℃. For consecutive missing values ​​of 30 minutes to 2 hours, the average of the same period over the previous 7 days is used for interpolation; for consecutive missing values ​​of 2 to 24 hours, historical data plus trend correction is used; for consecutive missing values ​​exceeding 24 hours, they are marked as missing and not included in the calculation.

[0083] Step 4.4, Unified Coding. Establish and map a unified coding standard: the coding for heat exchange stations is unified as HZ001. HZ082 format; the request type code is uniformly mapped to four standard labels: “Repair Report”, “Complaint”, “Consultation”, and “Suggestion”; the payment status is uniformly mapped to “Paid”, “Outstanding”, “Partially Paid”, and “Exempt”; the community code adopts “6-digit administrative division + 4-digit sequence number”, such as “1501050001”.

[0084] The standardized data is written into the DWD layer (detailed data layer). This layer of data has a unified format, a unified time base, and a unified coding system, and can be directly used for subsequent analysis and querying.

[0085] Step 5: Data association and fusion.

[0086] The association relationship between the three types of data is established using the heat exchange station code (station_code) and user code (user_id) as the core association primary keys: Production operation data Customer service work order data: linked via heat exchange station code. If the work order already records the heat exchange station code, it is directly matched; if the work order only records the user code, it is matched using the "user" data constructed in step S2. residential area The heat exchange station code is obtained by querying the "heat exchange station" mapping table. In terms of time, the production data windows of 1 hour before and after the work order creation time (a total of 2-hour windows) are correlated.

[0087] Customer service ticket data Billing data: Precise matching is achieved through user codes. The two systems share unified user master data, and the user_id is globally unique, enabling precise one-to-one association.

[0088] Production operation data Billing data: Aggregated billing data (total amount receivable, total amount actually received, and payment rate) for heat exchange stations, and linked with the corresponding heat exchange station's operating energy consumption data on an annual basis.

[0089] Based on the above relationships, a multidimensional data association graph was constructed using the Neo4j graph database. The graph contains the following node types: heat exchange stations (82 nodes), communities (approximately 500), users (approximately 150,000 households), equipment (approximately 500 units), work orders (approximately 20,000 per year), payment records (approximately 150,000 per year), and measuring points (approximately 2,500). The edge type in the graph includes: BELONGS_TO (user...). LOCATED_IN (Community) Heat exchange station), INSTALLED_AT (equipment) Heat exchange station), RELATED_TO (work order) Heat exchange station, calculated using time window matching algorithm), PAYS (user) (Payment records). Data conflicts are resolved according to preset priorities: if the user address is inconsistent, the billing system shall prevail; if the user's heat exchange station is inconsistent, the production scheduling system shall prevail.

[0090] Step 6: Actual test of linking customer service work orders with production data.

[0091] The complete processing flow of the method of the present invention is demonstrated using a real complaint case as an example: At 14:35 on January 15, 2026, Mr. Wang, a resident of Unit 2, Building 15, Green Garden Community, called the customer service hotline to report that "the heating in my home is not working, and the room temperature is only 17.5℃." The customer service staff entered the work order information (user code USER1002345678, request type "repair report", creation time 2026-01-15 14:35:00), and after saving, the system automatically triggered the linked processing flow of steps S3 to S5: 1. Spatial Location: Using user ID USER1002345678 as the query key, query the Redis space ownership mapping table and return the result: User Green Garden Community Heat exchange station HZ037.

[0092] 2. Parameter Extraction: Using HZ037 as the spatial constraint and 2026-01-15 14:35 as the reference time, ... Using a 1-hour time window, the current operating parameters of station HZ037 from 13:35 to 15:35 on January 15, 2026, were extracted from the DWD layer standardized data. Simultaneously, the parameters from 13:35 to 15:35 each day for the previous 7 days (January 8 to 14, 2026) were extracted, and the historical average values ​​of each parameter were calculated as a reference.

