An air conditioner operation state detection system

Abstract

The application discloses an air conditioner operation state detection system which is composed of an internet of things system, a data access platform, a service platform, a model archiving server and a business platform. Device data generated by air conditioner operation is reported to the data access platform, the data access platform generates a prediction request and pushes the prediction request to the service platform in a concurrent mode, the service platform accesses the prediction request of the platform in parallel, reads a prediction model from the model archiving server, predicts the air conditioner operation state in combination with the device data, and returns the prediction result to the data access platform, and the data access platform reports the prediction result to the business platform. The service platform also receives a training task of a new model from the business platform, trains the training task in an asynchronous multi-process mode, and stores the trained prediction model in the model archiving server. The effect of real-time receiving of device reported data, online prediction, online training of a prediction model, and real-time updating of a model is realized.

Description

Technical Field

[0001] This invention relates to the field of central air conditioning technology, and more specifically to an air conditioning operation status detection system. Background Technology

[0002] After-sales service for air conditioners is primarily based on customer reports of malfunctions. This approach relies on the customer's initiative to initiate service, with the service department responding passively. However, customers often lack technical expertise and can only report symptoms of the malfunction. This makes it difficult for service personnel to accurately pinpoint the location and type of the problem, requiring them to visit the customer's home for inspection. If the problem cannot be resolved on-site, a second visit is necessary. This cumbersome and inefficient process diminishes the customer experience and reduces the effectiveness of after-sales service.

[0003] In order to predict air conditioner malfunctions in advance and locate the type and location of the malfunction, several air conditioner malfunction reporting methods have been proposed in the prior art. These methods collect data on the operation of the air conditioner and predict and diagnose the operating status of the air conditioner based on predictive models or diagnostic algorithms. On the one hand, they can detect malfunctions in the early stage and proactively initiate maintenance procedures. On the other hand, they can locate the type and location of the malfunction that has occurred or is about to occur. This can prevent air conditioner malfunctions from accumulating until they become unusable before being discovered, thus reducing the user experience. It also helps maintenance personnel to perform targeted and well-prepared maintenance on malfunctions.

[0004] However, this fault reporting method requires the collection of a large amount of operational data and the support of multiple prediction or diagnostic models, which places higher demands on the fault prediction and diagnosis system.

[0005] The information disclosed in this background section is only intended to enhance the understanding of the background technology of this application, and therefore may include prior art that is not known to those skilled in the art. Summary of the Invention

[0006] To address the problems mentioned in the background art, this invention proposes an air conditioning operation status detection system. It constructs a detection system consisting of an Internet of Things system, a data access platform, a service platform, a model archiving server, and a business platform. It provides comprehensive system support for current fault diagnosis algorithms and achieves the effects of real-time reception of device-reported data, online prediction, online optimization of prediction models, and timely release of new models.

[0007] To achieve the above-mentioned objectives, the present invention employs the following technical solution:

[0008] An air conditioning operation status detection system is proposed, comprising:

[0009] The Internet of Things (IoT) system is used to access the air conditioning gateway and receive device data reported by the air conditioners.

[0010] The data access platform receives device data reported by air conditioners from the IoT system, generates prediction requests based on the device data and pushes them to the service platform; it receives prediction results from the service platform and reports them to the business platform.

[0011] The service platform receives concurrent prediction requests from the data access platform through a prediction interface, reads the prediction model to predict the air conditioner's operating status, and returns the prediction results to the data access platform; it also receives training tasks from the business platform through a training interface and trains the prediction model based on asynchronous multi-process.

[0012] A model archiving server, connected to the service platform, is used to store and archive prediction data and trained prediction models;

[0013] The business platform is used to obtain prediction results from the data access platform, implement target management based on the prediction results, and issue prediction model training instructions to the service platform.

[0014] Compared with existing technologies, the advantages and positive effects of this invention are as follows: The air conditioning operation status detection system proposed in this invention constructs a detection system composed of an Internet of Things (IoT) system, a data access platform, a service platform, a model archiving server, and a business platform. Equipment data generated by air conditioning operation is reported to the data access platform via the IoT system. The data access platform generates prediction requests based on the equipment data and pushes them to the service platform concurrently. The service platform receives the concurrent prediction requests from the data access platform through a prediction interface, reads the prediction model from the model archiving server, predicts the air conditioning operation status based on the equipment data, and returns the prediction results to the data access platform. The data access platform then reports the prediction results to the business platform, which implements target management such as alarms and maintenance process generation based on the prediction results. The service platform also receives training tasks for new models from the business platform through a training interface, trains the training tasks using an asynchronous multi-process approach, and stores the trained prediction models on the model archiving server. The air conditioning operation status detection system proposed in this invention provides comprehensive system support for current fault diagnosis algorithms, achieving the effects of real-time reception of equipment-reported data, online prediction, online training of prediction models, and timely updates of new models.

