Biological instrument equipment fault prediction method and system

Abstract

The invention discloses a biological instrument equipment fault prediction method and system, and relates to the technical field of biological instrument equipment predictive maintenance, and the method comprises the steps: collecting the operation data of biological instrument equipment, and carrying out the time sequence sampling of the operation data through a preset data sampling period, and obtaining time sequence sample data; based on a set feature weight coefficient, performing weighted fusion processing on the time sequence sample data to generate fused feature data; inputting the fused feature data into a pre-trained fault feature extraction model, and extracting a fault feature vector; and inputting the fault feature vector into a pre-constructed fault prediction model, and outputting early warning information. According to the invention, the accuracy of fault detection and the timeliness of early warning are improved, stable operation of equipment is effectively guaranteed, the maintenance cost is reduced, and the method has high engineering application and popularization values.

Description

Technical Field

[0001] The present invention relates to the technical field of predictive maintenance of biological instrument equipment, and particularly to a method and system for predicting faults in biological instrument equipment. Background Art

[0002] With the rapid development of biotechnology, biological instrument equipment is increasingly widely used in fields such as medical diagnosis, scientific research, and biopharmaceuticals. These devices generally have high-precision and high-reliability requirements, and their operating status directly affects the accuracy and repeatability of experimental results. Traditional maintenance of biological instrument equipment mostly adopts a passive management mode of regular maintenance or repair after a fault occurs. This method not only has high maintenance costs but also is difficult to effectively prevent sudden faults. In recent years, with the development of Internet of Things technology and artificial intelligence algorithms, data-driven device health management methods have gradually received attention. However, most of the existing fault prediction methods monitor and analyze single parameters or use simple threshold judgment mechanisms, and do not fully consider the multi-parameter coupling effect and dynamic characteristics of fault evolution of biological instrument equipment, resulting in insufficient prediction accuracy and prone to false alarms or missed alarms.

[0003] Currently, the main technical problems in the field of fault prediction of biological instrument equipment include: the mismatch between the data acquisition frequency and the dynamic characteristics of the equipment, resulting in the loss of important fault feature information; the lack of an effective fusion mechanism between data from different types of sensors, making it difficult to comprehensively reflect the overall operating status of the equipment; the traditional feature extraction methods have insufficient ability to capture long-term dependence relationships and mutation characteristics of time-series data; the prediction models generally have problems such as weak generalization ability and poor environmental adaptability. In addition, the existing technologies often ignore the periodic characteristics during the operation of biological instrument equipment and the relevance between components, making it difficult to accurately identify early signs of potential faults; To solve the above technical problems, the present invention proposes a method and system for predicting faults in biological instrument equipment. Summary of the Invention

[0004] In view of the problems of imperfect data acquisition and fusion, limited feature extraction ability, and poor generalization of prediction models in the existing fault prediction technology of biological instrument equipment, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is how to improve the accuracy of fault identification and the timeliness of early warning response, and achieve intelligent monitoring and risk prevention and control of the equipment operating status.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a method for predicting faults of biological instruments and equipment, which includes collecting operating data of the biological instruments and equipment, performing time series sampling on the operating data through a preset data sampling period to obtain time series sample data; performing weighted fusion processing on the time series sample data based on a set feature weight coefficient to generate fused feature data; inputting the fused feature data into a pre-trained fault feature extraction model to extract a fault feature vector; inputting the fault feature vector into a pre-built fault prediction model to output warning information.

[0008] As a preferred solution of the biological instrument equipment fault prediction method of the present invention, the pre-built fault prediction model adopts a hybrid prediction architecture combining a gradient boosting decision tree ensemble algorithm and a Markov chain prediction mechanism.

[0009] As a preferred solution of the biological instrument equipment fault prediction method of the present invention, it also includes: inputting the fault feature vector into the fault prediction model, identifying the abnormal pattern in the fault feature vector through the collaborative judgment of multiple decision trees, and determining the type of fault that has occurred; combining the determined fault type with the historical fault evolution data, using the Markov chain prediction mechanism, and at the same time calculating the fault development trend based on the state transition probability matrix to predict the probability of fault occurrence; setting differentiated fault thresholds in the fault prediction model, for different biological instrument equipment components, when the probability of a certain type of fault is greater than the fault threshold, triggering the generation of the warning information, wherein the warning information includes the fault type, the probability of fault occurrence, the expected time of fault occurrence, and the scope of fault impact.

