Abnormity processing method for carbon dioxide refrigeration equipment

Abstract

The invention discloses a carbon dioxide refrigeration equipment abnormity processing method, and belongs to the technical field of carbon dioxide refrigeration equipment. According to the invention, equipment operation parameters and environmental parameters are collected, the state of the refrigeration equipment can be evaluated more comprehensively, and a detection blind area caused by neglecting environmental factors is avoided; meanwhile, different time intervals are set to obtain the data, so that the change rates of different parameters can be more flexibly adapted, and the real-time performance and accuracy of the data are ensured; data are classified through a clustering algorithm, and abnormal data points can be identified more accurately by calculating the distance from the data points to the cluster center and setting a threshold value. The sensitivity of anomaly detection is improved, so that the anomaly recognition speed is improved; through the multi-objective optimization method, the relation between different objectives can be balanced, and efficient operation of equipment and effective processing of anomalies are realized. The problem of low abnormal detection sensitivity caused by incomplete detection data in the prior art is solved.

Description

Technical Field

[0001] The invention belongs to the technical field of carbon dioxide refrigeration equipment, and in particular relates to a method for handling an abnormality of carbon dioxide refrigeration equipment. Background Art

[0002] As global awareness of environmental protection increases, carbon dioxide as a refrigerant has gained attention due to its environmental friendliness. Carbon dioxide refrigeration technology has been widely used in refrigeration systems due to its natural, non-toxic and low global warming potential (GWP) characteristics. However, due to the high working pressure and low critical point of the carbon dioxide refrigeration system, higher requirements are placed on the safety and stability of the equipment operation.

[0003] The existing refrigeration equipment detection system has the defect of single detection parameters and cannot comprehensively evaluate the status of the refrigeration equipment. Furthermore, the traditional refrigeration equipment detection system monitors by detecting the status data of the refrigeration equipment itself. The environmental parameters around the refrigeration equipment also reflect the status of the refrigeration equipment to a certain extent. Since the detection data of the existing technology is not comprehensive, the existing detection system has the problem of low sensitivity in detecting abnormalities.

[0004] In summary, the existing refrigeration equipment testing has problems such as incomplete detection parameters, inability to accurately assess equipment status, and slow response to changes in environmental parameters. It is difficult to meet the high requirements of the carbon dioxide refrigeration system for equipment operation safety and stability. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a method for handling abnormalities of carbon dioxide refrigeration equipment in view of the deficiencies in the above-mentioned prior art, which is designed to be sensitive, reliable and easy to promote and use.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A method for handling an abnormality of a carbon dioxide refrigeration device, characterized in that the method comprises the following steps:

[0008] Step 1: Acquire environmental parameters at a first set time interval, and acquire equipment operating parameters at a second set time interval; the environmental parameters include ambient temperature, ambient humidity, air pressure, and air quality; the equipment operating parameters include voltage, current, vibration speed, compressor speed, condensing pressure, evaporation pressure, and refrigerant flow rate; the first set time interval is greater than the second set time interval;

[0009] Step 2: Perform data preprocessing on the acquired environmental parameters and equipment operation parameters, cluster the preprocessed data according to the set K value, and determine the cluster center; calculate the distance from each data point to the cluster center; compare the distance with the set distance threshold, and when the distance is greater than the set distance threshold, set the label of the data point as abnormal; when the distance is less than or equal to the set distance threshold, set the label of the data point as normal; the sum of the environmental parameter type and the equipment operation parameter type is consistent with the set K value;

[0010] Step 3: Input the distribution of abnormal labels within the set time length and the obtained environmental parameters and equipment operation parameters into the multi-objective optimization algorithm model, output a solution with the goal of minimizing energy consumption and maximizing refrigeration efficiency, and control the refrigeration equipment according to the solution to eliminate the abnormality;

[0011] The particle swarm optimization algorithm is used to solve the multi-objective optimization algorithm model. The corresponding execution steps are:

[0012] s1. Define the objective function to minimize energy consumption and maximize cooling efficiency, and set constraints;