[0015] In some embodiments of the present invention, the prediction interface includes: a real-time prediction interface for receiving concurrent prediction requests; and a prediction model acquisition interface for loading a prediction model from the model archive server.

[0016] For online real-time prediction, it is necessary to support a large number of concurrent requests and return prediction results as soon as possible. This detection system is designed with a prediction interface that can receive concurrent requests, receive several prediction requests from the data access platform at the same time, and return prediction results in a timely manner.

[0017] In some embodiments of the present invention, the system further includes: a caching platform connected to the data access platform for caching matching data; wherein the data access platform identifies and matches device data to obtain the matching data, and generates the prediction request based on the matching data.

[0018] In real-world scenarios, air conditioning systems are divided into indoor and outdoor units. When reporting equipment data, the data is reported separately for indoor and outdoor units. The data access platform identifies and matches the data from each device to obtain complete matching data for an air conditioning system. The matching data is then cached on the caching platform, and a prediction request is generated based on the matching data and reported to the service platform.

[0019] In some embodiments of the present invention, the service platform publishes the trained prediction model to the cache platform based on the new model publishing instruction issued by the business platform. Before reading the prediction model, the service platform determines whether the cache platform has a newly published prediction model. If so, it reads the newly published prediction model from the cache platform.

[0020] In addition to being stored in the model archive server for the service platform to access in real time, the newly trained prediction model is also used to improve the speed at which the service platform can access the new prediction model. After receiving feedback on the generation of the new prediction model, the business platform sends a new model release command to the service platform. The service platform then distributes the newly trained prediction model to the buffer platform, allowing the prediction process to quickly retrieve the latest model from the buffer platform, thereby improving the system's prediction capabilities.

[0021] In some embodiments of the present invention, the service platform reserves a prediction result buffer, trains the prediction model multiple times, caches the prediction results each time, and returns the prediction result when multiple prediction results are consistent.

[0022] To smooth prediction results, reduce fluctuations, or prevent outliers from compromising their stability, a prediction result buffer is designed for the service platform. The platform performs multiple predictions and reserves these multiple results in the buffer. Only when the multiple predictions are consistent is the final prediction returned, thus improving prediction accuracy. The prediction result buffer can be implemented in memory to mitigate its impact on prediction speed.

[0023] In some embodiments of the present invention, the system further includes a database for storing prediction data and prediction results, from which the business platform reads the prediction results.

[0024] Historical data is stored in a database so that the business platform can access, query, or apply it at any time.

[0025] In some embodiments of the present invention, the data access platform accesses the Internet of Things system through MQ, MQTT protocols or API interfaces.

[0026] In some embodiments of the present invention, the storage and archiving of prediction data and prediction results are implemented using MySQL, PostgreSQL, MongoDB, or ElasticSearch.

[0027] In some embodiments of the present invention, the pressure, gas, and / or temperature data in the device data are implemented by calling the CoolProp library.

[0028] In some embodiments of the present invention, the service platform uses Phthon's ASGI framework to implement concurrent access to prediction requests, and uses Phthon's Celery framework to implement asynchronous multi-process training tasks.

[0029] Other features and advantages of the present invention will become clearer after reading the detailed embodiments of the invention in conjunction with the accompanying drawings. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 This is a schematic diagram of the architecture of the air conditioning operation status detection system proposed in this invention;

[0032] Figure 2 This is a schematic diagram of the prediction process of the air conditioning operation status prediction system proposed in this invention;

[0033] Figure 3 This is a schematic diagram of the architecture of the air conditioning operation status detection system given in Embodiment 1 of the present invention;

[0034] Figure 4 This is a schematic diagram of the architecture of the air conditioning operation status detection system given in Embodiment 2 of the present invention;

[0035] Figure 5 This is a schematic diagram of the architecture of the air conditioning operation status detection system given in Embodiment 3 of the present invention;

[0036] Figure 6This is a schematic diagram of the prediction process of the air conditioning operation status prediction system given in Embodiment 3 of the present invention;

[0037] Figure 7 This is a schematic diagram of the architecture of the air conditioning operation status detection system given in Embodiment 4 of the present invention;