[0010] As a preferred solution of the biological instrument equipment fault prediction method described in the present invention, the fused feature data is input into a pre-trained fault feature extraction model to extract a fault feature vector, including: the pre-trained fault feature extraction model adopts an improved bidirectional long short-term memory network structure; the improved bidirectional long short-term memory network structure includes multiple hidden layers and a self-attention mechanism module.

[0011] As a preferred solution of the biological instrument fault prediction method described in the present invention, it also includes: dividing the fused feature data into a plurality of time window sequences according to the sliding window method, and the length of the time window is determined according to the working cycle characteristics of the biological instrument; performing deep feature learning on the fused feature data in the time window through the fault feature extraction model to extract implicit features including time dimension information, wherein the output layer of the fault feature extraction model maps the extracted deep features to a feature space of fixed dimension through nonlinear transformation to generate a fault feature vector.

[0012] As a preferred solution of the biological instrument device fault prediction method of the present invention, wherein: the method for obtaining the fused feature data is to calculate the correlation coefficient between each data type and the occurrence of a fault according to the historical fault correlation degree of the time-series sample data; normalize the correlation coefficient to obtain the corresponding feature weight coefficient, where the sum of the feature weight coefficients is one; based on the feature weight coefficient, use a non-linear feature fusion function to weight various types of data, and introduce a non-linear adjustment coefficient to adjust the contribution degree of the time-series sample data, and perform a non-linear transformation through the feature mapping index; combine and stack the weighted time-series sample data to generate fused feature data.

[0013] As a preferred solution of the biological instrument device fault prediction method of the present invention, wherein: the method for obtaining the time-series sample data is to collect operation data through multiple groups of sensors during the operation of the biological instrument device, and perform timed sampling on the operation data according to a preset data sampling period, where the operation data includes temperature data, voltage data, and vibration data; sort and integrate the operation data according to the time stamp to form time-series sample data with time-series characteristics, where the time-series sample data is stored in the form of a multi-dimensional matrix, with rows representing the sampling time and columns representing the sensor data types.

[0014] In a second aspect, an embodiment of the present invention provides a biological instrument device fault prediction system, which includes: a collection module for collecting operation data of the biological instrument device, performing time-series sampling on the operation data through a preset data sampling period to obtain time-series sample data; a generation module for performing weighted fusion processing on the time-series sample data based on a set feature weight coefficient to generate fused feature data; an extraction module for inputting the fused feature data into a pre-trained fault feature extraction model to extract a fault feature vector; and a prediction module for inputting the fault feature vector into a pre-constructed fault prediction model to output a warning message.

[0015] In a third aspect, an embodiment of the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and wherein: when the computer program instructions are executed by the processor, the steps of the biological instrument device fault prediction method as described in the first aspect of the present invention are implemented.

[0016] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and wherein: when the computer program instructions are executed by the processor, the steps of the biological instrument device fault prediction method as described in the first aspect of the present invention are implemented.

[0017] The beneficial effects of the present invention are as follows: By constructing a fault prediction method for biological instrument equipment that combines fusion feature acquisition, deep fault feature extraction, and multi-stage prediction models, multi-dimensional accurate modeling and trend deduction of the equipment operation state are realized; through feature weight weighting and non-linear fusion, the expression ability of key features is improved; based on the improved bidirectional long short-term memory network and self-attention mechanism, potential abnormal patterns in time series data are effectively captured; by combining the gradient boosting decision tree and the Markov chain model, the recognition and prediction ability of fault types and evolution paths is enhanced; the setting of differential fault thresholds makes the early warning response more targeted; this method overall improves the accuracy of fault detection and the timeliness of early warning, effectively ensures the stable operation of the equipment and reduces the maintenance cost, and has high engineering application and promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings according to these drawings without creative efforts. Among them:

[0019] Figure 1 It is a flowchart of the fault prediction method for biological instrument equipment in Embodiment 1. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention will be described in detail below with reference to the drawings of the specification.