[0013] s2. Initialize the particle swarm. Each particle represents a solution. The position and velocity of the particle are randomly generated and recorded as Y j =[y j1 ,y j2 ,…,y jm ] and V j =[v j1 ,v j2 ,…,v jm ], where m refers to the dimension of the problem, that is, the number of variables; at the same time, the following parameters are set: population size, maximum number of iterations, inertia weight, learning factor, random number;

[0014] s3. Calculate the fitness corresponding to the position of each particle, update the position and speed of each particle, and obtain the individual optimal position and the global optimal position;

[0015] s4. Repeat step s3 until the solution is output after the convergence condition is met. The convergence condition is that the maximum number of iterations is 100 times, the fitness value changes less than 0.1% for 10 consecutive iterations, or a solution that meets all constraints has been found and the fitness value is greater than 0.9.

[0016] Furthermore, in step 2, the distance adopts the Euclidean distance, and its calculation formula is as follows:

[0017]

[0018] Where x=(x 1 ,x 2 ,…,xn ) is the feature vector of the data point, c k =(c k1 ,c k2 ,…,c kn ) is the eigenvector of the kth cluster center, w i is the weight of the i-th feature, k is the index of the cluster, ranging from 1 to c, c is equal to the set K value, and n is the dimension of the data point and cluster center.

[0019] Furthermore, the distance threshold is set based on the distance distribution of each data point in the cluster to the cluster center, and its calculation formula is as follows:

[0020] Threshold k =μ k +α×σ k

[0021] Among them, μ k is the average of the Euclidean distances from all data points in the kth cluster to the cluster center, σ k is the standard deviation of the Euclidean distance from all data points in the kth cluster to the cluster center, and α∈[2,3] is a multiple of the standard deviation.

[0022] Furthermore, the standard deviation multiple is determined according to the abnormal rules corresponding to each type of data; the abnormal rules include: when the ambient temperature exceeds the set temperature range, it is determined as a temperature abnormality; when the ambient humidity exceeds the set humidity range, it is determined as a humidity abnormality; when the air pressure exceeds the set air pressure range, it is determined as an air pressure abnormality; when the air quality is less than the set quality threshold, it is determined as an air quality abnormality; when the voltage exceeds the set voltage range, it is determined as a voltage abnormality; when the current exceeds the set current range, it is determined as a current abnormality; when the compressor speed is greater than the rated speed, it is determined as a compressor speed abnormality; when the condensing pressure is less than the standard condensing pressure, it is determined as a condensing pressure abnormality; when the evaporating pressure is less than the standard evaporating pressure, it is determined as an evaporating pressure abnormality; when the refrigerant flow is less than the set flow threshold, it is determined as a refrigerant flow abnormality.

[0023] Furthermore, in step 2, preprocessing includes removing outliers, filling missing values, denoising, and data verification.

[0024] Furthermore, in step three, the formula for expressing the optimal position of an individual is:

[0025]

[0026] Among them, P best，j is the individual optimal position of the kth particle, Y j =[y j1 ,y j2 ,…,y jm] is the position of the jth particle, m is the dimension of the search space, f(Y j ) is the jth particle at the current position Y j The fitness value, f(P best，j ) is the jth particle at its individual optimal position P best，j The fitness value of

[0027] The expression formula of the global optimal position is:

[0028] G best =min(f(P best，1 ),f(P best，2 ),…,f(P best，j ))

[0029] Among them, G best is the global optimal position, min is the operation of selecting the minimum value, f(P best，1 ) is the first particle at its individual optimal position P best，1 The fitness value, f(P best，2 ) is the second particle at its individual optimal position P best，2 The fitness value of .

[0030] Furthermore, the calculation formula of the fitness value is:

[0031] f(Y j )=w a ×E(Y j )+w b ×C(Y j )+w b ×S(Y j )+w b ×P(Y j )

[0032] Among them, E(Y j ) is the energy efficiency ratio function, which means the jth particle is at the current position Y j Energy efficiency ratio, C(Y j ) is the system stability function, which indicates that the jth particle is at the current position Y j The system stability value, S(Y j ) is the safety margin function, which indicates that the jth particle is at the current position Y j The safety margin value, P(Y j ) energy consumption function, indicating that the jth particle is at the current position Y j Energy consumption value, w a 、w b 、w c 、w d All are weight coefficients;

[0033] The expression of the energy efficiency ratio function is:

[0034]

[0035] in, is the particle position Y j The cooling capacity, is the particle position Y j The input power, is the particle position Y j Refrigerant flow rate, c p is the specific heat capacity, is the particle position Y j The ambient temperature difference, is the particle position Y j The voltage, is the particle position Y j The current, is the power factor;

[0036] The expression of the system stability function is:

[0037]

[0038] Among them, s is the order of data collection, l is the total number of data collection within the set time length, is the particle position Y j The working pressure fluctuation range during the sth data collection, is the particle position Y j The ambient temperature fluctuation range during the sth data collection, is the particle position Y j The vibration fluctuation range during the sth data collection, P max is the maximum value of working pressure, T max is the maximum value of the ambient temperature, K max is the maximum value of vibration, w P 、w T 、w K All are weight coefficients;

[0039] The expression of the safety margin function is:

[0040]

[0041] Among them, min is the minimum value operation, P max is the maximum pressure, is the particle position Y j Pressure value, T max is the maximum value of the ambient temperature, is the particle position Y j The ambient temperature value;

[0042] The expression of energy consumption function is:

[0043]

[0044] in, is the particle position Y j The input power, is the particle position Y j The voltage, is the particle position Y j The current, is the power factor.

[0045] Furthermore, the formula for updating the velocity of each particle is:

[0046] V j t+1 =w×V j t +c 1 × 1 ×(P best，j -Y j t )+c 2 × 2 ×(G best -Y j t )

[0047] Among them, w is the inertia weight, c 1 、c 2 are learning factors, r 1 、r 2 All are random numbers, V j t is the velocity of the jth particle at time t, V j t+1 is the velocity of the jth particle at time t+1, Y j t is the position of the jth particle at time t;

[0048] The formula for updating the position of each particle is:

[0049] Y j t+1 =Y j t +V j t

[0050] Among them, Y j t+1 is the position of the jth particle at time t+1.

[0051] Furthermore, in step s1, the constraint condition includes a constraint on speed, which is expressed as:

[0052] V j t =clip(V j t , -V max ,V max )

[0053] Among them, V j t is the velocity of the jth particle at time t, and clip is to ensure that V j Will not exceed -V max ~V max Range, V max For maximum speed.

[0054] Furthermore, the position of the particle includes compressor speed, valve opening, fan speed, refrigerant flow, evaporation temperature setting value, exhaust temperature setting value, superheat setting value, subcooling setting value, condensation pressure setting value, and evaporation pressure setting value.

[0055] Compared with the prior art, the present invention has the following advantages:

[0056] The present invention not only collects equipment operating parameters, but also collects environmental parameters, which can more comprehensively evaluate the status of the refrigeration equipment and avoid detection blind spots caused by ignoring environmental factors; at the same time, setting different time intervals to obtain data can more flexibly adapt to the change rate of different parameters to ensure the real-time and accuracy of the data; by preprocessing the collected data, noise and outliers are eliminated to ensure data quality; by classifying the data through a clustering algorithm, by calculating the distance from the data point to the cluster center and setting a threshold, abnormal data points can be more accurately identified. The sensitivity of anomaly detection is improved, and the speed of anomaly identification is thereby improved; the relationship between different objectives can be balanced through a multi-objective optimization method, and efficient operation of the equipment and effective handling of anomalies can be achieved. The problem that the detection data of the prior art is not comprehensive enough, resulting in low sensitivity in detecting anomalies.

[0057] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 The figure is a flow chart of an embodiment of a method for handling an abnormality of a carbon dioxide refrigeration device according to the present invention. DETAILED DESCRIPTION

[0059] Example of abnormality handling method for carbon dioxide refrigeration equipment:

[0060] like Figure 1As shown, a method for handling an abnormality of a carbon dioxide refrigeration device comprises the following steps:

[0061] Step 1: Obtain environmental parameters at the first set time interval, and obtain equipment operating parameters at the second set time interval; environmental parameters include ambient temperature, ambient humidity, air pressure, and air quality, and equipment operating parameters include voltage, current, vibration speed, compressor speed, condensing pressure, evaporation pressure, and refrigerant flow; the first set time interval is greater than the second set time interval. The first set time interval is 10s, and the second set time interval is 1s. The above data is obtained through the corresponding sensors.