[0038] Figure 8 This is a schematic diagram of the architecture of the air conditioning operation status detection system given in Embodiment 5 of the present invention;

[0039] Figure 9 This is a schematic diagram of the architecture of the air conditioning operation status detection system given in Embodiment Six of the present invention;

[0040] Figure 10 This is a schematic diagram of the architecture of the air conditioning operation status detection system given in Embodiment 7 of the present invention;

[0041] Figure 11 This is a schematic diagram of the prediction process of the air conditioning operation status prediction system given in Embodiment Six of the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0043] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0044] 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. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0045] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0046] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0047] The following disclosure provides many different embodiments or examples for implementing various structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the invention. Furthermore, reference numerals and / or letters may be repeated in different examples; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed. In addition, examples of various specific processes and materials are provided in this invention, but those skilled in the art will recognize the application of other processes and / or the use of other materials.

[0048] In this application, the air conditioner performs a refrigeration cycle by using a compressor, condenser, expansion valve, and evaporator. The refrigeration cycle includes a series of processes involving compression, condensation, expansion, and evaporation to cool or heat an indoor space.

[0049] Low-temperature, low-pressure refrigerant enters the compressor, which compresses it into a high-temperature, high-pressure refrigerant gas and discharges the compressed refrigerant gas. The discharged refrigerant gas flows into the condenser. The condenser condenses the compressed refrigerant into a liquid phase, and the heat is released to the surrounding environment through the condensation process.

[0050] The expansion valve expands the high-temperature, high-pressure liquid refrigerant that condenses in the condenser into a low-pressure liquid refrigerant. The evaporator evaporates the expanded refrigerant in the expansion valve and returns the low-temperature, low-pressure refrigerant gas to the compressor. The evaporator achieves its cooling effect by utilizing the latent heat of refrigerant evaporation to exchange heat with the material being cooled. Throughout the cycle, the air conditioner regulates the temperature of the indoor space.

[0051] The outdoor unit of an air conditioner refers to the part of the refrigeration cycle that includes the compressor and the outdoor heat exchanger. The indoor unit of an air conditioner includes the indoor heat exchanger, and an expansion valve can be provided in either the indoor or outdoor unit.

[0052] The indoor and outdoor heat exchangers function as either condensers or evaporators. When the indoor heat exchanger is used as a condenser, the air conditioner functions as a heater in heating mode; when the indoor heat exchanger is used as an evaporator, the air conditioner functions as a cooler in cooling mode.

[0053] This invention proposes an air conditioning operation status detection system for fault diagnosis of central air conditioning systems, to support the increasing demands on the system as diagnostic algorithms are continuously upgraded, such as... Figure 1 As shown, the detection system includes:

[0054] IoT system 1 is used to access the air conditioning gateway and receive device data reported by the air conditioner.

[0055] Data access platform 2 receives device data reported by air conditioners from IoT system 1, generates prediction requests based on the device data and pushes them to service platform 3; it receives prediction results from service platform 3 and reports them to business platform; the prediction request includes at least device data.

[0056] Service platform 3 receives concurrent prediction requests from data access platform 2 through the prediction interface, reads the prediction model to predict the air conditioner's operating status, and returns the prediction results to data access platform 2; it also receives training tasks from business platform 5 through the training interface and trains the prediction model based on asynchronous multi-process.

[0057] Model archiving server 4, connected to service platform 3, is used to store and archive prediction data and trained prediction models.

[0058] Business platform 5 is used to issue prediction model training instructions to service platform 3 and to obtain prediction results from data access platform 2, and to implement target management based on the prediction results.

[0059] The air conditioning operation status detection system proposed in this invention constructs a detection system consisting of an Internet of Things (IoT) system 1, a data access platform 2, a service platform 3, a model archiving server 4, and a business platform 5. Equipment data generated during air conditioning operation is reported to the data access platform 2 based on the IoT system 1. Figure 2 As shown, data access platform 2 generates prediction requests based on device data and pushes them to service platform 3 concurrently. Service platform 3 receives the concurrent prediction requests from data access platform 2 through the prediction interface, reads the prediction model from model archiving server 4, predicts the air conditioner's operating status based on the device data, and returns the prediction results to data access platform 2. Data access platform 2 then reports the prediction results to business platform 5, which implements target management such as alarms and maintenance process generation based on the prediction results. Service platform 3 also receives training tasks for new models from business platform 5 through the training interface, trains the training tasks using an asynchronous multi-process approach, and stores the trained prediction model in model archiving server 4. The air conditioner operating status detection system proposed in this invention provides comprehensive system support for current fault diagnosis algorithms, achieving the effects of real-time reception of device-reported data, online prediction, online training of prediction models, and timely updates of new models.

