Intelligent detection method, device and equipment for refrigerant leakage

Abstract

The invention relates to the technical field of refrigerants, in particular to an intelligent refrigerant leakage detection method, device and equipment, and the method comprises the steps: taking an existing ventilation execution unit of the equipment as a detection excitation carrier; pre-defining ventilation disturbance in a template mode, and performing matching selection according to the current available ventilation capacity of the equipment and the operation stage during operation; removing the ventilation transition section in combination with a fan feedback state, and defining an effective detection time window corresponding to stable disturbance; based on the output structure response characteristics of the refrigerant concentration sensor, convergence with time position weights is carried out on the relative change of the concentration in the window; according to the method, a judgment index which is compared with a fixed normal response reference level in an equipment debugging stage is introduced, response magnitudes under different working conditions are unified to a comparable scale, and a judgment result is directly mapped into a pre-defined post-detection treatment action at an equipment side, so that a detection result can be instantly converted into a treatment result which can be executed by the equipment.

Description

Technical Field

[0001] This invention relates to the technical field of refrigerants, and more particularly to an intelligent method, apparatus, and equipment for detecting refrigerant leaks. Background Technology

[0002] As refrigeration and heat pump equipment using flammable refrigerants are increasingly used in residential and commercial settings, the safety risks posed by refrigerant leaks are receiving more and more attention.

[0003] Taking flammable refrigerants such as R290 as an example, they are flammable and easily diffused. Once a leak occurs in the cabin or surrounding space, it may form a flammable gas mixture locally, causing safety hazards. Existing leak detection solutions in engineering products mainly rely on concentration sensors placed in critical locations. These sensors trigger alarms or shutdowns by monitoring whether the concentration exceeds a fixed threshold. While this method works when the leak is large and the concentration rises rapidly, it still faces several challenges in actual operating conditions. Minor leaks or early leaks often manifest as slow concentration changes, easily masked or amplified by airflow disturbances generated by equipment operation, fan start-up and shutdown, cabin temperature changes, and external ventilation conditions, leading to both false alarms and missed alarms. At the same time, there is usually a significant spatial concentration gradient inside the equipment. Differences in sensor location can cause the same leak to exhibit different readings under different operating conditions, making it more difficult to set the threshold for single-point detection. To reduce false triggers, engineering often sets the threshold high, which ironically makes it difficult to detect early leaks in a timely manner. More importantly, leak detection is often in a passive observation state. The detection process is highly sensitive to the "airflow conditions at the time". Airflow conditions can change with factors such as fan commands, duct resistance, back pressure, and protection speed limits. Even if the same equipment performs the same detection action at different times, different concentration responses may occur due to differences in the fan transition process and actual ventilation capacity, resulting in a lack of consistent physical reference for the detection conclusions.

[0004] In flammable refrigerant systems, it is not feasible to frequently verify the leakage status by starting and stopping the compressor or by high-energy disturbances. How to ensure that the detection process has clear execution boundaries and that the concentration response corresponds to a specific physical disturbance event without introducing additional risks, thereby improving the early leakage identification capability and the stability of the judgment, remains a long-standing engineering challenge in the existing technology. Summary of the Invention

[0005] To address the above problems, this invention provides a method, apparatus, and equipment for intelligent detection of refrigerant leaks.

[0006] To achieve the above objectives, the present invention provides, in one aspect, a smart detection method for refrigerant leakage, comprising: Read all the pre-configured ventilation perturbation templates in the refrigeration equipment controller, calculate the ventilation matching score in the current operating ventilation mode and output the operation impact score of the refrigeration equipment in this mode. Then, the ventilation matching score and the operation impact score are combined into a comprehensive evaluation result by weighted scoring. After completing the comprehensive evaluation of all the pre-configured ventilation perturbation templates, the ventilation perturbation template with the largest comprehensive evaluation result is selected as the unique index identifier and output, which is recorded as the ventilation perturbation template identifier. The template parameters corresponding to the ventilation disturbance template identifier are read by looking up a table, and the start time of the ventilation disturbance is recorded. Then, a buffer count is introduced to eliminate the ventilation transition time. When the ventilation disturbance meets the end condition defined in the ventilation disturbance template identifier, the ventilation disturbance mode is exited and the end time of the ventilation disturbance is recorded. The detection time window used for this refrigerant leak detection is obtained and the ventilation disturbance execution identifier is output. Using the detection time window as the time filtering condition, the concentration sampling sequence within the window is extracted from the output of the refrigerant concentration sensor of the refrigeration equipment. By introducing a weight sequence that varies with the relative position of the sampling point within the window, the concentration sequence within the window is transformed into a relative change sequence. The refrigerant concentration response characteristics are obtained by weighted averaging. Leakage judgment index is obtained by comparing the refrigerant concentration response characteristics with a pre-configured response reference level. This index is used to determine refrigerant leakage by comparing it with a preset threshold. After the judgment, the corresponding post-detection handling operation is immediately executed.

