CN-122015409-A - Method and device for intelligently diagnosing refrigerant leakage, storage medium and refrigerator

Abstract

The invention provides a method, a device, a storage medium and a refrigerator for intelligently diagnosing refrigerant leakage, wherein the method is suitable for networking refrigeration equipment and comprises the following steps of collecting multidimensional serial port data in the operation process of the refrigerator; the method comprises the steps of carrying out validity check and encryption processing on collected data, uploading the data to a cloud big data platform, carrying out cleaning, missing value completion and abnormal value processing on the uploaded data at the cloud, extracting key features to construct a time sequence diagnosis segment, analyzing the time sequence diagnosis segment based on a machine learning decision tree model, judging whether a refrigerant leakage risk exists or not, generating a diagnosis result when judging that the refrigerant leakage risk exists, and pushing the diagnosis result to an operation and maintenance background and a user terminal. The invention can realize early detection and accurate early warning of refrigerant leakage on the premise of not adding any sensor.

Inventors

• Mu Zhoujie
• LIAO MIN
• LIU YUNTAO

Assignees

• 小米科技(武汉)有限公司
• 小米智能家电(武汉)有限公司
• 北京小米移动软件有限公司

Dates

Publication Date

20260512

Application Date

20260212

Claims (14)

1. 1. The method for intelligently diagnosing the refrigerant leakage is characterized by being suitable for networking refrigeration equipment and comprising the following steps of: collecting multidimensional serial port data in the running process of refrigeration equipment; The collected data is uploaded to a cloud big data platform after validity verification and encryption processing; Cleaning the uploaded data at the cloud end, complementing the missing value and processing the abnormal value; extracting key features to construct a time sequence diagnosis segment, analyzing the time sequence diagnosis segment based on a machine learning decision tree model, and judging whether a refrigerant leakage risk exists or not; And when judging that the refrigerant leakage risk exists, generating a diagnosis result and pushing the diagnosis result to the operation and maintenance background and the user terminal.
2. 2. The method of claim 1, wherein the multi-dimensional serial data comprises one or more of an ambient temperature, a freezer temperature, a refrigerator temperature, a freezer defrost temperature, a refrigerator defrost temperature, a real-time power, a compressor operating state, a refrigerator set operating mode, a freezer door open/close state, a refrigerator door open/close state, a compressor on/off ratio, and a compartment refrigeration state.
3. 3. The method of claim 1, wherein each of the timing diagnostic segments comprises sampled data over a continuous period of time.
4. 4. The method of claim 1, wherein the machine learning decision tree model is XGBoost model, and the typical operation mode change rules before and after refrigerant leakage are learned by training historical normal operation data and known refrigerant leakage sample data.
5. 5. The method of claim 1, wherein the determining logic of the refrigerant leakage is to indirectly determine whether the refrigerant quantity meets the standard through the operation characteristic without directly measuring the refrigerant quantity.
6. 6. The method of claim 5 wherein the operating characteristics include fitting a linear regression to the temperature, the slope of the change in power, and reflecting the continuous change in refrigeration performance of the device.
7. 7. The method of claim 5, wherein the operating characteristics further comprise reflecting whether plant refrigeration performance is stable with standard deviation quantization fluctuations.
8. 8. The method of claim 5, wherein the operating characteristics further comprise counting the number of state switches with a difference to reflect whether device cooling is frequently adjusted.
9. 9. The method of claim 1, wherein the diagnostic result comprises at least one of a risk level, an abnormality index, and a suspected leak time.
10. 10. The method of claim 9, wherein the diagnosis results are synchronously pushed to a user APP or notified to the user through a short message form, and the risk level and the temporary coping advice are notified, and simultaneously, the diagnosis results and the key abnormal data are pushed to a factory operation and maintenance management background for after-sales service personnel to check and make an overhaul plan.
11. 11. The method of claim 1, further comprising performing a feature importance analysis for multi-dimensional cross-validation and interference rejection.
12. 12. An intelligent refrigerator refrigerant leakage diagnosis device, which is characterized in that the intelligent refrigerator refrigerant leakage diagnosis device adopts the method as claimed in any one of claims 1 to 11, and comprises: the data reporting module is used for acquiring multidimensional operation parameters from the networked refrigerator controller and carrying out digital signal encryption and verification processing; The data transmission and storage module is used for uploading the data to the cloud server; The intelligent diagnosis module comprises a data arrangement module, a feature importance analysis module, a refrigerant leakage algorithm model and a risk judgment conclusion module; And the early warning and feedback module is used for triggering an early warning mechanism and distributing the diagnosis result to the operation and maintenance management device and the user terminal.
13. 13. A computer readable storage medium having stored thereon a computer program, which when executed by a processor performs the method of any of claims 1 to 11.
14. 14. A refrigerator characterized in that the intelligent diagnosis of refrigerant leakage is performed by adopting the method as claimed in any one of claims 1 to 11.