[0060] The air conditioning operation status detection system proposed in this invention will be described in detail below with several specific embodiments. Example 1

[0061] like Figure 3 As shown in the figure, the air conditioning operation status detection system provided in this embodiment includes:

[0062] IoT system 1 is used to access the air conditioning gateway and receive device data reported by the air conditioner.

[0063] Data access platform 2 receives device data reported by air conditioners from IoT system 1, generates prediction requests based on the device data and pushes them to service platform 3; it receives prediction results from service platform 3 and reports them to business platform; the prediction request includes at least device data.

[0064] Service platform 3 receives concurrent prediction requests from data access platform 2 through the prediction interface, reads the prediction model to predict the air conditioner's operating status, and returns the prediction results to data access platform 2; it also receives training tasks from business platform 5 through the training interface and trains the prediction model based on asynchronous multi-process.

[0065] Model archiving server 4, connected to service platform 3, is used to store and archive prediction data and trained prediction models.

[0066] Business platform 5 is used to issue prediction model training instructions to service platform 3 and to obtain prediction results from data access platform 2, and to implement target management based on the prediction results.

[0067] The prediction interface includes a real-time prediction interface and a prediction model acquisition interface. The real-time prediction interface is used to receive prediction requests concurrently from the data access platform 2, while the prediction model acquisition interface is used to load prediction models from the model archive server 4.

[0068] The equipment data generated by the air conditioner operation is reported to the data access platform 2 via the IoT system 1. The data access platform 2 generates prediction requests based on the equipment data and pushes them to the service platform 3 concurrently. The service platform 3 receives the concurrent prediction requests from the data access platform 2 through the prediction interface, reads the prediction model from the model archive server 4, predicts the air conditioner's operating status based on the equipment data, and returns the prediction results to the data access platform 2. The data access platform 2 then reports the prediction results to the business platform 5, which implements target management such as alarms and maintenance process generation based on the prediction results. The service platform 3 also receives training tasks for new models from the business platform 5 through the training interface, trains the training tasks using an asynchronous multi-process approach, and stores the trained prediction model in the model archive server 4. For online real-time prediction, it is necessary to support a large number of concurrent requests and return prediction results as quickly as possible. In this embodiment, a prediction interface capable of receiving concurrent requests is designed for the detection system, allowing it to receive several prediction requests from the data access platform 2 concurrently and return prediction results in a timely manner. Example 2

[0069] like Figure 4 As shown in the figure, the air conditioning operation status detection system provided in this embodiment includes:

[0070] IoT system 1 is used to access the air conditioning gateway and receive device data reported by the air conditioner.

[0071] Data access platform 2 receives device data reported by air conditioners from IoT system 1, generates prediction requests based on the device data and pushes them to service platform 3; it receives prediction results from service platform 3 and reports them to business platform; the prediction request includes at least device data.

[0072] Service platform 3 receives concurrent prediction requests from data access platform 2 through the prediction interface, reads the prediction model to predict the air conditioner's operating status, and returns the prediction results to data access platform 2. It also receives training tasks from business platform 5 through the training interface and trains the prediction model based on asynchronous multi-processing. The prediction interface includes a real-time prediction interface and a prediction model acquisition interface; the real-time prediction interface receives concurrent prediction requests from data access platform 2, while the prediction model acquisition interface loads the prediction model from model archiving server 4.

[0073] Model archiving server 4, connected to service platform 3, is used to store and archive prediction data and trained prediction models.

[0074] Business platform 5 is used to issue prediction model training instructions to service platform 3 and to obtain prediction results from data access platform 2, and to implement target management based on the prediction results.

[0075] The caching platform 6 is connected to the data access platform 2 and is used to cache matching data. The data access platform 2 identifies and matches device data to obtain matching data, and generates prediction requests based on the matching data.