[0062] Step 2: Perform data preprocessing on the acquired environmental parameters and equipment operation parameters, cluster the preprocessed data according to the set K value, and determine the cluster center; calculate the distance from each data point to the cluster center; compare the distance with the set distance threshold, when the distance is greater than the set distance threshold, set the label of the data point to abnormal, when the distance is less than or equal to the set distance threshold, set the label of the data point to normal; the sum of the environmental parameter type and the equipment operation parameter type is consistent with the set K value.

[0063] Among them, preprocessing includes removing outliers, filling missing values, denoising, and data verification. Among them, denoising is based on discrete wavelet transform (DWT). Removing outliers is to screen abnormal fluctuation values ​​according to the set upper and lower thresholds to ensure the validity of the data. When filling missing values, for missing data, mean interpolation, nearest neighbor interpolation and other methods can be used to fill in. The purpose of data verification is to ensure the consistency of data format and meet the predetermined standards.

[0064] In order to reduce the system overhead of this method and improve the calculation speed, the above distance uses the Euclidean distance, and its calculation formula is as follows:

[0065]

[0066] Where x=(x 1 ,x 2 ,…,x n ) is the feature vector of the data point, c k =(c k1 ,c k2 ,…,c kn ) is the eigenvector of the kth cluster center, w i is the weight of the i-th feature, k is the index of the cluster, ranging from 1 to c, c is equal to the set K value, and n is the dimension of the data point and the cluster center. According to the feature type of the data point, the corresponding w iThe value of can be set to multiple values, for example: the pressure feature weight is 0.3; the temperature feature weight is 0.25; the flow feature weight is 0.2; the electrical feature weight is 0.15; and the vibration feature weight is 0.1.

[0067] According to actual needs, the distance can also be Euclidean distance, Manhattan distance or Chebyshev distance.

[0068] In order to improve the flexibility of the judgment, the set distance threshold is determined based on the distance distribution of each data point in the cluster to the cluster center. The calculation formula is as follows:

[0069] Threshold k =μ k +α×σ k

[0070] Among them, μ k is the average of the Euclidean distances from all data points in the kth cluster to the cluster center, σ k is the standard deviation of the Euclidean distance from all data points in the kth cluster to the cluster center, and α∈[2,3] is a constant.

[0071] Specifically, the standard deviation multiple α is determined according to the abnormal rules corresponding to each type of data; the abnormal rules include: when the ambient temperature exceeds the set temperature range, it is determined as a temperature abnormality; when the ambient humidity exceeds the set humidity range, it is determined as a humidity abnormality; when the air pressure exceeds the set air pressure range, it is determined as an air pressure abnormality; when the air quality is less than the set quality threshold, it is determined as an air quality abnormality; when the voltage exceeds the set voltage range, it is determined as a voltage abnormality; when the current exceeds the set current range, it is determined as a current abnormality; when the compressor speed is greater than the rated speed, it is determined as a compressor speed abnormality; when the condensing pressure is less than the standard condensing pressure, it is determined as a condensing pressure abnormality; when the evaporating pressure is less than the standard evaporating pressure, it is determined as an evaporating pressure abnormality; when the refrigerant flow is less than the set flow threshold, it is determined as a refrigerant flow abnormality.

[0072] Step 3: Input the distribution of abnormal labels within the set time length and the obtained environmental parameters and equipment operation parameters into the multi-objective optimization algorithm model, with the goal of minimizing energy consumption and maximizing cooling efficiency, output a solution, and control the refrigeration equipment according to the solution to eliminate the abnormality. The set time length range is 10 minutes to 2 hours, because the abnormality is generally not instantaneous, but has precursors. By analyzing the data over a period of time, it is possible to more accurately determine the abnormal failure that is about to occur or has occurred.

