Fault detection method and system for medical waste treatment equipment based on data analysis

Abstract

The present application relates to the field of data processing technology, and in particular to a method and system for fault detection of medical waste treatment equipment based on data analysis. The method comprises the following steps: calculating the data volatility of data points in real time; constructing a correlation sequence; calculating the degree of abnormality in equipment operation based on the length of the correlation sequence in which the data of the real-time data points updated in real time are located and the data volatility of the data points in the correlation sequence; and judging the equipment operating status based on the degree of abnormality. The step of constructing the correlation sequence comprises setting an initial sequence; calculating the data correlation of data points of different dimensions at each moment in the initial sequence; setting a correlation threshold, and iterating the initial sequence based on the correlation threshold until an iteration stop condition is reached, with the initial sequence after reaching the stop condition being the correlation sequence. The present application has the effect of improving the accuracy of abnormality detection in medical waste treatment equipment.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a fault detection method and system for medical waste treatment equipment based on data analysis. Background Art

[0002] Medical waste refers to waste generated by patients during diagnosis, treatment, and nursing care. Unlike ordinary waste, medical waste typically contains large quantities of pathogens and hazardous chemicals. Medical waste can be incinerated in incinerators or treated using high-temperature and high-pressure sterilization. Due to the environmental impact of incinerators, high-temperature and high-pressure sterilization has become the mainstream method for medical waste treatment. During the high-temperature and high-pressure sterilization process, temperature and humidity are two major factors affecting the effectiveness of medical waste treatment. Excessively high temperatures can cause equipment overheating and failure, while excessively low temperatures risk incomplete pathogen inactivation. Excessive humidity can easily cause equipment corrosion and the formation of aerosols carrying pathogens. Therefore, to ensure the effectiveness of medical waste treatment and the proper operation of medical waste treatment equipment, it is necessary to monitor the temperature and humidity of medical waste treatment equipment during treatment and issue warnings if the temperature or humidity deviates from the abnormal range. Common warning methods involve installing temperature and humidity sensors in the treatment equipment to monitor the temperature and humidity of the medical waste treatment environment. Abnormal detection of medical equipment is achieved through a simple threshold alarm method.

[0003] During the treatment of medical waste, the temperature in the treatment environment is usually around 120 As mentioned above, if the temperature sensor and humidity sensor work in this environment for a long time, the data collected by the sensors will drift (for example, the output value of the temperature sensor decays over time in a high temperature environment), which will lead to inaccurate monitoring data and ultimately false alarms. Therefore, the accuracy of the warning during the use of the abnormality detection method in the related technology is low. Summary of the Invention

[0004] In order to improve the accuracy of early warning of medical waste treatment equipment, the present application provides a fault detection method and system for medical waste treatment equipment based on data analysis.

[0005] In a first aspect, the present application provides a method for monitoring abnormal operation of sensors in medical waste treatment equipment, which adopts the following technical solution:

[0006] A method for fault detection of medical waste treatment equipment based on data analysis includes the following steps: calculating the data fluctuation of data points in real time; constructing a correlation sequence; calculating the abnormality of equipment operation based on the length of the correlation sequence in which the data of the real-time data points updated in real time are located and the data fluctuation of the data points in the correlation sequence; and judging the operating status of the equipment based on the abnormality.

[0007] The steps of constructing the correlation sequence include setting an initial sequence; calculating the data correlation of data points of different dimensions at each moment in the initial sequence; setting a correlation threshold, and iterating the initial sequence based on the correlation threshold until the iteration stop condition is reached, and taking the initial sequence after reaching the stop condition as the correlation sequence; wherein the calculation formula for data correlation is:

[0008] Where, Represents the data correlation of data points of different dimensions collected at the same time in the initial sequence; Indicates the length of the initial sequence; Indicates the first Dimension The data fluctuation of each data point; Indicates the first Dimension The data fluctuation of each data point; Represents an exponential function with a natural constant as its base.