[0076] The equipment data generated by the air conditioner operation, including outdoor unit data and indoor unit data, is reported to the data access platform 2 via the IoT system 1. The reported equipment data is reported separately for indoor and outdoor units. After receiving the equipment data from the IoT system 1, the data access platform 2 identifies and matches the data of each device to obtain complete matching data for an air conditioning system. The matching data is cached in the caching platform 6. Based on the matching data, a prediction request is generated and pushed to the service platform 3 in a concurrent manner. The service platform 3 receives the concurrent prediction requests from the data access platform 2 through the prediction interface, reads the prediction model from the model archiving server 4, combines it with the equipment data to predict the air conditioner's operating status, and returns the prediction result to the data access platform 2. The data access platform 2 then reports the prediction result to the business platform 5, which implements target management such as alarms and maintenance process generation based on the prediction result. At the same time, the service platform 3 also receives training tasks for new models from the business platform 5 through the training interface, trains the training tasks in an asynchronous multi-process manner, and stores the trained prediction model in the model archiving server 4. Example 3

[0077] like Figure 5 As shown, the air conditioning operation status detection system provided in this embodiment includes:

[0078] IoT system 1 is used to access the air conditioning gateway and receive device data reported by the air conditioner.

[0079] Data access platform 2 receives device data reported by air conditioners from IoT system 1, generates prediction requests based on the device data and pushes them to service platform 3; it receives prediction results from service platform 3 and reports them to business platform; the prediction request includes at least device data.

[0080] Service platform 3 receives concurrent prediction requests from data access platform 2 through the prediction interface, reads the prediction model to predict the air conditioner's operating status, and returns the prediction results to data access platform 2. It also receives training tasks from business platform 5 through the training interface and trains the prediction model based on asynchronous multi-processing. The prediction interface includes a real-time prediction interface and a prediction model acquisition interface; the real-time prediction interface receives concurrent prediction requests from data access platform 2, while the prediction model acquisition interface loads the prediction model from model archiving server 4.

[0081] Model archiving server 4, connected to service platform 3, is used to store and archive prediction data and trained prediction models.

[0082] Business platform 5 is used to issue prediction model training instructions to service platform 3 and to obtain prediction results from data access platform 2, and to implement target management based on the prediction results.

[0083] The caching platform 6 is connected to the data access platform 2 and is used to cache matching data. The data access platform 2 identifies and matches device data to obtain matching data, and generates prediction requests based on the matching data.

[0084] As described above, based on the new model release instruction issued by the business platform 5, the service platform 3 releases the trained prediction model to the cache platform 6. Before reading the prediction model, the service platform 3 first determines whether the cache platform 6 has a newly released prediction model. If so, it reads the newly released prediction model from the cache platform 6.

[0085] The equipment data generated by the air conditioner operation, including outdoor unit data and indoor unit data, is reported to the data access platform 2 based on the Internet of Things system 1. The reported equipment data is reported separately for indoor and outdoor units. After receiving the equipment data from the Internet of Things system 1, the data access platform 2 identifies and matches the data of each device to obtain complete matching data for an air conditioning system. The matching data is cached in the caching platform 6. Based on the matching data, a prediction request is generated and pushed to the service platform 3 in a concurrent manner. The service platform 3 receives the concurrent prediction requests from the data access platform 2 through the prediction interface, reads the prediction model from the model archiving server 4, combines it with the equipment data to predict the air conditioner's operating status, and returns the prediction result to the data access platform 2. The data access platform 2 then reports the prediction result to the business platform 5, which implements target management such as alarms and maintenance process generation based on the prediction result.

[0086] like Figure 6As shown, service platform 3 also receives training tasks for new models from business platform 5 through the training interface, trains the tasks using an asynchronous multi-process approach, stores the trained prediction model in model archive server 4, and then sends a message to business platform 5 indicating that the training of the new prediction model is complete. In addition to being stored in model archive server 4 for real-time access by service platform 3, to improve the speed at which service platform 3 can access the new prediction model, business platform 5, upon receiving feedback from the generated new prediction model, issues a new model release command to service platform 3. Service platform 3 then distributes the newly trained prediction model to buffer platform 6, allowing the prediction process to quickly retrieve the latest model from buffer platform 6, thereby improving the system's prediction capabilities. Example 4

[0087] The air conditioning operation status detection system provided in this embodiment includes:

[0088] IoT system 1 is used to access the air conditioning gateway and receive device data reported by the air conditioner.

[0089] Data access platform 2 receives device data reported by air conditioners from IoT system 1, generates prediction requests based on the device data and pushes them to service platform 3; it receives prediction results from service platform 3 and reports them to business platform; the prediction request includes at least device data.

[0090] Service platform 3 receives concurrent prediction requests from data access platform 2 through the prediction interface, reads the prediction model to predict the air conditioner's operating status, and returns the prediction results to data access platform 2. It also receives training tasks from business platform 5 through the training interface and trains the prediction model based on asynchronous multi-processing. The prediction interface includes a real-time prediction interface and a prediction model acquisition interface; the real-time prediction interface receives concurrent prediction requests from data access platform 2, while the prediction model acquisition interface loads the prediction model from model archiving server 4.