[0007] Preferably, the ventilation matching score is constrained to arrive Within the range, the higher the value of the ventilation matching score, the higher the degree of matching between the ventilation perturbation template and the current ventilation capacity.

[0008] Preferably, the operational impact score is determined through engineering testing or empirical analysis during the template design phase and stored as a dimensionless score. When the refrigeration equipment enters different operational phases, the controller reads the operational impact score of the corresponding template for that phase from a pre-configured mapping relationship based on the current operational phase status. The operational impact score is located in... arrive Within a certain range, it is used to indicate the acceptability of the refrigeration equipment under the current operating stage of the ventilation perturbation template.

[0009] Preferably, the detection time window is bounded by the actual start and end times of the ventilation disturbance, and items are excluded from both sides. The control cycle consists of several control cycles, represented as follows: ; Where W represents the detection time window used in this refrigerant leak detection. This marks the start of the ventilation disturbance. This is the moment when the ventilation disturbance ends; This is the buffer count calculated based on the fan transition deviation.

[0010] Preferably, the ventilation perturbation execution identifier includes a ventilation perturbation template identifier, a ventilation perturbation start time, a ventilation perturbation end time, a buffer count, and a detection time window.

[0011] Preferably, the refrigerant concentration response characteristics are used for subsequent leak detection, and their construction process is kept consistent with the time boundary of ventilation disturbance events.

[0012] Preferably, after obtaining the leakage detection index, the controller compares it with a pre-configured threshold to complete the determination. When the leakage detection index is greater than the threshold, the controller determines that there is a refrigerant leak; when the leakage detection index is less than the threshold, the controller determines that there is no refrigerant leak.

[0013] Preferably, after the determination is completed, the handling operation is implemented in a predefined action sequence: if a leak is determined to exist, the controller generates a leak detection flag and writes it to the operation status register, writes the event record to the operation log cache, and triggers a pre-configured alarm output, including lighting up the alarm indicator light or driving the buzzer, and outputting a leak indication signal to the safety control module inside the device; if no leak is determined to exist, the controller clears the temporary flag of this detection and ends the detection process.

[0014] To achieve the above objectives, the present invention provides a second aspect of an intelligent refrigerant leak detection device, comprising: The ventilation disturbance template identifier generation module is used to read all the ventilation disturbance templates pre-configured in the refrigeration equipment controller, calculate the ventilation matching score in the current operating ventilation mode and output the operation impact score of the refrigeration equipment in this mode. Then, the ventilation matching score and the operation impact score are combined into a comprehensive evaluation result by weighted scoring. After completing the comprehensive evaluation of all the preset ventilation disturbance templates, the ventilation disturbance template with the largest comprehensive evaluation result is selected as the unique index identifier and output, which is recorded as the ventilation disturbance template identifier. The ventilation disturbance time calculation module is used to read the template parameters corresponding to the ventilation disturbance template identifier by looking up a table, record the start time of the ventilation disturbance, and then introduce a buffer count to eliminate the ventilation transition time. When the ventilation disturbance meets the end condition defined in the ventilation disturbance template identifier, the ventilation disturbance mode is exited and the end time of the ventilation disturbance is recorded. The detection time window used for this refrigerant leak detection is obtained and the ventilation disturbance execution identifier is output. The refrigerant concentration feature extraction module is used to extract the concentration sampling sequence within the window from the output of the refrigerant concentration sensor of the refrigeration equipment, using the detection time window as the time filtering condition. By introducing a weight sequence that changes with the relative position of the sampling point within the window, the concentration sequence within the window is transformed into a relative change sequence. The refrigerant concentration response features are then calculated by weighted averaging. The refrigerant leak detection module is used to obtain a leak judgment index by comparing the refrigerant concentration response characteristics with a pre-configured response reference level. It is used to complete the refrigerant leak judgment with a preset threshold and immediately executes the post-detection handling operation corresponding to the judgment result after the judgment.

[0015] To achieve the above objectives, the present invention provides, in a third aspect, an electronic device comprising: a processor and a memory; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform a refrigerant leak intelligent detection method as described in any of the first aspects.