[0021] Many specific details are set forth in the following description in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention, so the present invention is not limited by the specific embodiments disclosed below.

[0022] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation manner of the present invention. The "in one embodiment" that appears in different places in this specification does not all refer to the same embodiment, nor is it a separate or alternative embodiment that mutually excludes other embodiments.

[0023] Embodiment 1

[0024] Refer to Figure 1 , which is the first embodiment of the present invention. This embodiment provides a fault prediction method for biological instrument equipment, including the following specific steps as shown in Figure 1 :

[0025] S1: Collect the operation data of the biological instrument and equipment, perform time-series sampling on the operation data through a preset data sampling period to obtain time-series sample data;

[0026] S2: Based on the set feature weight coefficients, perform weighted fusion processing on the time-series sample data to generate fusion feature data;

[0027] S3: Input the fusion feature data into a pre-trained fault feature extraction model to extract fault feature vectors;

[0028] S4: Input the fault feature vectors into a pre-constructed fault prediction model to output warning information.

[0029] In the embodiment of the present application, the above step S1 includes:

[0030] Specifically, during the operation of the biological instrument and equipment, collect operation data through multiple groups of sensors and perform regular sampling on the operation data according to the preset data sampling period;

[0031] It should be noted that the operation data includes temperature data, voltage data, and vibration data; install multiple groups of sensors at key parts of the biological instrument and equipment, where the sensors include thermistor temperature sensors, voltage fluctuation detection sensors, and piezoelectric vibration sensors; the thermistor temperature sensors are set at the reaction cavity, motor heat dissipation area, and circuit control unit of the biological instrument and equipment to collect temperature data; the voltage fluctuation detection sensors are set at the power input end, signal processing module, and actuator connection of the biological instrument and equipment to collect voltage data; the piezoelectric vibration sensors are set at the centrifugal component, liquid delivery pump, and robotic arm connection of the biological instrument and equipment to collect vibration data, and the preset data sampling period is set to be adjustable from 10 ms to 500 ms according to the working characteristics of the biological instrument and equipment, and synchronous sampling is performed on the operation data.

[0032] Furthermore, sort and integrate the operation data according to the time stamp to form time-series sample data with time-series characteristics, where the time-series sample data is stored in the form of a multi-dimensional matrix, with rows representing sampling times and columns representing sensor data types.

[0033] It should be noted that if the collected temperature data exceeds the equipment safety threshold, immediately trigger a temperature anomaly warning; if the temperature continues to rise and exceeds the critical value, automatically start the equipment cooling program and notify the maintenance personnel; if the voltage fluctuation amplitude exceeds 3 times the standard deviation, mark this time point as a voltage anomaly point; if more than 3 anomaly points are continuously detected, include the voltage instability situation in the fault feature vector; if the frequency characteristics of the vibration data deviate, it is determined as a potential mechanical fault; if accompanied by a temperature anomaly at the same time, increase the fault warning level and shorten the predicted fault occurrence time.

[0034] In the embodiment of the present application, the above step S2 includes:

[0035] Specifically, according to the historical fault correlation degree of the timing sample data, calculate the correlation coefficient between each data type and the occurrence of the fault;

[0036] Furthermore, normalize the correlation coefficient to obtain the corresponding feature weight coefficient, where the sum of the feature weight coefficients is one;

[0037] It should be noted that the feature weight coefficients include the temperature feature weight coefficient, the voltage feature weight coefficient, and the vibration feature weight coefficient, where the temperature feature weight coefficient assigns the importance weight to the temperature data, the voltage feature weight coefficient assigns the importance weight to the voltage data, and the vibration feature weight coefficient assigns the importance weight to the vibration data.

[0038] Even further, based on the feature weight coefficients, use a non-linear feature fusion function to weight various types of data, and introduce a non-linear adjustment coefficient to adjust the contribution degree of the timing sample data, and perform non-linear transformation through the feature mapping index;

[0039] Specifically, combine and stack the weighted timing sample data to generate fused feature data.