[0093] 3. Deviation Calculation: The comparison results between the current parameters at the time of complaint (14:35) and the historical reference average are shown in Table 1: Table 1

[0094] 4. Rule Matching: The deviation indicator matches the fault judgment rule set. Specifically, the matched rule includes: rule "Secondary network water supply temperature deviation". And the temperature difference between supply and return water C (pointing to "insufficient heating at the heat exchange station", weight 0.30), rule "opening deviation of the primary network electric regulating valve" (This refers to "abnormal opening of primary network valves", weight 0.35). Both refer to "abnormal opening of primary network valves". The fault chain of "insufficient heating at the heat exchange station" has a comprehensive confidence score of 0.30 + 0.35 = 65%. Meanwhile, the rule "primary network water supply temperature deviation..." "Missed (actual deviation -5.2%), eliminating 'insufficient heat source output' as a candidate cause."

[0095] 5. Report Generation and Push: Generate a fault analysis report with the following content: "Fault Type: Abnormal opening of primary network valves (confidence level 65%), accompanied by overall low secondary network heating parameters; Recommended Measures: Remotely check the opening of the electric regulating valve of the primary network at station HZ037. If the abnormality is confirmed, adjust it to the normal opening of 65% and observe the recovery of the water supply temperature." The report is automatically attached to the work order and pushed to the customer service terminal and the dispatch center's large screen.

[0096] 6. Handling and Closure: After seeing the alarm on the large screen, the dispatcher remotely checked and found that the electric regulating valve of the primary network at station HZ037 was showing an opening of 35% (normally it should be 65%), indicating a valve actuator malfunction. The dispatcher remotely sent a control command to adjust the valve opening to 65%. 40 minutes later (15:20), the secondary network water supply temperature rose to 49℃, close to the normal level. Customer service personnel followed up with the user, who reported that the room temperature had risen to 19.5℃, and the work order was closed. The actual fault was attributed to "valve actuator malfunction" and entered into the system, serving as a sample to be sent back to the data platform for rule optimization.

[0097] It should be understood that, in comparison, the traditional method requires customer service personnel to log into the production SCADA system to query the operating data of station HZ037 (approximately 3 minutes), log into the billing system to check the user's payment status (approximately 2 minutes), and review historical work orders within the customer service system (approximately 2 minutes), followed by manual comparison and judgment (approximately 8 minutes), taking approximately 15 minutes to complete the initial location. The method of this invention automatically completes all analysis after the work order is created, outputting a fault analysis report within 30 seconds, improving the initial location efficiency by approximately 96.7% (from 15 minutes to 30 seconds). The entire process from complaint to restoration of heating, under the traditional method, averages 4-6 hours (requiring the initial complaint handling process). Send a repairman On-site investigation Return to warehouse to retrieve spare parts replace The complete process of "restoration" is described in this invention. By remotely diagnosing and adjusting the valve opening, the processing time is reduced to 1.5 hours. The remote diagnosis process saves about 2 hours of on-site troubleshooting by maintenance workers, improving efficiency by about 70%.

[0098] Example 3 To achieve the above method, such as Figures 2 to 6 As shown, this invention further integrates the above methods into a complete converged management and control platform at the system level, specifically: like Figure 2 As shown, the overall platform architecture of this invention adopts a four-layer architecture of "cloud-pipe-edge-terminal". From bottom to top, the layers are as follows: The terminal layer (sensing and execution layer) consists of the heat exchange station PLC controller, temperature sensor, pressure sensor, flow meter, heat meter, electric regulating valve and frequency converter, etc., responsible for real-time acquisition of on-site operation data and equipment control execution; The edge layer (edge ​​computing layer) consists of the heat exchange station local control cabinet and edge computing gateway, undertaking the functions of closed-loop automatic control, data preprocessing and network interruption resume transmission within the station; The pipe layer (data transmission layer) consists of industrial switch, 4G / 5G wireless communication module and VPN security gateway, ensuring real-time, secure and reliable data transmission between the heat exchange station and the central platform; The cloud layer (platform application layer) consists of unified data middleware, comprehensive management and control platform, mobile terminal application and large screen visualization system, realizing data aggregation and storage, intelligent analysis and decision-making, operation scheduling and customer service closed-loop management.