[0016] The beneficial effects of this invention are as follows: This invention proposes an engineering-oriented detection organization method for refrigerant leakage detection in refrigeration equipment, advancing detection from simple threshold observation to response discrimination linked to controlled ventilation disturbances. Its core lies in utilizing the existing ventilation actuators of the equipment as the detection excitation carrier. Ventilation disturbances are predefined in a template-based manner and matched to the equipment's current available ventilation capacity and operating stage during runtime, ensuring the detection excitation has executable capabilities and consistent control constraints under different operating conditions. During actual ventilation disturbance execution, not only is the start and end of the control command recorded, but the ventilation transition phase is also eliminated based on fan feedback status, thereby defining an effective detection time window corresponding to stable disturbances. This ensures that subsequent concentration sampling strictly falls within the stable stage of the same physical event. Within this time window, the response is constructed solely based on the refrigerant concentration sensor output. Based on the characteristics, by converging the relative concentration changes within the window with time and location weights, the characteristics are made to better fit the main period of stable airflow on leakage diffusion, reducing the dilution of characteristics by transient disturbances. In the judgment stage, a judgment index is introduced that is compared with the normal response reference level solidified during the equipment commissioning stage, unifying the response magnitude under different operating conditions to a comparable scale, and directly mapping the judgment result to the predefined post-detection handling actions on the equipment side, including event recording, status flag setting, and alarm output, so that the detection result can be instantly transformed into an actionable result of the equipment and provide a reliable trigger basis for subsequent safety management. Attached Figure Description

[0017] Figure 1 This is a flowchart of a smart refrigerant leakage detection method in a specific embodiment of the present invention; Figure 2 This is a block diagram of a refrigerant leakage intelligent detection system according to a specific embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some 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 are within the scope of protection of the present invention.

[0019] refer to Figure 1 As shown, one embodiment of this application proposes a smart detection method for refrigerant leakage, including: S101: Read all pre-configured ventilation perturbation templates in the refrigeration equipment controller, calculate the ventilation matching score for the current operating ventilation mode and output the operational impact score of the refrigeration equipment under this mode. Then, use a weighted scoring method to synthesize the ventilation matching score and operational impact score into a comprehensive evaluation result. After completing the comprehensive evaluation of all preset ventilation perturbation templates, select the ventilation perturbation template with the largest comprehensive evaluation result as a unique index identifier and output it, denoted as the ventilation perturbation template identifier, which specifically includes: This step determines the ventilation perturbation method to be used for subsequent detection before the refrigeration equipment enters the refrigerant leak detection process. This ventilation perturbation method is not generated ad hocly during operation, nor does it rely on dynamic adjustment or adaptive optimization. Instead, it is selected from several ventilation perturbation templates pre-configured during the equipment design or commissioning phase. The purpose of this is to ensure that subsequent observations of refrigerant concentration changes are based on controlled, stable, and comparable physical conditions, thereby avoiding interference from random ventilation or inconsistent operation on the detection results.

[0020] During the equipment design or commissioning phase, several ventilation perturbation templates are pre-configured in the controller's non-volatile storage unit. Each ventilation perturbation template corresponds to a fixed ventilation execution mode, which is explicitly described in the template, for example, by a ventilation intensity parameter characterizing the control requirements of the template on the fan during execution. This ventilation intensity parameter is expressed in the form of a ratio or a level, making different templates comparable and facilitating matching with the available fan capacity during operation. In actual operation, the controller first reads the currently allowed maximum ventilation capacity level from the fan drive unit or internal control register. This capacity level is also expressed in the form of a ratio or a level, reflecting the actual upper limit of the fan's dispatchable capacity under the current operating state and protection strategy constraints.

[0021] After obtaining the currently available ventilation capacity, the controller calculates the feasibility of each candidate ventilation perturbation template under the current operating state. This calculation is based on the deviation fitting concept commonly used in engineering, which involves comparing the deviation between the required ventilation intensity of the template and the currently available ventilation capacity to obtain a dimensionless score for measuring the template's fit. The calculation method is as follows: ; in, Indicates the first The ventilation matching score of a ventilation perturbation template under the current equipment condition is used to reflect the feasibility of the template under the current conditions; Indicates the first The ventilation intensity parameters are pre-cured in each ventilation perturbation template. These parameters are determined by engineers based on the equipment structure and ventilation path during the template configuration phase and stored in the controller. This indicates the maximum permissible ventilation capacity level under the current operating conditions of the equipment. This value is directly given by the fan control logic or protection strategy. Represents a pre-defined minimum positive value, used in... To avoid calculation errors when taking the minimum value, the above calculations are performed. Constrained arrive Within the range, the larger the value, the higher the degree of matching between the template and the current ventilation capacity.