[0009] It represents the difference in data fluctuations of data points of different dimensions at the same moment in the initial sequence. This difference reflects the consistency of the changing trends of the data of the two dimensions, that is, whether the data of the two dimensions change synchronously. If the data of the two dimensions change synchronously, the initial sequence is continuously updated and iterated, resulting in a longer final correlation sequence. The longer the correlation sequence, the more consistent the data changes of multiple dimensions in the correlation sequence. The shorter the length of the correlation sequence, the more inconsistent the data changes of the two dimensions. Combined with the length of the correlation sequence and the data fluctuations of the data points in the correlation sequence, the degree of abnormality of the equipment is calculated, and the status of the equipment is judged based on the degree of abnormality. Compared with the traditional threshold method judgment, this application takes into account the consistency of data changes in multiple dimensions; suppresses the impact of sensor data drift, thereby improving the accuracy of the final equipment detection and early warning.

[0010] Optionally, the iteration stopping condition is: the data correlation is less than a correlation threshold.

[0011] The greater the data correlation, the more consistent the changes in the data values of the two dimensions. Therefore, data points with consistent changes should be grouped into the same correlation sequence. If the calculated data consistency of the two dimensions is low, it means that the two dimensions are currently less correlated. Perhaps the data in one dimension has undergone a sudden change. In this case, the iteration should be stopped, the correlation sequence should be constructed, and the degree of abnormality of the correlation sequence should be calculated.

[0012] Optionally, the steps for calculating data volatility include: The data is the adjacent data set at that moment; the data fluctuation is calculated, and the calculation formula of data fluctuation is:

[0013] Where, Indicates the Dimension The data fluctuation of the data points collected at the moment; Indicates the Dimension The data value collected at the moment; Indicates the Dimension The mean of all data values in the adjacent data set corresponding to the data point at the moment; Indicates the Dimension The standard deviation of all data values in the adjacent data set corresponding to the data point at the moment; Represents the standard normalization function.

[0014] Indicates the Dimension The difference between the mean of a data point at a given moment and the mean of the data in the corresponding adjacent datasets reflects the distribution of that data point relative to the data in the adjacent datasets, and thus the data volatility at that moment. If the mean difference between the data at that moment and the data in the adjacent datasets is large, it indicates that the data at that moment has experienced significant fluctuations; conversely, if the mean difference between the data at that moment and the data in the adjacent datasets is small, it indicates that the data at that moment has experienced minimal fluctuations.

[0015] Optional, any time before The data is the adjacent data set at that moment; the data fluctuation is calculated, and the calculation formula of data fluctuation is: Where, Indicates the Dimension The data fluctuation of the data points collected at the moment; Indicates the Dimension The data value collected at the moment; Indicates the Dimension The data point at time t in the neighboring dataset data values; Represents the number of data points in the neighboring dataset; Indicates the Dimension The standard deviation of all data values in the adjacent data set corresponding to the data point at the moment; Represents the standard normalization function.

[0016] Indicates the Dimension The difference between a data point at a given moment and one of the corresponding data in the adjacent dataset is calculated by summing these differences to obtain the overall difference between the data point and the corresponding adjacent datasets. This method provides a more stable assessment of volatility and avoids the bias caused by a single mean.

[0017] Optionally, the step of setting the initial sequence includes: determining the initial data point, setting the initial length of the initial sequence , taking the initial data point as the benchmark to obtain the The data points constitute the initial sequence.

[0018] Optionally, in response to the absence of a complete correlation sequence in the collected data, the first collected data point is used as the initial data point; in response to the presence of a complete correlation sequence in the collected data, the first data point after the correlation sequence is used as the initial data point.

[0019] The calculation steps of the abnormal degree of equipment operation include: calculating the data abnormality of the correlation sequence where the real-time data points are located; The calculation result of the negative value of the length of the base correlation sequence as the exponent is taken as the interval anomaly; the product of the interval anomaly and the data anomaly is taken as the degree of anomaly.