[0099] like Figure 3 As shown, the entire data processing workflow of the data platform is divided into six layers from bottom to top: (i) Data access layer: Kafka (real-time message queue, carrying SCADA time-series data), Canal (database binlog parsing, near real-time incremental synchronization of customer service work order data), DataX (batch data synchronization tool, T+1 offline synchronization of billing data) and Flink CDC (real-time change data capture) are used to realize the access of multi-source heterogeneous data.

[0100] (ii) Data storage layer: Hadoop HDFS (distributed file storage, storing raw data and historical details), Apache Hive (Hadoop-based offline data warehouse, ODS / DWD / DWS / ADS layered), TDengine (time series database, storing SCADA real-time running data), Apache Doris (MPP analytical database, report query and ad-hoc analysis), Neo4j (graph database, multidimensional data association graph) and Redis (caching, spatial mapping table and hot data cache).

[0101] (III) Data computation layer: Apache Flink (real-time computation, real-time data cleaning and real-time alerting) and Apache Spark (offline computation, batch ETL and model training) are used.

[0102] (iv) Data Governance Layer: This layer covers metadata management, data quality management (five-dimensional verification: integrity, validity, consistency, accuracy, and timeliness) and data security management (RBAC-based permission model, HTTPS / TLS transmission encryption, HDFS transparent encryption, AES-256 encrypted storage of sensitive fields, data anonymization, and retention of audit logs for no less than 180 days).

[0103] (v) Data Service Layer: Provides RESTful API (unified API gateway, OAuth2.0 Token + API Key two-factor authentication), SQL query interface, and real-time WebSocket push (data latency) (5 seconds).

[0104] (vi) Upper application layer: comprehensive management and control platform (including five modules: operation monitoring, intelligent scheduling, customer service, business analysis and system management), mobile APP and third-party systems.

[0105] like Figure 4 As shown, the integrated management and control platform of this invention includes five primary functional modules: The operation monitoring module displays the real-time operating status of all heat exchange stations using a Geographic Information System (GIS) map, including key parameters such as supply water temperature, return water temperature, supply water pressure, return water pressure, instantaneous flow rate, and cumulative heat. It employs WebSocket full-duplex communication for real-time data push with minimal data latency. 5 seconds; Configure over-limit alarm judgment rules to achieve hierarchical alarm push and closed-loop processing.

[0106] The intelligent scheduling module predicts heating load based on the XGBoost regression model. It takes 15 features, including outdoor temperature, wind speed, solar radiation intensity, and historical heat load, as input and outputs hourly heat load predictions for the next 24 hours. On this basis, with the objective function of minimizing the overall operating cost, and under the conditions of meeting the upper and lower limits of water supply temperature, the rate of temperature change, flow constraints, room temperature constraints, and heat source output constraints, it uses sequential quadratic programming or genetic algorithms to solve for the hourly optimal water supply temperature setpoint curve for the next 24 hours.

[0107] Customer Service Module: This module links customer service work orders with production data to implement the automatic fault location method described in claim 1. When a user initiates a repair request, the system automatically retrieves historical operating data from the corresponding heat exchange station, calculates the deviation between the operating parameters during the complaint period and the historical average for the same period, matches the fault determination rules, and outputs a fault analysis report to assist customer service personnel in quickly locating the fault.

[0108] Operational Analysis Module: Integrates billing data and heating cost data to generate operational analysis reports and trend forecasts; uses the Prophet time series forecasting model to predict operational indicator trends.

[0109] Large-screen visualization module: Through a large data visualization screen, the panoramic data of heating operation is displayed in multiple layers in the form of charts, heat maps, dashboards and other formats.

[0110] like Figure 5 As shown, the core entity relationship of the multidimensional data association graph is: the central node is the heat station, and its association relationships include: heat station (LOCATED_IN) — Community (BELONGS_TO) — User; User — (CREATES) Work Order (RELATED_TO, time window matching) Heat exchange station; User — (PAYS) Payment record; Heat exchange station — (INSTALLED_AT) Devices Meter Point Runtime sequence data; heat exchange stations — (SERVES) Users (heating service relationships). The correlation between the three types of data is reflected in: production operation data (measuring points) equipment Heat exchange station), customer service data (work orders) user Heat exchange station), billing data (payment) user The three entities (heat exchange station, user, and heat exchange station) are linked through two core entities: the heat exchange station and the user.