[0022] In addition to the matching degree of ventilation capacity, each ventilation perturbation template is assigned an inherent parameter during the configuration phase to characterize its perturbation properties, reflecting the degree of impact of the template on the stability of equipment operation during execution. This parameter is not calculated in real time during operation, but is determined through engineering testing or empirical analysis during the template design phase and stored as a dimensionless score. When the equipment enters different operating phases, the controller reads the corresponding template's operational impact score for that phase from the pre-configured mapping relationship based on the current operating phase status, thus obtaining... This rating is also located in arrive Within a certain range, it is used to characterize the acceptability of the template at the current operational stage.

[0023] Obtain the ventilation matching score for each template. and operational impact score Subsequently, the controller employs a weighted scoring method commonly used in engineering decision-making to combine the two scores into a single evaluation result, which is used for ranking and selection among multiple ventilation perturbation templates. The calculation method is shown below: ; in, Indicates the first The comprehensive evaluation results of the ventilation perturbation template under the current equipment condition; and This represents a pre-defined weighting coefficient used to balance the trade-off between ventilation matching and operational stability. This weight is determined during the equipment model or product design phase and remains unchanged during operation. Through the above calculations, the controller can effectively sort a limited number of ventilation perturbation templates without introducing complex optimization or learning processes.

[0024] After all candidate ventilation perturbation templates have undergone comprehensive evaluation, the controller selects the ventilation perturbation template with the highest comprehensive evaluation result and outputs its unique index in the template table as the ventilation perturbation template identifier. The ventilation perturbation template marking Used to determine the specific implementation method of ventilation perturbation in subsequent steps.

[0025] S102: The template parameters corresponding to the ventilation disturbance template identifier are read by looking up a table, the start time of the ventilation disturbance is recorded, and then a buffer count is introduced to eliminate the ventilation transition time. When the ventilation disturbance meets the end condition defined in the ventilation disturbance template identifier, the ventilation disturbance mode is exited and the end time of the ventilation disturbance is recorded. The detection time window used for this refrigerant leak detection is obtained and the ventilation disturbance execution identifier is output, specifically including: Ventilation Disturbance Template Marking A unique ventilation perturbation template record corresponds to each controller. This template record is written to the controller's non-volatile storage unit during equipment commissioning or design phases to describe a standardized ventilation perturbation execution. The template record contains at least the following: normalized parameters characterizing the target ventilation intensity, execution rules defining the duration of the ventilation perturbation, and control information describing the recovery method after the ventilation perturbation ends. In this step, no further filtering or calculation of the template is performed; it is simply identified as a ventilation perturbation template. Use it as an index to directly call the corresponding template content.

[0026] The controller receives the ventilation perturbation template identifier Then, the ventilation perturbation template identifier is first read from the non-volatile memory cell using a lookup table. The corresponding template parameters are used, and the target ventilation intensity parameters defined in the template are converted into control quantities recognizable by the fan drive. These control quantities can be sent to the fan actuator unit by writing to the fan drive register, outputting the control duty cycle, or setting control status bits; the specific form depends on the fan control structure of the equipment. Within the same control cycle after issuing the control command and confirming that the fan has entered the ventilation disturbance mode, the controller records the start time of the ventilation disturbance. This moment originates from the controller's internal system clock or counter.

[0027] After the ventilation disturbance begins, the fan moves from its original working point to the formwork. The defined target ventilation state transition occurs. During this transition, the actual ventilation state of the fan gradually approaches the target state. To distinguish between the "ventilation transition phase" and the "ventilation stabilization phase" in time, the controller reads the fan feedback sequence from the fan driver at fixed control cycles after the ventilation perturbation begins. The fan feedback can come from the speed feedback count, Hall signal count results provided by the fan driver, or a normalized feedback value estimated internally by the driver based on the control input, as long as the feedback can be updated synchronously with changes in the fan state. Simultaneously, the controller uses template identifiers... Read the ventilation intensity parameters of the template from the template table, and convert the ventilation intensity parameters into the normalized target value of the fan feedback domain according to the pre-fixed conversion relationship. This is used for subsequent deviation calculations.