[0073] The particle swarm optimization algorithm is used to solve the multi-objective optimization algorithm model. The corresponding execution steps are:

[0074] s1. Define the objective function to minimize energy consumption and maximize refrigeration efficiency, and set constraints. The constraints include the safety range of various parameters of the refrigeration equipment, such as exhaust temperature ≤ 120℃; suction superheat ≥ 5℃; working pressure ≤ 90bar, etc. In order to avoid excessive update speed, skipping the individual optimal position, resulting in a large deviation between the determined individual optimal position and the actual individual optimal position, the constraints also include constraints on the update speed, which can be expressed as:

[0075] V j t =clip(V j t , -V max ,V max )

[0076] Among them, V j t is the velocity of the jth particle at time t, and clip is to ensure that V j Will not exceed -V max ~V max Range, V max For maximum speed.

[0077] s2. Initialize the particle swarm. Each particle represents a solution. The position and velocity of the particle are randomly generated and recorded as Y j =[y j1 ,y j2 ,…,y jm ] and V j =[v j1 ,v j2 ,…,v jm ], where m refers to the dimension of the problem, that is, the number of variables; at the same time, the following parameters are set: population size, maximum number of iterations, inertia weight, learning factor, and random number.

[0078] The positions of the particles in step s2 include compressor speed, valve opening, fan speed, refrigerant flow, evaporation temperature setting value, exhaust temperature setting value, superheat setting value, subcooling setting value, condensation pressure setting value, and evaporation pressure setting value. The different states of these positions are combined into a particle group to form a solution. The solution of each refrigeration equipment will have a slight deviation. The corresponding particle position is set according to the actual refrigeration equipment, and the solution is solved by using this method to achieve the processing of various abnormalities.

[0079] s3. Calculate the fitness corresponding to the position of each particle, update the position and speed of each particle, and obtain the individual optimal position and the global optimal position.

[0080] s4. Repeat step s3 until the solution is output after the convergence condition is met. The convergence condition is that the maximum number of iterations is 100 times, the fitness value changes less than 0.1% for 10 consecutive iterations, or a solution that meets all constraints has been found and the fitness value is greater than 0.9.

[0081] Among them, the expression formula of the individual optimal position in step three is:

[0082]

[0083] Among them, P best，j is the individual optimal position of the jth particle, Y j =[y j1 ,y j2 ,…,y jm ] is the position of the jth particle, m is the dimension of the search space, f(Y j ) is the jth particle at the current position Y j The fitness value, f(P best，j ) is the jth particle at its individual optimal position P best，j The fitness value of

[0084] The expression formula of the global optimal position is:

[0085] G best =min(f(P best，1 ),f(P best，2 ),…,f(P best，j ))

[0086] Among them, G best is the global optimal position, min is the operation of selecting the minimum value, f(P best，1 ) is the first particle at its individual optimal position P best，1 The fitness value, f(P best，2 ) is the second particle at its individual optimal position P best，2 The fitness value of .

[0087] The formula for updating the velocity of each particle is:

[0088] V j t+1 =w×V j t +c 1 × 1 ×(P best，j -Y j t )+c 2 × 2 ×(G best -Y j t )

[0089] Among them, w is the inertia weight, c 1 、c 2 are learning factors, r 1 、r 2 All are random numbers, V j t is the velocity of the jth particle at time t, V j t+1 is the velocity of the jth particle at time t+1, Y j t is the position of the jth particle at time t;

[0090] The formula for updating the position of each particle is:

[0091] Y j t+1 =Y j t +V j t

[0092] Among them, Y j t+1 is the position of the jth particle at time t+1.

[0093] The calculation formula for the above fitness value is:

[0094] f(Y j )=w a ×E(Y j )+w b ×C(Y j )+w b ×S(Y j )+w b ×P(Y j )

[0095] Among them, f(Y j ) is the jth particle at the current position Y j The fitness value, E(Y j ) is the energy efficiency ratio function, which means the jth particle is at the current position Y j Energy efficiency ratio, C(Y j ) is the system stability function, which indicates that the jth particle is at the current position Y j The system stability value, S(Y j ) is the safety margin function, which indicates that the jth particle is at the current position Y j The safety margin value, P(Y j ) energy consumption function, indicating that the jth particle is at the current position Y j Energy consumption value, w a 、w b 、w c、w d are weight coefficients. a 、w b 、w c 、w d The values ​​are: 0.4, 0.3, 0.2, 0.1 respectively.