[0111] like Figure 6 As shown, the process for linking customer service work orders with production data (automatic fault location) is as follows: User complaint Customer service creates a work order (enters user information / request type). The system automatically triggers associated location. Query the heat exchange station to which the user belongs Retrieve the station's operational data for the past 24 hours. Calculate parameter deviation Matching fault determination rules Community-based complaint aggregation Output a fault analysis report (fault cause + confidence level + recommended measures). Automatic work order dispatch (remote troubleshooting / on-site repair handling) Real-time feedback of processing Processing complete User feedback Work order closed loop Data is written back to the data platform (for model optimization and statistical analysis). If the system determines that the problem is at the user end (the heat exchange station is operating normally), a work order is directly dispatched to the user end maintenance team; if the problem is determined to be a heat exchange station malfunction, an alarm is simultaneously pushed to the dispatch center's large screen.

[0112] In this embodiment, the data flow and dependencies between modules are as follows: the spatial mapping module outputs the constructed spatial attribution mapping table to the parameter extraction module; the data access module writes standardized data into the detailed data layer of the data platform for the parameter extraction module to query; the parameter extraction module outputs the extracted current operating parameters and historical reference operating parameters to the fault determination module; the fault determination module outputs the generated confidence scores of each candidate fault cause to the linkage output module; the linkage output module pushes the generated fault analysis report to the customer service terminal and the dispatch terminal, and sends the processed actual fault attribution data back to the sample database of the data platform. The above data flow relationships constitute the "spatial mapping construction..." Data access and standardization Work order trigger extraction Deviation Judgment Scoring A complete data pipeline of "report push and closed-loop iteration".

[0113] Example 4 To implement the methods of the above embodiments, the present invention also provides an electronic device, including a processor and a memory; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the various steps of the above-described intelligent heating multi-system integrated management and control method based on a data middle platform.

[0114] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0115] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for integrated management and control of multiple intelligent heating systems based on a data platform, characterized in that, include: Establish a spatial attribution mapping relationship between users and heat exchange stations; Heterogeneous data from the production scheduling system and customer service work order system are integrated into the data platform and standardized to obtain standardized data. When a customer service work order creation event is captured, the mapping relationship is queried according to the user code in the work order to determine the associated heat exchange station. The heat exchange station is used as a spatial limit and the preset time window is extended based on the work order creation time. The current operating parameters and historical reference operating parameters of the heat exchange station are extracted from the standardized data. The extracted current operating parameters are compared with the historical reference operating parameters to obtain the deviation index. The deviation index is then input into the preset fault judgment rule set for matching, and the confidence score corresponding to each candidate fault cause is output. Based on the output confidence scores, the causes of failures that meet the preset conditions are selected, a failure analysis report associated with the corresponding work order is generated, and the failure analysis report is pushed to the business terminal.

2. The method as described in claim 1, characterized in that, Constructing a spatial affiliation mapping relationship between users and heat exchange stations, including: Obtain heat exchange station data from the production scheduling system and user profile data from the customer service work order system; A spatial attribution mapping table containing the hierarchical attribution relationship between users, communities, and heat exchange stations is established, with the heat exchange station code in the production scheduling system as the first primary key and the user code in the customer service work order system as the second primary key. The spatial affiliation mapping table is stored in the cache layer of the data platform.

3. The method as described in claim 1, characterized in that, Integrate heterogeneous data from the production scheduling system and customer service work order system into the data platform, including: The time-series operational data generated by the production scheduling system is streamed into the data platform via a real-time message channel. For transaction change data generated by the customer service ticket system, the data platform is accessed incrementally through the database log parsing channel; The accessed time-series running data and transaction change data are written into the source data layer of the data platform, preserving the original data format.