[0040] In the embodiment of the present application, the above step S3 includes:

[0041] Specifically, the pre-trained fault feature extraction model adopts an improved bidirectional long short-term memory network structure, and the improved bidirectional long short-term memory network structure includes multiple hidden layers and a self-attention mechanism module for capturing long-term dependence relationships and key time point features in the fused feature data;

[0042] Preferably, the specific formula of the fault feature extraction model is as follows:

[0043]

[0044] Where, F t is the fault feature vector extracted at time t, is the output of the forward LSTM, is the output of the backward LSTM, α i is the self-attention weight coefficient of the i-th feature dimension, W f is the weight matrix of the forward LSTM layer, W f is the weight matrix of the backward LSTM layer, b f is the bias vector of the forward LSTM layer, b b is the bias vector of the backward LSTM layer, W r is the weight of the residual connection layer, b r is the bias of the residual connection layer, xt The fused feature data input at time t, σ(*) is the activation function, and n is the feature dimension;

[0045] It should be noted that if a certain dimension component of the fault feature vector exceeds 2 times the standard deviation of the historical mean, a feature layer warning is triggered; if this component remains abnormal within 3 consecutive time windows, it is upgraded to an intermediate warning; if the Euclidean norm of the fault feature vector breaks through the preset threshold, a feature layer warning is triggered and the abnormal feature dimension is recorded; if several feature dimensions are abnormal at the same time, it is determined as a compound fault risk and the warning level is increased; if the self-attention weight coefficient distribution of the fault feature vector shows a highly concentrated phenomenon, it is identified as an abnormality of the key component; if the concentration area has a high degree of matching with the historical fault features, a targeted warning is triggered and it is recommended to check the corresponding hardware components.

[0046] Furthermore, the fused feature data is segmented into several time window sequences according to the sliding window method, and the length of the time window is determined according to the working cycle characteristics of the biological instrument equipment; the fused feature data within the time window is subjected to deep feature learning through the fault feature extraction model to extract the hidden features including time dimension information.

[0047] Even further, by introducing a residual connection mechanism, the original information of the fused feature data is retained, and the sensitivity of the fault feature extraction model to small fault symptoms is enhanced;

[0048] Specifically, the output layer of the fault feature extraction model maps the extracted deep features to a feature space of a fixed dimension through a non-linear transformation to generate a fault feature vector.

[0049] It should be noted that the dimension components of the fault feature vector correspond to the feature representations of different types of potential faults of the biological instrument equipment.

[0050] In the embodiment of the present application, the above step S4 includes:

[0051] Specifically, the pre-constructed fault prediction model adopts a hybrid prediction architecture combining the gradient boosting decision tree integration algorithm and the Markov chain prediction mechanism for multi-dimensional analysis and reasoning of the fault feature vector;

[0052] Preferably, the specific formula of the fault prediction model is as follows:

[0053]

[0054] Among them, P(F t+Δt ) is the fault occurrence probability at the future Δt moment, θ is the hybrid weight coefficient, β k is the weight coefficient of the kth decision tree, T k (F t ) is the prediction output of the kth decision tree, πi is the probability distribution of the current state i, is the probability that state i transfers to state j after time Δt, K is the number of decision trees, and m is the number of Markov states;

[0055] Furthermore, input the fault feature vector into the fault prediction model. Through the collaborative judgment of multiple decision trees, identify the abnormal patterns in the fault feature vector and determine the type of fault that occurs;

[0056] Even further, combine the determined type of fault that has occurred with the historical fault evolution data. Utilize the Markov chain prediction mechanism, and at the same time, based on the state transition probability matrix, deduce the fault development trend and predict the probability of fault occurrence;

[0057] Specifically, set different fault thresholds in the fault prediction model. For different components of the biological instrument device, when the occurrence probability of a certain type of fault is greater than the fault threshold, trigger the generation of a warning message, where the warning message includes the type of fault, the probability of fault occurrence, the estimated time of fault occurrence, and the scope of fault impact;