[0096] Wherein, the expression of the energy consumption function is:

[0097] E(Y j )+w b ×C(Y j )+w b ×S(Y j )+w b ×P(Y j )

[0098] The expression of the energy efficiency ratio function is:

[0099]

[0100] in, is the particle position Y j The cooling capacity, is the particle position Y j The input power, is the particle position Y j Refrigerant flow rate, c p is the specific heat capacity, which is a known physical parameter of the device, is the particle position Y j The ambient temperature difference, is the particle position Y j The voltage, is the particle position Y j The current, is the power factor.

[0101] The expression of the system stability function is:

[0102]

[0103] Among them, s is the order of data collection, l is the total number of data collection within the set time length, is the particle position Y j The working pressure fluctuation range during the sth data collection, is the particle position Y j The ambient temperature fluctuation range during the sth data collection, is the particle position Y j The vibration fluctuation range during the sth data collection, P max is the maximum value of working pressure, T maxis the maximum value of the ambient temperature, K max is the maximum value of vibration, w P 、w T 、w K are weight coefficients, w P 、w T 、w K are 0.5, 0.3, and 0.2 respectively, where P max , T max , K max It can also be replaced with the corresponding normal operating value, the purpose of which is to normalize these data to facilitate the quantification of system stability. The working pressure fluctuation range refers to the difference between the maximum and minimum values, and the working pressure range is 0 to 100 bar; the ambient temperature fluctuation range refers to the difference between the maximum and minimum values, and the ambient temperature range is -40℃ to 120℃; the vibration fluctuation range refers to the difference between the maximum and minimum values, and the vibration range is 0 to 1000Hz.

[0104] The expression of the safety margin function is:

[0105]

[0106] Among them, min is the minimum value operation, P max is the maximum pressure, is the particle position Y j Pressure value, T max is the maximum value of the ambient temperature, is the particle position Y j The ambient temperature value.

[0107] The expression of energy consumption function is:

[0108]

[0109] in, is the particle position Y j The input power, is the particle position Y j The voltage, is the particle position Y j The current, The power factor can be calculated by dividing the actual power by the total power provided by the power supply.

[0110] The fitness value calculation formula is the fitness function, where maximizing the cooling efficiency is the positive weighted part of the objective function. The fitness value will increase with the increase of the energy efficiency ratio. If the weight w aIf the weight w is high enough, the fitness function can push the particle swarm optimization goal toward maximizing COP. Minimizing energy consumption is the negative weighted part of the function, and the fitness value decreases as energy consumption increases. d High enough, optimization will tend to reduce energy consumption. a 、w b 、w c 、w d These weight coefficients can be adjusted autonomously to find a balance between the two. The system stability and safety margin in the fitness function mainly play a role in ensuring system stability and operational safety. The goal of maximizing cooling efficiency and minimizing energy consumption is achieved under the premise of ensuring system stability and operational safety.

[0111] The present invention realizes comprehensive data collection through step one. Specifically, the present solution not only obtains equipment operating parameters, but also obtains environmental parameters, including ambient temperature, humidity, air pressure, air quality, etc. These environmental parameters are closely related to the equipment operating status and can more comprehensively reflect the actual operating environment and status of the refrigeration equipment. At the same time, setting different time intervals to obtain data can more flexibly adapt to the change rate of different parameters and ensure the real-time and accuracy of the data. Through step two, efficient and accurate data processing and abnormality identification are realized. In step two, the collected data is preprocessed to eliminate noise and abnormal values ​​to ensure data quality, and then the data is classified by using a clustering algorithm. By calculating the distance from the data point to the cluster center and setting a threshold, the abnormal data point can be more accurately identified. Compared with the traditional single parameter threshold judgment, this method can more effectively capture the abnormal changes of data and improve the sensitivity of abnormality detection. Through step three, the corresponding optimal solution is determined according to the abnormality identified in step two. Specifically, the distribution of the abnormal label and the obtained parameters are input into the multi-objective optimization algorithm model, and multiple goals such as energy consumption and refrigeration efficiency are comprehensively considered to output the optimal solution, and the refrigeration equipment is accurately controlled to eliminate the abnormality. This multi-objective optimization method can balance the relationship between different objectives and achieve efficient operation of equipment and effective handling of anomalies.