[0020] The calculation of the degree of anomaly combines the length of the correlation sequence and the data volatility of the data points in the correlation sequence. If a correlation sequence is short and has high data volatility, it indicates that the data changes in multiple dimensions in the correlation sequence are inconsistent and the data changes are large. This indicates that the equipment is anomaly, improving the accuracy of equipment anomaly alarms.

[0021] Optionally, the step of calculating the data anomaly of the correlation sequence where the real-time data point is located includes: calculating the average value of the data fluctuations of all data points in different dimensions in the correlation sequence where the real-time data point is located to obtain the average anomaly; and taking the average of the average anomalies of multiple dimensions as the data anomaly.

[0022] Calculate the mean of data fluctuation anomalies in different dimensions to represent the average anomaly of the dimension, and combine the average anomalies of different dimensions to analyze the overall data anomaly of the correlation sequence.

[0023] Optionally, the step of judging the operating status of the device includes: setting an abnormality threshold, and issuing an alarm in response to the abnormality level being greater than the abnormality threshold.

[0024] In a second aspect, the present application provides a fault detection system for medical waste treatment equipment based on data analysis, which adopts the following technical solutions:

[0025] A fault detection system for medical waste treatment equipment based on data analysis includes a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, the fault detection method for medical waste treatment equipment based on data analysis is implemented.

[0026] The beneficial effect is: the above-mentioned fault detection method for medical waste treatment equipment based on data analysis is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a system based on the memory and the processor for easy use.

[0027] The present application has the following technical effects: the present application combines the data correlation of data points in multiple dimensions to construct a correlation sequence. According to the correlation sequence, the occurrence of false alarms caused by excessive differences between the previous and subsequent data is reduced, thereby improving the accuracy of equipment operation detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a method flow chart of a fault detection method for medical waste treatment equipment based on data analysis in an embodiment of the present application.

[0029] Figure 2 4 is a method flow chart of step S2 of the method for fault detection of medical waste treatment equipment based on data analysis in an embodiment. DETAILED DESCRIPTION

[0030] The embodiment of the present application discloses a fault detection method for medical waste treatment equipment based on data analysis, which calculates the fluctuation data volatility of data points at each dimension at each time; analyzes the data correlation of data changes in different dimensions, and constructs a correlation sequence based on the data correlation. The abnormality of the equipment is calculated based on the data volatility of each data point in the correlation sequence and the number of data points in the correlation sequence, thereby realizing the detection of the equipment. Compared with the threshold alarm method in traditional technology, the collected data is divided into a correlation sequence in this application, and the data in the correlation sequence is analyzed separately, which reduces the situation where false alarms are generated due to inaccurate measurement values caused by drift of measurement values, thereby improving the accuracy of the alarm.

[0031] Reference Figure 1 The fault detection method of medical waste treatment equipment based on data analysis includes steps S1 to S4.

[0032] S1: Calculate the data fluctuation of data points in real time;

[0033] The environmental parameters in the real-time process are collected by the sensor. In this embodiment, the temperature and humidity data in the environment are collected by the temperature sensor and the humidity sensor. The frequency of collection is 10 ;

[0034] The data fluctuation of the real-time collected data point is calculated based on the data of the data points near the real-time collected data point.

[0035] During the operation of the device, data is continuously accumulated. When a certain amount of data is accumulated, the data fluctuation of the data points collected in real time is calculated. Of course, in other embodiments, the data fluctuation of the data points can also be calculated from the beginning of data collection.

[0036] First, build a neighboring dataset of data points collected in real time: Determine the amount of data in the neighboring dataset ; Take the front of the real-time collected data points The data in the adjacent datasets are composed of data points. The fluctuation of the data corresponding to any data point is calculated based on the fluctuation of the data in the adjacent datasets.

[0037] In one embodiment, the calculation formula for data volatility is:

[0038] At any time before The data is the adjacent data set at that moment; the data fluctuation is calculated, and the calculation formula of data fluctuation is:

[0039] Where, Indicates the Dimension The data fluctuation of the data points collected at the moment; Indicates the Dimension The data value collected at the moment; Indicates the Dimension The mean of all data values in the adjacent data set corresponding to the data point at the moment; Represents the standard normalization function.