[0058] It should be noted that if P(F t+Δt ) ∈ [0.3, 0.6), it is a first-level warning, and mark a yellow warning sign on the device display interface, record the current type of fault and the occurrence probability, and increase the data sampling frequency of this component; if P(F t+Δt ) ∈ [0.6, 0.8), it is a second-level warning, and mark an orange warning sign on the device display interface, push the warning message to the remote monitoring platform, automatically reduce the device operation load, and at the same time calculate the estimated time of fault occurrence and evaluate the scope of fault impact; if P(F t+Δt ) ≥ 0.8, it is a third-level warning, mark a red warning sign on the device display interface, immediately send an emergency notice to the maintenance personnel, start the emergency protection mechanism, and record detailed fault feature data; if the prediction results show an increase in the fault probability three times in a row, the warning level is raised by one level; if the estimated time of fault occurrence is less than the device working cycle, the warning level is raised by one level.

[0059] Furthermore, the warning message is displayed through the human-machine interface of the instrument device and is synchronously pushed to the remote monitoring platform and the mobile terminal of the maintenance personnel to achieve the full-process closed-loop management of fault prevention.

[0060] Embodiment 2

[0061] This is the second embodiment of the present invention. This embodiment also provides a biological instrument device fault prediction system, including: an acquisition module, which is used to acquire the operation data of the biological instrument device, perform time-series sampling on the operation data through a preset data sampling period to obtain time-series sample data;

[0062] A generation module, which performs weighted fusion processing on the time-series sample data based on set feature weight coefficients to generate fusion feature data;

[0063] An extraction module, which is used to input the fusion feature data into a pre-trained fault feature extraction model to extract fault feature vectors;

[0064] A prediction module, which is used to input the fault feature vectors into a pre-constructed fault prediction model and output early warning information.

[0065] It should be noted that the technical solution of this system for predicting faults in a biological instrument device belongs to the same concept as the technical solution of the above-mentioned method for predicting faults in a biological instrument device. For the details not described in detail in the technical solution of the system for predicting faults in a biological instrument device in this embodiment, reference can be made to the description of the technical solution of the above-mentioned method for predicting faults in a biological instrument device.

[0066] The above-mentioned unit modules can be embedded in the processor in the computer device in hardware form or be independent of the processor, or can be stored in the memory in the computer device in software form, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0067] This embodiment also provides an electronic device, which includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a carrier network, NFC (Near Field Communication) or other technologies. When the computer program is executed by the processor, it realizes a method for evaluating the maximum access capacity of new energy considering the vulnerability of the power grid. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, a trackball or a touchpad set on the shell of the computer device, or an external keyboard, a touchpad or a mouse, etc.

[0068] This embodiment also provides a storage medium, on which a computer program is stored. When the program is executed by a processor, the following steps are implemented: collecting the operation data of a biological instrument device, performing sequential sampling on the operation data through a preset data sampling period to obtain sequential sample data; performing weighted fusion processing on the sequential sample data based on set feature weight coefficients to generate fusion feature data; inputting the fusion feature data into a pre-trained fault feature extraction model to extract a fault feature vector; and inputting the fault feature vector into a pre-constructed fault prediction model to output a warning message.

[0069] In summary, the present invention realizes multi-dimensional accurate modeling and trend deduction of the device operation state by constructing a biological instrument device fault prediction method combining fusion feature acquisition, deep fault feature extraction and multi-stage prediction models; improves the expression ability of key features through feature weight weighting and non-linear fusion; effectively captures potential abnormal patterns in sequential data based on an improved bidirectional long short-term memory network and self-attention mechanism; enhances the recognition and prediction ability of fault types and evolution paths by combining gradient boosting decision tree and Markov chain model; the differential fault threshold setting makes the warning response more targeted; the method as a whole improves the accuracy of fault detection and the timeliness of warning, effectively guarantees the stable operation of the device and reduces the maintenance cost, and has high engineering application and promotion value.