[0091] Model archiving server 4, connected to service platform 3, is used to store and archive prediction data and trained prediction models.

[0092] Business platform 5 is used to issue prediction model training instructions to service platform 3 and to obtain prediction results from data access platform 2, and to implement target management based on the prediction results.

[0093] The caching platform 6 is connected to the data access platform 2 and is used to cache matching data. The data access platform 2 identifies and matches device data to obtain matching data, and generates prediction requests based on the matching data.

[0094] The above, such as Figure 7As shown, service platform 3 reserves a buffer for prediction results, trains the prediction model multiple times, caches the prediction results each time, and returns the prediction result when multiple prediction results are consistent.

[0095] To smooth prediction results, reduce fluctuations, or prevent outliers from disrupting their stability, a prediction result buffer is designed for service platform 3. Service platform 3 performs multiple predictions and reserves multiple prediction results in the buffer. The final prediction result is only returned when all predictions are consistent, thus improving prediction accuracy. The prediction result buffer can be implemented in memory to mitigate its impact on prediction speed. Example 5

[0096] like Figure 8 As shown in the figure, the air conditioning operation status detection system provided in this embodiment includes:

[0097] IoT system 1 is used to access the air conditioning gateway and receive device data reported by the air conditioner.

[0098] Data access platform 2 receives device data reported by air conditioners from IoT system 1, generates prediction requests based on the device data and pushes them to service platform 3; it receives prediction results from service platform 3 and reports them to business platform; the prediction request includes at least device data.

[0099] Service platform 3 receives concurrent prediction requests from data access platform 2 through the prediction interface, reads the prediction model to predict the air conditioner's operating status, and returns the prediction results to data access platform 2. It also receives training tasks from business platform 5 through the training interface and trains the prediction model based on asynchronous multi-processing. The prediction interface includes a real-time prediction interface and a prediction model acquisition interface; the real-time prediction interface receives concurrent prediction requests from data access platform 2, while the prediction model acquisition interface loads the prediction model from model archiving server 4.

[0100] Model archiving server 4, connected to service platform 3, is used to store and archive prediction data and trained prediction models.

[0101] Business platform 5 is used to issue prediction model training instructions to service platform 3 and to obtain prediction results from data access platform 2, and to implement target management based on the prediction results.

[0102] The caching platform 6 is connected to the data access platform 2 and is used to cache matching data. The data access platform 2 identifies and matches device data to obtain matching data, and generates prediction requests based on the matching data.

[0103] Database 7 is used to store prediction results, and business platform 3 reads prediction results from it.

[0104] Historical data is stored in a database so that the business platform can access, query, or apply it at any time. Example 6

[0105] This implementation presents an air conditioning operation status monitoring system deployed as a single unit, which can be used to monitor the operation status of all air conditioning equipment in a user's room, including:

[0106] IoT Unit 1 is used to connect to the air conditioning gateway and receive device data reported by the air conditioner.

[0107] Data access unit 2 receives device data reported by the air conditioner from IoT unit 1, generates a prediction request based on the device data and pushes it to service unit 3; it receives the prediction result from service unit 3 and reports it to business unit; the prediction request includes at least device data.

[0108] Service Unit 3 receives concurrent prediction requests from Data Access Unit 2 via the prediction interface, reads the prediction model to predict the air conditioner's operating status, and returns the prediction results to Data Access Unit 2. It also receives training tasks from Business Unit 5 via the training interface and trains the prediction model based on asynchronous multi-processing. The prediction interface includes a real-time prediction interface and a prediction model acquisition interface; the real-time prediction interface receives concurrent prediction requests from Data Access Unit 2, while the prediction model acquisition interface loads the prediction model from Model Archiving Unit 4.

[0109] Model archiving unit 4 and connection service unit 3 are used to store and archive prediction data and trained prediction models.

[0110] Business Unit 5 is used to issue prediction model training instructions to Service Unit 3 and obtain prediction results from Data Access Unit 2, and to implement target management based on the prediction results.

[0111] The caching unit 6 is connected to the data access unit 2 and is used to cache matching data. The data access unit 2 identifies and matches device data to obtain matching data and generates a prediction request based on the matching data.