[0040] It represents the difference between the mean of the data in a certain dimension at a certain moment and the data in the corresponding adjacent dataset. The mean of the data in the adjacent dataset reflects the distribution of the data in the adjacent dataset. This section represents the difference between the data at that moment and the data in the corresponding adjacent dataset. The smaller the difference, the more consistent the data value at that moment is with the distribution of the data in the adjacent dataset, and therefore the data fluctuation at that moment is relatively small. Conversely, if the difference between the data of a certain dimension at a certain moment and the mean of the data in the corresponding adjacent dataset is large, it indicates that the data at that moment has experienced significant fluctuations and may be abnormal data.

[0041] At the same time, the standard deviation is introduced in this formula. By analyzing the distribution of data points in the adjacent data sets, if the standard deviation is smaller, it means that the distribution of data points in the adjacent data sets is more concentrated. The greater the confidence level, the higher the accuracy of data fluctuation calculation.

[0042] In another embodiment, the calculation formula for data volatility is:

[0043] Where, Indicates the Dimension The data fluctuation of the data points collected at the moment; Indicates the Dimension The data value collected at the moment; Indicates the Dimension The data point at time t in the neighboring dataset data values; Represents the number of data points in the neighboring dataset; Indicates the Dimension The standard deviation of all data values in the adjacent data set corresponding to the data point at the moment; Represents the standard normalization function.

[0044] This part represents the difference between the data of a certain dimension at a certain moment and one of the data in the adjacent data set. The difference is accumulated to obtain the overall difference between the data at that moment and the data in the adjacent data set, and then Perform normalization processing to limit the value range of the final calculation result.

[0045] S2: Construct correlation sequence.

[0046] Reference Figure 2 , step S2 includes steps S21 to S23;

[0047] S21: Set the initial sequence:

[0048] Determine the initial data point and set the initial length of the initial sequence , taking the initial data point as the benchmark to obtain the The data points constitute the initial sequence.

[0049] In the process of determining the initial data point, a data point collected by the sensor can be used as the initial data point. However, considering the initial stage of data collection, the amount of data accumulated is small and cannot accurately reflect the operation of the equipment. Therefore, in this embodiment, the 50th data is used as the initial data point. As the data points are continuously collected and updated, when the initial data point is updated After obtaining the data The length of the sequence of data is The initial sequence of

[0050] S22: Calculate the data correlation of data of different dimensions at each moment in the initial sequence;

[0051] The calculation formula for data correlation is:

[0052] Where, Represents the data correlation of data of different dimensions collected at the same time in the initial sequence; Indicates the length of the initial sequence; Indicates the first Dimension The data fluctuation of each data point; Indicates the first Dimension The data fluctuation of each data point; Expressed as a natural constant The exponential function of base .

[0053] It should be noted that the data collected at each data point in this embodiment includes data in two dimensions: temperature and humidity. In other embodiments, other information of the device may also be collected according to actual conditions, such as current, internal pressure of the device, and other data.

[0054] This represents the difference between data from different dimensions at the same moment in the correlation sequence. During normal device operation, the device's temperature and humidity should be stable, with similar or equal data fluctuations. A change in the data for either dimension will alter the trend of the data between the two, affecting the data fluctuations. By traversing the differences in data fluctuations for the different dimensions at all moments in the correlation sequence, the total difference is calculated and then averaged to obtain the average difference in data fluctuations between the two dimensions in the initial sequence. The average difference reflects whether the data changes in the two dimensions are consistent over time. The smaller the difference between the data in the two dimensions, the more consistent the changes, and thus the greater the data correlation between the two dimensions.