[0112] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural change made to the above embodiment based on the technical essence of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A method for handling abnormalities of carbon dioxide refrigeration equipment, characterized in that: The method comprises the following steps: Step 1: Acquire environmental parameters at a first set time interval, and acquire equipment operating parameters at a second set time interval; the environmental parameters include ambient temperature, ambient humidity, air pressure, and air quality; the equipment operating parameters include voltage, current, vibration speed, compressor speed, condensing pressure, evaporation pressure, and refrigerant flow rate; the first set time interval is greater than the second set time interval; Step 2: Perform data preprocessing on the acquired environmental parameters and equipment operation parameters, cluster the preprocessed data according to the set K value, and determine the cluster center; calculate the distance from each data point to the cluster center; compare the distance with the set distance threshold, and when the distance is greater than the set distance threshold, set the label of the data point as abnormal; when the distance is less than or equal to the set distance threshold, set the label of the data point as normal; the sum of the environmental parameter type and the equipment operation parameter type is consistent with the set K value; Step 3: Input the distribution of abnormal labels within the set time length and the obtained environmental parameters and equipment operation parameters into the multi-objective optimization algorithm model, output a solution with the goal of minimizing energy consumption and maximizing refrigeration efficiency, and control the refrigeration equipment according to the solution to eliminate the abnormality; The particle swarm optimization algorithm is used to solve the multi-objective optimization algorithm model. The corresponding execution steps are: s1. Define the objective function to minimize energy consumption and maximize cooling efficiency, and set constraints; s2. Initialize the particle swarm. Each particle represents a solution. The position and velocity of the particle are randomly generated and recorded as Y j =[y j1 ,y j2 ,…,y jm ] and V j =[v j1 ,v j2 ,…,v jm ], where m refers to the dimension of the problem, that is, the number of variables; at the same time, the following parameters are set: population size, maximum number of iterations, inertia weight, learning factor, random number; s3. Calculate the fitness corresponding to the position of each particle, update the position and speed of each particle, and obtain the individual optimal position and the global optimal position; s4. Repeat step s3 until the solution is output after the convergence condition is met. The convergence condition is that the maximum number of iterations is 100 times, the fitness value changes less than 0.1% for 10 consecutive iterations, or a solution that meets all constraints has been found and the fitness value is greater than 0.

9.

2. A method for handling abnormalities of carbon dioxide refrigeration equipment according to claim 1, characterized in that: In step 2, the distance is calculated using the Euclidean distance, which is calculated using the following formula: Where x=(x1,x2,…,x n ) is the feature vector of the data point, c k =(c k1 ,c k2 ,…,c kn ) is the eigenvector of the kth cluster center, w i is the weight of the i-th feature, k is the index of the cluster, ranging from 1 to c, c is equal to the set K value, and n is the dimension of the data point and cluster center.

3. A method for handling abnormalities of carbon dioxide refrigeration equipment according to claim 2, characterized in that: The distance threshold is set based on the distance distribution of each data point in the cluster to the cluster center, and its calculation formula is as follows: Threshold k =μ k +α×σ k Among them, μ k is the average of the Euclidean distances from all data points in the kth cluster to the cluster center, σ k is the standard deviation of the Euclidean distance from all data points in the kth cluster to the cluster center, and α∈[2,3] is a multiple of the standard deviation.

4. A method for handling abnormality of carbon dioxide refrigeration equipment according to claim 3, characterized in that: The standard deviation multiple is determined according to the abnormal rules corresponding to each type of data; the abnormal rules include: when the ambient temperature exceeds the set temperature range, it is judged as temperature abnormality; when the ambient humidity exceeds the set humidity range, it is judged as humidity abnormality; when the air pressure exceeds the set air pressure range, it is judged as air pressure abnormality; when the air quality is less than the set quality threshold, it is judged as air quality abnormality; when the voltage exceeds the set voltage range, it is judged as voltage abnormality; when the current exceeds the set current range, it is judged as current abnormality; when the compressor speed is greater than the rated speed, it is judged as compressor speed abnormality; when the condensing pressure is less than the standard condensing pressure, it is judged as condensing pressure abnormality; when the evaporating pressure is less than the standard evaporating pressure, it is judged as evaporating pressure abnormality; when the refrigerant flow is less than the set flow threshold, it is judged as refrigerant flow abnormality.