[0112] Service Unit 3 reserves a buffer for prediction results, trains the prediction model multiple times, caches the prediction results each time, and returns the prediction result when multiple prediction results are consistent.

[0113] like Figure 9As shown, this embodiment uses Phthon to design the operating program of the detection system. In the service of data access unit 2, the reported device data can be converted into matching data and cached in memory (caching unit 6). SQLite is used to store the prediction results. Service unit 3 uses FastAPI to implement the prediction interface and training interface functions. The information of the currently used prediction model (model ID, version, download address, training time, etc.) is cached in memory. A list is used in memory as a prediction result buffer. The prediction model can be obtained from the cloud, or a simple prediction model archiving service can be implemented using nginx and the lua-retry-upload module.

[0114] Celery can be used to implement an asynchronous multi-process training service, and Celery and the interface can be deployed on the same server.

[0115] The equipment data generated during air conditioning operation, including data from both the outdoor and indoor units, is reported to the data access unit 2 via IoT unit 1. The reported equipment data is submitted separately for indoor and outdoor units. After receiving the equipment data from IoT unit 1, data access unit 2 identifies and matches the data from each device to obtain complete matching data for the air conditioning system. This matching data is then cached in memory. Based on the matching data, prediction requests are generated and pushed concurrently to service unit 3. Service unit 3 receives these concurrent prediction requests from data access unit 2 through a prediction interface, reads the prediction model from the nginx and lua-retry-upload modules, and combines it with the equipment data according to a set sequence. The system predicts the air conditioner's operating status and uses a List in memory as a buffer for the prediction results. Each prediction result is recorded, and consistency is checked. If consistent, the prediction result is returned to data access unit 2; otherwise, training continues until all prediction results within a specified number of iterations are consistent. Data access unit 2 then reports the prediction results to business unit 5, which performs target management such as alarms and maintenance process generation based on the prediction results. Service unit 3 also receives new model training tasks from business unit 5 through a training interface, uses Celery to train the model, and stores the trained prediction model in the nginx and lua-retry-upload modules. Prediction data and results are stored in the sqlite module. Example 7

[0116] This embodiment presents an air conditioning operation status monitoring system deployed in a distributed manner. It can be used by enterprises to monitor the operation status of air conditioning equipment purchased by all users, achieving high availability in practical applications, including:

[0117] IoT system 1 is used to access the air conditioning gateway and receive device data reported by the air conditioner.

[0118] Data access platform 2 receives device data reported by air conditioners from IoT system 1, generates prediction requests based on the device data and pushes them to service platform 3; it receives prediction results from service platform 3 and reports them to business platform; the prediction request includes at least device data.

[0119] Service platform 3 receives concurrent prediction requests from data access platform 2 through the prediction interface, reads the prediction model to predict the air conditioner's operating status, and returns the prediction results to data access platform 2. It also receives training tasks from business platform 5 through the training interface and trains the prediction model based on asynchronous multi-processing. The prediction interface includes a real-time prediction interface and a prediction model acquisition interface; the real-time prediction interface receives concurrent prediction requests from data access platform 2, while the prediction model acquisition interface loads the prediction model from model archiving platform 4.

[0120] Model archiving platform 4 and connection service platform 3 are used to store and archive prediction data and trained prediction models.

[0121] Business platform 5 is used to issue prediction model training instructions to service platform 3 and to obtain prediction results from data access platform 2, and to implement target management based on the prediction results.

[0122] The caching platform 6 is connected to the data access platform 2 and is used to cache matching data. The data access platform 2 identifies and matches device data to obtain matching data, and generates prediction requests based on the matching data.

[0123] Service platform 3 reserves a buffer for prediction results, trains the prediction model multiple times, caches the prediction results each time, and returns the prediction result when multiple prediction results are consistent.

[0124] like Figure 10 As shown, in this embodiment, the data access platform 2 connects to the IoT system 1 via MQ, MQTT protocols, or API interfaces. It uses Redis to cache the matching data converted from reported device data. Predictive data and prediction results are stored using MySQL, PostgreSQL, MongoDB, and Elasticsearch. The service platform 3 is deployed on a high-performance server, using Redis to cache the currently used prediction model information and as a buffer for prediction results. When the service platform 3 starts, the latest prediction model is preloaded to increase the responsiveness of the prediction interface. Additionally, the training interface can be deployed separately to reduce its impact on the prediction interface.