[0028] To determine the ventilation transition length that needs to be eliminated in time, this step introduces a buffer count. The design of this buffer count is derived from methods used in control engineering and signal processing to handle system transient processes, specifically by proactively avoiding regions where the system's dynamics are not yet stable before feature extraction or decision-making. Specifically, the buffer count is determined by the minimum buffer size preset in the template and the actual transient deviation, and its calculation method is as follows: ; in, This indicates the count of buffers that need to be removed, and its unit is the number of control cycles. This indicates the minimum buffer count preset in the template, used to cover typical wind turbine startup or switching processes; This represents the preset deviation amplification factor in the template, used to map ventilation transition deviations into additional buffer lengths; This indicates the number of feedback samples used to assess ventilation transition deviations; its value is determined by the control cycle and the preset assessment duration in the template. Indicates the first day after the start of ventilation disturbance. The feedback normalization value of the secondary air blower; template The corresponding target ventilation intensity normalization value; A preset minimum positive value, used in The calculation remains stable when taking smaller values. The above calculation results are converted to integer counts through a rounding operation, so that they can be directly used to control the periodic index.

[0029] When ventilation disturbance meets the template When the termination condition is defined, the controller sends a recovery command to the fan driver to exit the ventilation disturbance mode, and records the end time of the ventilation disturbance within the same control cycle in which the fan exits the disturbance state. Based on this, the controller uses the actual start and end times of ventilation disturbances as boundaries, and eliminates disturbances on both sides. A control cycle is established to construct a detection time window that strictly corresponds to the ventilation disturbance: ; in, This indicates the detection time window used in this refrigerant leak detection; This marks the start of the ventilation disturbance. This is the moment when the ventilation disturbance ends; This is the buffer count calculated based on the fan transient deviation. Using this method, the detection time window naturally focuses on the stable execution phase of ventilation disturbances, thus avoiding the inclusion of transient disturbances during fan start-up, shutdown, or switching phases in subsequent refrigerant concentration response analysis.

[0030] Ventilation Disturbance Implementation Marking At least include ventilation perturbation template identification Ventilation disturbance start time End of ventilation disturbance Buffer count and the detection time window determined by the above parameters. This execution identifier is recorded as a whole in the controller and used for the acquisition and determination of refrigerant concentration in subsequent steps.

[0031] S103: Using the detection time window as the time filtering condition, the concentration sampling sequence within the window is extracted from the output of the refrigerant concentration sensor of the refrigeration equipment. By introducing a weighted sequence that varies with the relative position of the sampling point within the window, the concentration sequence within the window is transformed into a relative change sequence. The refrigerant concentration response characteristics are then calculated through weighted averaging, specifically including: Around the detection time window given in S102 The concentration sequence within a window is extracted from the refrigerant concentration sensor output of the refrigeration equipment, and this sequence is transformed into a response feature that characterizes the "concentration-time response pattern" during the stable phase of ventilation perturbation. This feature is used for leak detection in subsequent steps, and its construction process is consistent with the time boundary of the ventilation perturbation event, giving the feature a clear physical orientation.

[0032] By window As a time-based filtering condition, the concentration sampling sequence within the window is extracted from the controller's continuous sampling buffer of the refrigerant concentration sensor; the ventilation control is not changed, nor is the sensor sampling frequency.

[0033] The refrigerant concentration sensor outputs concentration readings to the controller at fixed sampling periods. The controller timestamps each reading and writes it to a circular buffer. The controller receives a ventilation perturbation execution flag. Then, using the detection time window By filtering the readings in the buffer at the boundary, a concentration sequence sorted by time is obtained within the window. ,in This represents the number of sampling points within the window. This is the first valid sampled value near the start of the window. Because of the window... As defined in step two based on the stable segment of ventilation perturbation, the concentration sequence obtained in this step naturally corresponds to the concentration response under stable airflow perturbation.

[0034] Within the stable phase of ventilation perturbation, the carrying and diffusion effects of airflow on leaked refrigerant are typically more pronounced in the middle of the window. To reflect this temporal distribution characteristic, this step introduces a weighted sequence that varies with the relative position of the sampling point within the window. The relative change is calculated using a weighted average. This construction originates from the classic practice of weighted aggregation of time-series samples in signal processing and statistics, namely, the weighted mean. Unlike the common "equal weighting," this uses symmetrical linear triangular weights, ensuring that samples near the center of the window contribute more to the features, thus focusing the features more on the core period of stable disturbances. The weights are defined as follows: ; in, Indicates the first in the window Time weights for each sample; The sample number. ; This represents the total number of samples within the window. This weight is... and The time value is ,exist When the value is close to the center of the window, it is close to the value. Furthermore, it is symmetric about the center of the window. The calculation of this weight depends only on the sample index and the total number of samples, making it easy to implement in the controller. When the number of samples within the window... When the value is less than the preset minimum, the controller will process all samples in the window with equal weight and continue to complete the response feature calculation.