[0070] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A method for predicting failure of biological instrument equipment, characterized in that: include, Collecting operation data of biological instruments and equipment, and performing time series sampling on the operation data through a preset data sampling period to obtain time series sample data; Based on the set feature weight coefficient, weighted fusion processing is performed on the time series sample data to generate fused feature data; Inputting the fused feature data into a pre-trained fault feature extraction model to extract a fault feature vector; The fault feature vector is input into a pre-built fault prediction model to output warning information.

2. The method for predicting biological instrument failure according to claim 1, characterized in that: The pre-built fault prediction model adopts a hybrid prediction architecture that combines a gradient boosting decision tree ensemble algorithm with a Markov chain prediction mechanism.

3. The method for predicting biological instrument failure according to claim 2, characterized in that: Also includes, Inputting the fault feature vector into the fault prediction model, identifying the abnormal pattern in the fault feature vector through the collaborative judgment of multiple decision trees, and determining the type of fault that occurred; Combine the confirmed fault types with the historical fault evolution data, use the Markov chain prediction mechanism, and calculate the fault development trend based on the state transition probability matrix to predict the probability of fault occurrence; Differentiated fault thresholds are set in the fault prediction model. For different biological instrument equipment components, when the probability of a certain type of fault is greater than the fault threshold, the generation of the warning information is triggered, wherein the warning information includes the fault type, the probability of fault occurrence, the expected time of fault occurrence and the scope of fault impact.

4. The method for predicting biological instrument failure according to claim 3, characterized in that: Inputting the fused feature data into a pre-trained fault feature extraction model to extract a fault feature vector, including: The pre-trained fault feature extraction model adopts an improved bidirectional long short-term memory network structure; the improved bidirectional long short-term memory network structure includes multiple hidden layers and a self-attention mechanism module.

5. The method for predicting biological instrument failure according to claim 4, characterized in that: Also includes, The fused feature data is divided into a number of time window sequences according to the sliding window method, and the length of the time window is determined according to the working cycle characteristics of the biological instrument equipment; The fault feature extraction model is used to perform deep feature learning on the fused feature data within the time window to extract implicit features including time dimension information, wherein the output layer of the fault feature extraction model maps the extracted deep features to a feature space of fixed dimension through nonlinear transformation to generate a fault feature vector.

6. The method for predicting biological instrument failure according to claim 5, characterized in that: The method for obtaining the fusion feature data is: According to the historical fault correlation of time series sample data, the correlation coefficient between each data type and fault occurrence is calculated; Normalizing the correlation coefficient to obtain a corresponding feature weight coefficient, wherein the sum of the feature weight coefficients is one; Based on the feature weight coefficient, a nonlinear feature fusion function is used to perform weighted processing on various types of data, and a nonlinear adjustment coefficient is introduced to adjust the contribution of the time series sample data, and a nonlinear transformation is performed through a feature mapping index; The weighted time series sample data are combined and superimposed to generate fused feature data.

7. The method for predicting biological instrument failure according to claim 6, characterized in that: The method for obtaining the time series sample data is: During the operation of the biological instrument, operating data is collected by using a plurality of sets of sensors, and the operating data is sampled regularly according to a preset data sampling period, wherein the operating data includes temperature data, voltage data and vibration data; The operation data are sorted and integrated according to timestamps to form time series sample data with time series characteristics, wherein the time series sample data is stored in a multi-dimensional matrix form, with rows representing sampling moments and columns representing sensor data types.

8. A biological instrument equipment failure prediction system, based on the biological instrument equipment failure prediction method according to any one of claims 1 to 7, characterized in that: include, The acquisition module is used to acquire the operation data of the biological instrument and equipment, and perform time series sampling on the operation data through a preset data sampling period to obtain time series sample data; A generation module performs weighted fusion processing on the time series sample data based on a set feature weight coefficient to generate fused feature data; An extraction module, used for inputting the fused feature data into a pre-trained fault feature extraction model to extract a fault feature vector; The prediction module is used to input the fault feature vector into a pre-built fault prediction model and output warning information.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the biological instrument equipment failure prediction method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the biological instrument equipment failure prediction method according to any one of claims 1 to 7 are implemented.

Images