Description

Method and device for intelligently diagnosing refrigerant leakage, storage medium and refrigerator Technical Field The invention relates to the technical field of refrigerator fault diagnosis, in particular to a method and device for intelligently diagnosing refrigerant leakage, a storage medium and a refrigerator. Background Currently, most household refrigerators on the market mainly rely on additionally installed special sensor hardware, such as a pressure sensor, a gas concentration sensor or an infrared leakage detection device, for detecting whether leakage occurs to the refrigerant. The method judges whether leakage exists or not by monitoring the pressure change of the refrigerant pipeline or the concentration of refrigerant molecules in the surrounding environment. However, such conventional detection approaches rely on hardware sensors, are costly, and have poor accuracy. More importantly, the traditional detection mechanism is generally lagged, and a user can sense abnormality and report repair only after a large amount of refrigerant is lost and obvious refrigeration failure occurs in the refrigerator, so that a maintainer can confirm the position and degree of leakage by on-site disassembly detection. The 'post maintenance' mode is slow in response and long in service period, and the best intervention time is often missed, so that serious consequences such as compressor burning, complete machine scrapping and the like can be caused, and user experience and brand praise are greatly influenced. Disclosure of Invention The invention aims to provide a method, a device, a storage medium and a refrigerator for intelligently diagnosing refrigerant leakage, which fully utilize multidimensional serial port operation data output by a refrigerator controller, perform cleaning, feature extraction and machine learning modeling through a cloud big data platform, construct a XGBoost decision tree model capable of identifying a refrigerant leakage typical mode, and realize early detection and accurate early warning of refrigerant leakage on the premise of not adding any sensor. In order to achieve the above purpose, the invention provides a method and a device for intelligently diagnosing refrigerant leakage, a storage medium and a refrigerator. The technical scheme of the invention is realized as follows: a method of intelligent diagnosis of refrigerant leakage, the method being adapted to a networked refrigeration appliance, comprising the steps of: collecting multidimensional serial port data in the running process of refrigeration equipment; The collected data is uploaded to a cloud big data platform after validity verification and encryption processing; Cleaning the uploaded data at the cloud end, complementing the missing value and processing the abnormal value; extracting key features to construct a time sequence diagnosis segment, analyzing the time sequence diagnosis segment based on a machine learning decision tree model, and judging whether a refrigerant leakage risk exists or not; And when judging that the refrigerant leakage risk exists, generating a diagnosis result and pushing the diagnosis result to the operation and maintenance background and the user terminal. Further, the multi-dimensional serial data comprises one or more of an ambient temperature, a freezing chamber temperature, a refrigerating chamber temperature, a freezing defrosting temperature, a refrigerating defrosting temperature, real-time power, a compressor running state, a refrigerator set running mode, a refrigerating door opening and closing state, a compressor on-off ratio and a compartment refrigerating state. Further, each of the timing diagnostic segments contains sampling data over a continuous period of time. Furthermore, the machine learning decision tree model is XGBoost model, training is carried out through historical normal operation data and known refrigerant leakage sample data, and typical operation mode change rules before and after refrigerant leakage are learned. Further, the judging logic of the refrigerant leakage is that whether the refrigerant quantity reaches the standard is indirectly judged through the operation characteristics, and the refrigerant quantity is not required to be directly measured. Further, the operation characteristics comprise that the continuous change condition of the refrigeration performance of the equipment is reflected by using linear regression to fit the change slope of temperature and power. Further, the operation characteristics further comprise whether the refrigeration performance of the equipment is stable or not by using the standard deviation quantification fluctuation of the power. Further, the operation characteristics further comprise the step of using the difference value to count the state switching times to reflect whether the refrigeration of the device is frequently regulated. Further, the diagnosis result includes at least one of a risk level, an abnormality i