[0035] After defining the weights, this step uses the first sample in the window. As a reference benchmark, the concentration sequence within the window is converted into a relative change sequence. The refrigerant concentration response characteristics were obtained by calculating their weighted average. This structure is also derived from the weighted average absolute deviation expression in statistics. Its purpose is to summarize the overall change in concentration relative to the baseline within a window using a single quantitative indicator, while reducing the impact of individual outliers. Its calculation is as follows: ; in, This indicates the refrigerant concentration response characteristics; Indicates the first in the window Each concentration sample value is directly output by the refrigerant concentration sensor; The first concentration sample value within the window is used as a baseline for relative change. The time weights are calculated using the previous formula; This represents the total number of samples within the window. Because... For dimensionless weights, the dimensions of the numerator and denominator are determined by... Decision, therefore The dimensional meaning is consistent with the concentration sampling value, which is in line with common sense.

[0036] Refrigerant concentration response characteristics By execution identifier Detection time window Concentration sampling sequence within and their corresponding weights The calculated value is used for leakage determination in subsequent steps.

[0037] The logical relationship between the two equations above is as follows: The first equation generates a set of symmetrical linear weight sequences on the discrete-time index to reflect the time distribution of the ventilation perturbation stability segment, which is "more representative in the middle section"; the second equation, based on this, performs weighted aggregation on the relative change sequence to obtain a single response feature. This is used for subsequent leak detection.

[0038] In one feasible embodiment, a detection time window is set. The Internal Union obtained The concentration samples, in chronological order, are: .

[0039] First calculate according to the weighting formula : , , , , The denominator is .

[0040] Then calculate the relative change: , , , , The molecule is .

[0041] therefore The controller calculates accordingly. Then, this output is used as the judgment for subsequent steps. For ease of engineering debugging and verification, this step can write key results of a single detection process to the runtime log in the controller: the window boundary referenced in the execution identifier, and the number of sampling points within the window. Calculated response characteristics This log does not alter the algorithm flow; it is used solely for on-site debugging or experimental archiving to facilitate comparative analysis of calculation results under different operating conditions.

[0042] S104: Based on the refrigerant concentration response characteristics, a leak detection index is obtained by comparing it with a pre-configured response reference level. This index is used to determine refrigerant leaks against a preset threshold. Immediately after the detection, post-detection handling procedures corresponding to the detection result are executed, specifically including: The controller pre-configures a set of reference parameters for decision-making during the device design or commissioning phase and writes them into non-volatile memory. These reference parameters include the response reference level. With the judgment threshold .in, The typical response amplitude used to characterize the stable range of ventilation perturbations under leak-free conditions is derived from multiple executions of the same testing procedure on the equipment under leak-free conditions, using the same calculation method as in step three. The results were statistically analyzed, and representative levels were selected as reference values. The threshold is used to transform the difference between the current response and the reference level into a definite boundary. To ensure consistency in the decision-making process and avoid direct comparisons between response values ​​of different magnitudes, this step adopts the normalized ratio decision-making approach commonly used in engineering decision-making. This approach originates from the relative deviation and ratio test expressions in statistics and engineering measurement, which obtain a dimensionless ratio by dividing the observed quantity by the reference quantity, and then compares it with the threshold to complete the decision.

[0043] Based on the above idea, the controller will use the currently detected response features Mapped to leakage detection indicators The calculation method is as follows: ; in, This represents a leakage determination index, used as a dimensionless ratio for comparison. The refrigerant concentration response characteristics output in step three; A pre-configured response reference level. Because... and The ratio represents the magnitude of the same type of response. It is a dimensionless quantity, which is convenient for unified processing in the equipment controller and conforms to common sense.

[0044] After obtaining the leakage determination index Then, the controller compares it with a pre-configured threshold. Comparisons are made to determine the leak detection criteria. When the controller detects a refrigerant leak, it determines that a leak has occurred; when the leak detection index... At this time, the controller determines that there is no refrigerant leak. Threshold Similarly, for solidified storage, the determination method can be based on experience settings during the equipment commissioning phase or on coverage settings based on test data, so that it maintains a stable judgment boundary under common environmental disturbances and sensor noise conditions.