[0055] S23: Setting a correlation threshold, and iterating the initial sequence based on the correlation threshold until an iteration stop condition is reached, and taking the initial sequence after the stop condition is reached as the correlation sequence;

[0056] In this embodiment, the correlation threshold is set to 0.8. In response to the correlation of data across different dimensions in the initial sequence exceeding the correlation threshold, the length of the initial sequence is increased by one, and subsequent real-time data points are added to the initial sequence. Finally, the data correlation of the initial sequence is recalculated. Once the correlation of the initial sequence falls below the correlation threshold, the update iterations of the initial sequence cease, and the initial sequence is used as the correlation sequence.

[0057] Specifically, for example, a set of data points 1, 2, 3, 4, and 5 have been collected during the real-time operation of the device. The first data point collected during the operation of the device is used as the initial data point. The initial length of the initial sequence is set to 5. An initial sequence including data points 1, 2, 3, 4, and 5 is then constructed. The data correlation of the initial sequence is calculated. If the calculated data correlation is 0.6, the device operation data is continued to be collected to generate new data points.

[0058] Then, a set of data is generated: 1, 2, 3, 4, 5, and 6. This set of data is used as the initial sequence, and the data correlation of the initial sequence is calculated again and compared with the correlation threshold. If it is greater than the correlation threshold, the data of the device operation is continuously collected to generate new data points.

[0059] Then there is a group of data: 1, 2, 3, 4, 5, 6, 7; take this group of data as the initial sequence, and calculate the data correlation of the initial sequence again, and compare it with the correlation threshold; if it is less than the correlation threshold, then take this group of data as the correlation sequence, and the data in the correlation sequence are 1, 2, 3, 4, 5, 6, 7.

[0060] S3: Calculate the abnormality of the device operation according to the length of the correlation sequence in which the data of the real-time data point updated in real time is located and the data fluctuation of the data point in the correlation sequence.

[0061] Calculate the data anomaly of the correlation series where the real-time data points are located; The calculation result of the negative value of the length of the base correlation sequence as the exponent is taken as the interval anomaly; the product of the interval anomaly and the data anomaly is taken as the degree of anomaly.

[0062] Calculate the average value of the data fluctuation of all data points in different dimensions in the correlation sequence where the real-time data point is located to obtain the average anomaly; the average of the average anomalies of multiple dimensions is taken as the data anomaly;

[0063] Specifically, the calculation of the abnormality degree can be expressed by the following formula:

[0064] Where, Indicates the degree of abnormality of the correlation series; Indicates the first The mean value of data fluctuation for all data points in the dimension; Indicates the first The mean value of data fluctuation for all data points in the dimension; Represents a correlation sequence length; Expressed as a natural constant The exponential function of base .

[0065] When constructing a correlation sequence, if the data volatility of the two dimensions remains consistent over a long period of time, this indicates that the data in the two dimensions are less likely to undergo sudden changes over that long period, leading to a longer correlation sequence. Similarly, a longer correlation sequence indicates a lower likelihood of data mutation during that period, and therefore a lower likelihood of current device anomalies. Conversely, a shorter correlation sequence indicates a greater difference in the data volatility of the two dimensions over a shorter period of time, reflecting a significant, abnormal change in the data during the time corresponding to the correlation sequence, and therefore a greater likelihood of current device anomalies.

[0066] It represents the data fluctuation of the data points in the correlation sequence. The larger the fluctuation, the greater the possibility of abnormal mutation of the data in the correlation sequence, and therefore the greater the possibility of abnormal operation of the current equipment.

[0067] The degree of anomaly is calculated based on the degree of fluctuation and the length of the correlation sequence. The length of the correlation sequence can reflect the changing trends of the data in different dimensions within the correlation sequence. Synchronous changes in data across different dimensions indicate that the data changes are likely due to sensor drift, not equipment anomalies. Therefore, the longer the correlation sequence, the more likely the alarm is suppressed. If data across different dimensions do not change synchronously, this indicates a sudden change in data in one dimension, making the anomaly more likely to be an equipment anomaly. Combining the data fluctuations of the data points in the correlation sequence accurately calculates the degree of anomaly in the current equipment operation, improving the accuracy of equipment anomaly alarms.