4. The method as described in claim 1, characterized in that, Standardization of heterogeneous data includes: The timestamps from different data sources are aligned according to a preset uniform time granularity. The alignment formula is as follows: The division uses integer division, the original time is in Unix timestamp format, and the aligned time is in the same unit as the original time, generating a time series data sequence with a unified time base; Based on a pre-set encoding mapping dictionary, the heterogeneous encoding fields in each system are converted into a unified standard encoding. After time alignment and encoding conversion, the data is stored in the detailed data layer of the data platform.

5. The method as described in claim 1, characterized in that, Standardization of heterogeneous data includes: For short-term missing data, linear interpolation is used to fill in the missing values. The filling formula is as follows: in, The interpolated value for the missing point. and These are the nearest valid values ​​before and after the missing point. The timestamp of the missing point. and These are the timestamps corresponding to the most recent valid values ​​before and after; Data sequences that are continuously missing for more than a preset duration are marked as missing and will not be included in subsequent deviation index calculations.

6. The method as described in claim 1, characterized in that, Based on the work order creation time, a preset time window is extended, and the current operating parameters corresponding to the heat exchange station are extracted from the standardized data, including: Using the creation time of the target work order as the base time, a sliding time window is formed by extending forward by a first preset time and backward by a second preset time. Using the associated heat exchange station code as a spatial filtering condition, the operating value sequence of each parameter type of the heat exchange station within the sliding time window is extracted from the standardized data and used as the current operating parameter.

7. The method as described in claim 1, characterized in that, The historical reference operating parameters for the heat exchange station corresponding to the same period are extracted from the standardized data, including: Based on the creation time of the target work order, a historical time interval is determined; wherein, the historical time interval is a continuous period of time within the previous 7 natural days that is the same as the creation time, and the duration of the continuous period is the same as the duration of the preset time window. Using the associated heat exchange station code as a spatial filtering condition, the operating value sequence of each parameter type of the heat exchange station within the historical time interval is extracted from the standardized data, and the arithmetic mean of the same parameter type within the historical time interval is calculated as the historical reference operating parameter.

8. The method as described in claim 1, characterized in that, The deviation index is input into a preset fault determination rule set for matching, and the confidence score corresponding to each candidate fault cause is output. A fault analysis report is then generated based on each confidence score, including: The deviation index is matched step by step with multiple fault determination rules in the fault determination rule set. Each fault determination rule corresponds to a candidate fault cause and has a preset weight coefficient. The weight coefficients corresponding to the cumulative hits of each fault determination rule are used to calculate the comprehensive confidence score of each candidate fault cause. Candidate fault causes with a comprehensive confidence score higher than a preset threshold are selected, and a fault analysis report is generated that includes fault type labels, confidence percentage values, and handling suggestions.

9. A smart heating multi-system integrated management and control platform based on a data middleware platform, characterized in that, include: The spatial mapping module is used to construct the spatial affiliation mapping relationship between users and heat exchange stations; The data access module is used to connect heterogeneous data from the production scheduling system and the customer service work order system to the data platform and perform standardized processing to obtain standardized data. The parameter extraction module is used to determine the associated heat exchange station by querying the mapping relationship based on the user code in the work order when a customer service work order creation event is captured. It also uses the heat exchange station as a spatial limit and the work order creation time as a benchmark to expand a preset time window to extract the current operating parameters and historical reference operating parameters of the heat exchange station from the standardized data. The fault determination module is used to compare the extracted current operating parameters with the historical reference operating parameters to obtain the deviation index, and input the deviation index into the preset fault determination rule set for matching, and output the confidence score corresponding to each candidate fault cause. The linkage output module is used to filter out the causes of failures that meet preset conditions based on the confidence scores of the output, generate a fault analysis report associated with the corresponding work order, and push the fault analysis report to the business terminal.

10. An electronic device, characterized in that, include: processor; The memory stores executable instructions; when the processor executes the instructions, it implements the intelligent heating multi-system fusion management and control method based on a data middle platform as described in any one of claims 1-8.