[0125] Service platform 3 uses the Celery framework to implement asynchronous multi-process training services. The Celery server is deployed as a separate service, and the specific service configuration can be selected according to the number of training sessions. The trained model objects are exported as *.joblib files and uploaded to model archiving platform 4. Because it is an asynchronous task, after completing an assigned task, a callback to the training interface is needed to notify the model of the training results, such as... Figure 10 As shown.

[0126] Model archiving platform 4 can be implemented using file management servers such as Alibaba Cloud OSS to ensure the security and high availability of model files.

[0127] In practical applications, data access platform 2 also needs to process some characteristic parameters in the reported device data, including but not limited to pressure, gas, and temperature data, which can be achieved by calling the CoolProp library.

[0128] Service platform 3 can use Phthon's ASGI framework (such as starlette, fastapi, etc.) to achieve concurrent access to prediction requests.

[0129] like Figure 11 As shown, the equipment data generated by the air conditioner operation, including outdoor unit data and indoor unit data, is reported to the data access platform 2 via IoT platform 1. The reported equipment data is reported separately for indoor and outdoor units. After receiving the equipment data from IoT platform 1, data access platform 2 identifies and matches the data of each device to obtain complete matching data for the air conditioning system. This matching data is then cached in Redis. Based on the matching data, prediction requests are generated and pushed concurrently to service platform 3. Service platform 3 uses PowerPoint's ASGI framework (such as Starlette, FastAPI, etc.) to implement concurrent access to prediction requests, reading prediction data from file management servers such as Alibaba Cloud OSS. The model, combined with equipment data, predicts the air conditioner's operating status a set number of times. Redis is used to store each prediction result. The model checks if all prediction results are consistent. If they are, the prediction result is returned to the data access platform 2; otherwise, training continues until all prediction results within the specified number of attempts are consistent. The data access platform 2 then reports the prediction results to the business platform 5, which implements target management such as alarms and maintenance process generation based on the prediction results. The service platform 3 also receives training tasks for new models from the business platform 5 through a training interface, uses Celery to train the model, and exports the trained prediction model as a *.joblib file, uploading it to a file management server such as Alibaba Cloud OSS.

[0130] In the description of the above embodiments, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0131] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An air conditioning operation status detection system, characterized in that, include: The Internet of Things (IoT) system is used to access the air conditioning gateway and receive device data reported by the air conditioners. The data access platform receives device data reported by air conditioners from the Internet of Things system, and generates prediction requests based on the device data and pushes them to the service platform. Receive prediction results from the service platform and report them to the business platform; A caching platform, connected to the data access platform, is used to cache matching data; The data access platform identifies and matches device data to obtain the matching data, and generates the prediction request based on the matching data. The service platform receives concurrent prediction requests from the data access platform through a prediction interface, reads the prediction model to predict the air conditioner's operating status, and returns the prediction results to the data access platform. It also receives training tasks from the business platform through a training interface and trains the prediction model based on asynchronous multi-processing. Based on a new model release instruction issued by the business platform, the service platform publishes the trained new prediction model to the cache platform. Before reading the prediction model, the service platform determines whether the cache platform already has the newly published prediction model; if so, it reads the newly published prediction model from the cache platform. A model archiving server, connected to the service platform, is used to store and archive prediction data and trained prediction models; The business platform is used to obtain prediction results from the data access platform, implement target management based on the prediction results, and issue prediction model training instructions to the service platform. The prediction interface includes: The real-time prediction interface is used to receive concurrent prediction requests; A prediction model acquisition interface is used to load prediction models from the model archive server.

2. The air conditioning operation status detection system according to claim 1, characterized in that, The service platform reserves a buffer for prediction results, performs multiple predictions and caches the prediction results each time, and returns the prediction result when the multiple prediction results are consistent.

3. The air conditioning operation status detection system according to claim 1, characterized in that, The system also includes: A database is used to store prediction data and prediction results, from which the business platform reads the prediction results.

4. The air conditioning operation status detection system according to claim 1, characterized in that, The data access platform connects to the IoT system via MQ, MQTT protocols, or API interfaces.

5. The air conditioning operation status detection system according to claim 3, characterized in that, The storage and archiving of forecast data and forecast results are implemented using MySQL, PostgreSQL, MongoDB, or ElasticSearch.

6. The air conditioning operation status detection system according to claim 1, characterized in that, The pressure, gas, and / or temperature data in the device data are implemented using the CoolProp library.

7. The air conditioning operation status detection system according to claim 1, characterized in that, The service platform uses Phthon's ASGI framework to enable concurrent access to prediction requests, and Phthon's Celery framework to implement asynchronous multi-process training tasks.

Images