[0045] In one feasible embodiment, the response reference level fixed during the device commissioning phase is: The threshold for judgment is In this detection process, the output response characteristics of step three are as follows: The judgment indicator is... ,because The controller then issues a "no leakage" conclusion. For another example, if a detection yields... ,but ,because The controller then determines that a leak exists. The above calculation involves only one division and one comparison, facilitating stable execution in embedded controllers.

[0046] After determining a leak, the controller immediately executes post-detection handling operations. These operations are implemented using a predefined sequence of actions, which is fixed in the program using a state machine or procedure call, requiring no additional calculations during execution. If a leak is detected, the controller generates a leak detection flag and writes it to the operating status register. Simultaneously, it writes the event record to the operating log cache and triggers pre-configured alarm outputs, such as illuminating an alarm indicator light, activating a buzzer, or outputting a leak indication signal to the device's internal safety control module. If no leak is detected, the controller clears the temporary flag for this detection and terminates the detection process. To facilitate on-site debugging and test archiving, the controller can write critical calculations into a debugging log, such as recording... , , These records, along with the judgment results, are used for experimental analysis and do not change the judgment logic.

[0047] refer to Figure 2 As shown, a second aspect of this application provides a refrigerant leak intelligent detection system, comprising: The ventilation disturbance template identifier generation module is used to read all the ventilation disturbance templates pre-configured in the refrigeration equipment controller, calculate the ventilation matching score in the current operating ventilation mode and output the operation impact score of the refrigeration equipment in this mode. Then, the ventilation matching score and the operation impact score are combined into a comprehensive evaluation result by weighted scoring. After completing the comprehensive evaluation of all the preset ventilation disturbance templates, the ventilation disturbance template with the largest comprehensive evaluation result is selected as the unique index identifier and output, which is recorded as the ventilation disturbance template identifier. The ventilation disturbance time calculation module is used to read the template parameters corresponding to the ventilation disturbance template identifier by looking up a table, record the start time of the ventilation disturbance, and then introduce a buffer count to eliminate the ventilation transition time. When the ventilation disturbance meets the end condition defined in the ventilation disturbance template identifier, the ventilation disturbance mode is exited and the end time of the ventilation disturbance is recorded. The detection time window used for this refrigerant leak detection is obtained and the ventilation disturbance execution identifier is output. The refrigerant concentration feature extraction module is used to extract the concentration sampling sequence within the window from the output of the refrigerant concentration sensor of the refrigeration equipment, using the detection time window as the time filtering condition. By introducing a weight sequence that changes with the relative position of the sampling point within the window, the concentration sequence within the window is transformed into a relative change sequence. The refrigerant concentration response features are then calculated by weighted averaging. The refrigerant leak detection module is used to obtain a leak judgment index by comparing the refrigerant concentration response characteristics with a pre-configured response reference level. It is used to complete the refrigerant leak judgment with a preset threshold and immediately executes the post-detection handling operation corresponding to the judgment result after the judgment.

[0048] A third aspect of this application provides an electronic device, including: a processor and a memory; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform a refrigerant leak intelligent detection method as described in the first aspect of the present invention.

[0049] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for intelligent detection of refrigerant leakage, characterized in that, Comprising: Reading all the ventilation perturbation templates pre-configured in the refrigeration equipment controller, calculating the ventilation matching score in the current running ventilation mode and outputting the running impact score on the refrigeration equipment in the mode, then weighting the scores to synthesize the comprehensive evaluation results, after completing the comprehensive evaluation of all pre-set ventilation perturbation templates, selecting the ventilation perturbation template with the largest comprehensive evaluation result as the unique index identifier and outputting, recorded as the ventilation perturbation template identifier; Reading the template parameters corresponding to the ventilation perturbation template identifier through table lookup, recording the start time of the ventilation perturbation, then introducing a buffer count to remove the ventilation transition time, when the ventilation perturbation meets the end condition defined in the ventilation perturbation template identifier, exiting the ventilation perturbation mode and recording the end time of the ventilation perturbation, obtaining the detection time window used for this refrigerant leakage detection and outputting the ventilation perturbation execution identifier; Taking the detection time window as the time screening condition, extracting the concentration sampling sequence within the window from the refrigerant concentration sensor output of the refrigeration equipment, converting the concentration sequence within the window into a relative change sequence by introducing a weight sequence that changes with the relative position of the sampling point within the window, and calculating the refrigerant concentration response feature by weighted average; Comparing the refrigerant concentration response feature with the pre-configured response reference level to obtain a leakage judgment index, which is used to complete the refrigerant leakage judgment with the pre-set threshold, and immediately executes the post-detection disposal operation corresponding to the judgment result after the judgment.