5. A method for handling abnormality of carbon dioxide refrigeration equipment according to claim 1, characterized in that: In step 2, preprocessing includes removing outliers, filling missing values, denoising, and data verification.

6. A method for handling abnormality of carbon dioxide refrigeration equipment according to claim 1, characterized in that: In step three, the formula for expressing the optimal position of an individual is: Among them, P best，j is the individual optimal position of the jth particle, Y j =[y j1 ,y j2 ,…,y jm ] is the position of the jth particle, m is the dimension of the search space, f(Y j ) is the jth particle at the current position Y j The fitness value, f(P best，j ) is the jth particle at its individual optimal position P best，j The fitness value of The expression formula of the global optimal position is: G best =min(f(P best，1 ),f(P best，2 ),…,f(P best，j )) Among them, G best is the global optimal position, min is the operation of selecting the minimum value, f(P best，1 ) is the first particle at its individual optimal position P best，1 The fitness value, f(P best，2 ) is the second particle at its individual optimal position P best，2 The fitness value of .

7. A method for handling abnormality of carbon dioxide refrigeration equipment according to claim 6, characterized in that: The calculation formula of the fitness value is: f(Y j )=w a ×E(Y j )+w b ×C(Y j )+w b ×S(Y j )+w b ×P(Y j ) Among them, E(Y j ) is the energy efficiency ratio function, which means the jth particle is at the current position Y j Energy efficiency ratio, C(Y j ) is the system stability function, which indicates that the jth particle is at the current position Y j The system stability value, S(Y j ) is the safety margin function, which indicates that the jth particle is at the current position Y j The safety margin value, P(Y j ) energy consumption function, indicating that the jth particle is at the current position Y j Energy consumption value, w a 、w b 、w c 、w d All are weight coefficients; The expression of the energy efficiency ratio function is: in, is the particle position Y j The cooling capacity, is the particle position Y j The input power, is the particle position Y j Refrigerant flow rate, c p is the specific heat capacity, is the particle position Y j The ambient temperature difference, is the particle position Y j The voltage, is the particle position Y j The current, is the power factor; The expression of the system stability function is: Among them, s is the order of data collection, l is the total number of data collection within the set time length, is the particle position Y j The working pressure fluctuation range during the sth data collection, is the particle position Y j The ambient temperature fluctuation range during the sth data collection, is the particle position Y j The vibration fluctuation range during the sth data collection, P max is the maximum value of working pressure, T max is the maximum value of the ambient temperature, K max is the maximum value of vibration, w P 、w T 、w K All are weight coefficients; The expression of the safety margin function is: Among them, min is the minimum value operation, P max is the maximum pressure, is the particle position Y j Pressure value, T max is the maximum value of the ambient temperature, is the particle position Y j The ambient temperature value; The expression of energy consumption function is: in, is the particle position Y j The input power, is the particle position Y j The voltage, is the particle position Y j The current, is the power factor.

8. A method for handling abnormality of carbon dioxide refrigeration equipment according to claim 6, characterized in that: The formula for updating the velocity of each particle is: Among them, w is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers, V j t is the velocity of the jth particle at time t, V j t+1 is the velocity of the jth particle at time t+1, Y j t is the position of the jth particle at time t; The formula for updating the position of each particle is: AND j t+1 =And j t +V j t Among them, Y j t+1 is the position of the jth particle at time t+1.

9. A method for handling abnormality of carbon dioxide refrigeration equipment according to claim 1, characterized in that: In step s1, the constraint condition includes a constraint on speed, which is expressed as: In j t =clip(V j t ,-V max ,V max ) Among them, V j t is the velocity of the jth particle at time t, and clip is to ensure that V j Will not exceed -V max ~V max Range, V max For maximum speed.

10. A method for handling abnormalities of carbon dioxide refrigeration equipment according to claim 1, characterized in that: The position of the particles includes compressor speed, valve opening, fan speed, refrigerant flow, evaporation temperature setting value, exhaust temperature setting value, superheat setting value, subcooling setting value, condensation pressure setting value, and evaporation pressure setting value.

Images