[0068] S4: Determine the equipment operating status based on the degree of abnormality.

[0069] Set an abnormality threshold and issue an alarm when the abnormality level exceeds the abnormality threshold;

[0070] In this embodiment, the anomaly threshold is set to 0.75. When the degree of anomaly in the correlation sequence exceeds this threshold, an alarm is issued, indicating that the equipment is operating abnormally. When the degree of anomaly in the correlation sequence is less than the anomaly threshold, step S3 is repeated. During the repetition of step S3, the initial data point is the first data point after the correlation sequence.

[0071] An embodiment of the present application also discloses a fault detection system for medical waste treatment equipment based on data analysis, including a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, a fault detection method for medical waste treatment equipment based on data analysis according to the present application is implemented.

[0072] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0073] The above are all preferred embodiments of the present application, and are not intended to limit the scope of protection of the present application. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A method for fault detection of medical waste treatment equipment based on data analysis, characterized in that: The steps include: real-time calculation of the data fluctuation of the data point, including: The data is the adjacent data set at that moment; the data fluctuation is calculated using the formula: , Indicates the Dimension The data fluctuation of the data points collected at the moment, Indicates the Dimension The data value collected at the moment, Indicates the Dimension The mean of all data values in the adjacent data set corresponding to the data point at time , Indicates the Dimension The standard deviation of all data values in the adjacent data set corresponding to the data point at the moment, Represents a standard normalization function; the data points collected include temperature and humidity or current and internal pressure of the device; Constructing a correlation sequence includes: setting an initial sequence; calculating the data correlation of data points of different dimensions at each moment in the initial sequence; setting a correlation threshold, and iterating the initial sequence based on the correlation threshold until the iteration stop condition is reached, and taking the initial sequence after the stop condition is reached as the correlation sequence; the calculation formula for data correlation is: , Represents the data correlation of data points of different dimensions collected at the same time in the initial sequence, represents the length of the initial sequence, Indicates the first Dimension The data fluctuation of the data points, Indicates the first Dimension The data fluctuation of the data points, represents an exponential function with a natural constant as its base; The abnormality degree of the equipment operation is calculated according to the length of the correlation sequence of the real-time data point updated in real time and the data fluctuation of the data point in the correlation sequence, including: calculating the data abnormality of the correlation sequence where the real-time data point is located, the method is: calculating the average value of the data fluctuation of all data points in different dimensions in the correlation sequence where the real-time data point is located to obtain the average abnormality, and taking the average of the average abnormalities of multiple dimensions as the data abnormality; The negative value of the length of the base correlation sequence is used as the exponential calculation result as the interval anomaly; the product of the interval anomaly and the data anomaly is the degree of anomaly; Determine the equipment operating status based on the degree of abnormality.

2. A method for fault detection of medical waste treatment equipment based on data analysis according to claim 1, characterized in that: The iteration stopping condition is: the data correlation is less than the correlation threshold.

3. A method for fault detection of medical waste treatment equipment based on data analysis according to claim 1, characterized in that: The steps of setting the initial sequence include: determining the initial data point, setting the initial length of the initial sequence , taking the initial data point as the benchmark to obtain the The data points constitute the initial sequence.

4. A method for fault detection of medical waste treatment equipment based on data analysis according to claim 3, characterized in that: In response to the absence of a complete correlation sequence in the collected data, the first collected data point is used as the initial data point; in response to the presence of a complete correlation sequence in the collected data, the first data point after the correlation sequence is used as the initial data point.

5. The method for fault detection of medical waste treatment equipment based on data analysis according to claim 1, characterized in that: The step of judging the operating status of the equipment includes: setting an abnormality threshold, and issuing an alarm in response to the abnormality level being greater than the abnormality threshold.

6. A sensor abnormality monitoring system for medical waste treatment equipment, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a fault detection method for medical waste treatment equipment based on data analysis according to any one of claims 1 to 5 is implemented.

Images