2. The method of claim 1, wherein, The ventilation match score is constrained to be in the range of to where a higher numerical value of the ventilation match score indicates a higher degree of match between the ventilation perturbation template and the current ventilation capabilities.

3. The method of claim 1, wherein, The operation influence score is determined by engineering test or empirical analysis through the template design stage and is stored in the form of a dimensionless score. When the refrigeration equipment enters different operation stages, the controller reads the operation influence score of the corresponding template in the stage from the pre-configured mapping relationship according to the current operation stage state. The operation influence score is located in the range of 0 to 1, indicating the acceptable degree of the refrigeration equipment in the current operation stage under the ventilation perturbation template. to .

4. The method of claim 1, wherein, The detection time window is bounded by the actual start and end time of the ventilation perturbation and is truncated on both sides is composed of one control period and is expressed as: ； wherein W represents a detection time window used for this time refrigerant leakage detection, is a ventilation perturbation start time; is a ventilation perturbation end time; is a buffer count calculated from a fan transition deviation.

5. The method of claim 1, wherein, The ventilation perturbation execution identifier includes the ventilation perturbation template identifier, the ventilation perturbation start time, the ventilation perturbation end time, the buffer count and the detection time window.

6. The method of claim 1, wherein, The refrigerant concentration response feature is used for subsequent leakage judgment, and its construction process remains consistent with the time boundary of the ventilation perturbation event.

7. The method of claim 1, wherein, After obtaining the leakage judgment index, the controller compares it with the pre-configured threshold to complete the judgment, when the leakage judgment index is greater than the threshold, the controller judges that there is refrigerant leakage; when the leakage judgment index is less than the threshold, the controller judges that there is no refrigerant leakage.

8. The method of claim 1, wherein, After completing the judgment, the disposal operation is implemented in a predefined action sequence: if the judgment exists leakage, the controller generates a leakage detection flag and writes it into the running state register, at the same time, writes the event record into the running log buffer, and triggers the pre-configured alarm output, including lighting the alarm indicator or driving the buzzer, outputting the leakage indication signal to the safety control module inside the equipment; if the judgment does not exist leakage, the controller clears the temporary flag of this detection and ends the detection process.

9. A refrigerant leakage intelligent detection device, characterized in that, Applied to a refrigerant leakage intelligent detection method as claimed in any one of claims 1-8, comprising: The ventilation perturbation template identification generation module is configured to read all ventilation perturbation templates pre-configured in the refrigeration equipment controller, calculate a ventilation matching score in a current running ventilation mode and an output running impact score on the refrigeration equipment in the mode, then synthesize the ventilation matching score and the running impact score into a comprehensive evaluation result in a weighted score manner, select a ventilation perturbation template with the largest comprehensive evaluation result as a unique index identification and output, and record as a ventilation perturbation template identification. The ventilation perturbation time calculation module is configured to read template parameters corresponding to the ventilation perturbation template identification through a table lookup manner, record a start time of the ventilation perturbation, then introduce a buffer count to remove a ventilation transition time, exit the ventilation perturbation mode when the ventilation perturbation satisfies an end condition defined in the ventilation perturbation template identification, record an end time of the ventilation perturbation, obtain a detection time window used for the current refrigerant leakage detection, and output a ventilation perturbation execution identification. The refrigerant concentration feature extraction module is configured to extract a concentration sampling sequence in the window from the refrigerant concentration sensor output of the refrigeration equipment with the detection time window as a time screening condition, convert the concentration sequence in the window into a relative change amount sequence by introducing a weight sequence changing with a relative position of the sampling point in the window, and obtain a refrigerant concentration response feature through weighted average calculation. The refrigerant leakage detection module is configured to obtain a leakage judgment index by comparing the refrigerant concentration response feature with a pre-configured response reference level, complete refrigerant leakage judgment with a pre-set threshold, and immediately execute a post-detection disposal operation corresponding to the judgment result after the judgment.

10. An electronic device, comprising: The refrigerant leakage detection module is configured to obtain a leakage judgment index by comparing the refrigerant concentration response feature with a pre-configured response reference level, complete refrigerant leakage judgment with a pre-set threshold, and immediately execute a post-detection disposal operation corresponding to the judgment result after the judgment. It comprises: a processor and a memory; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory, so that the processor executes the refrigerant leakage intelligent detection method in any one of claims 1-8.

Images