A method and system for monitoring the operation of a medical printer

Abstract

The present application belongs to the technical field of equipment state monitoring, and particularly relates to a running monitoring method and system based on a medical printer, which method comprises: collecting multi-source sensor data and extracting a multi-dimensional feature vector therefrom; based on a pre-constructed normal state model, evaluating the degree to which the multi-dimensional feature vector deviates from the normal state model to obtain a degradation drift distance; calculating a degradation feature signature and a degradation trend index; based on the degradation feature signature, determining the occurrence probability of each preset failure mode, and combining the degradation trend index to calculate a predictive maintenance score, and finally giving a warning according to the predictive maintenance score. By analyzing the degradation trajectory composed of continuous task state points, the present application can not only evaluate the degradation degree of the medical printer equipment, but also effectively identify the potential failure mode thereof.

Description

Technical Field

[0001] This invention relates to the field of equipment condition monitoring technology. More specifically, this invention relates to a method and system for monitoring the operation of a medical printer. Background Technology

[0002] In modern medical environments, medical printers are critical information output devices that ensure smooth diagnostic and treatment processes, primarily used for printing medical imaging films, patient information labels, and other information. Compared to ordinary printers, medical printers have unique application characteristics: First, they require extremely high operational reliability; any unexpected downtime can directly lead to delays in diagnosis and treatment, or even trigger medical risks, with consequences far more serious than the inability to print office documents. Second, their workload and consumables are special; for example, when printing medical imaging films, the heating module needs to operate stably at a specific temperature for extended periods, placing more stringent demands on energy consumption and the heat dissipation system. Finally, the data is highly correlated; abnormal printing tasks are not merely equipment problems but may also be linked to specific patient information and treatment procedures.

[0003] Existing technologies detect abnormal states by monitoring sensor data during printer operation, such as power consumption and temperature. One common technique is to extract a set of features from the energy consumption data after each printing job to form a task feature vector. Then, anomaly detection algorithms such as Local Outlier Factor (LOF) are used to calculate the degree of deviation of the task feature vector from a large number of historical normal tasks to determine whether the current task is abnormal.

[0004] However, current technologies analyze each print job in an isolated and static manner, only able to determine whether the current job is abnormal, but unable to effectively monitor the gradual performance degradation of the equipment. In practical applications, many serious failures do not occur suddenly, but are the result of long-term, slow wear or aging accumulation of critical components such as motors, heating modules, and transmission components. This gradual degradation is not reflected in energy consumption data as a single, huge anomaly, but rather as a continuous, slow increasing trend across multiple consecutive print jobs. Furthermore, even if current technologies detect a significant change in the trend, they cannot determine which specific type of component is deteriorating. Summary of the Invention

[0005] To address the technical problems that existing technologies cannot effectively monitor the progressive performance degradation of equipment and cannot diagnose the specific root causes of failures, this invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides an operation monitoring method based on a medical printer, comprising: collecting multi-source sensor data during a single printing task of the medical printer and extracting a multi-dimensional feature vector from it, wherein the multi-dimensional feature vector integrates the global resource consumption, dynamic operation mode, and cross-domain physical quantity correlation state information of the printing task; evaluating the degree of deviation of the multi-dimensional feature vector from the normal state model based on a pre-constructed normal state model to obtain a drift vector representing the direction of deviation and a degradation drift distance representing the magnitude of deviation; calculating a degradation feature signature based on the drift vector of each task within a time window consisting of multiple consecutive printing tasks, and calculating a degradation trend index based on the degradation drift distance of each task, wherein the degradation feature signature represents the recent dominant degradation direction of the device, and the degradation trend index represents the recent severity of degradation of the device; determining the occurrence probability of each preset fault mode based on the degradation feature signature, weighting the risk cost of each fault mode using the occurrence probability, and combining it with the degradation trend index to calculate a predictive maintenance score, and finally issuing an early warning based on the predictive maintenance score.

[0007] This invention, by constructing a normal state model and analyzing deviation trajectories, can understand the degradation process of equipment. At the same time, the introduction of hierarchical features and cross-domain correlation features enables the model to capture the equipment status from richer dimensions, effectively identify the root cause of anomalies, and provide a direct basis for precise maintenance. It solves the problem that existing technologies can only determine whether there is an anomaly but cannot diagnose the specific root cause of the fault.

[0008] Preferably, the global resource consumption status information includes total task duration, peak power, average power, total power consumption, and average operating temperature; the dynamic operation status information is obtained by calculating the energy ratio through discrete wavelet transform of the power time series; the cross-domain physical quantity correlation status information includes power consumption-temperature rise time delay and power consumption-vibration correlation.

[0009] This invention constructs hierarchical features to monitor device status from multiple dimensions, providing rich data support that is closely related to physical meaning for subsequent accurate diagnosis.

[0010] Preferably, the normal state model is constructed based on principal component analysis. The step of evaluating the degree to which the multidimensional feature vector deviates from the normal state model includes: projecting the multidimensional feature vector onto a hyperplane composed of the principal components of the normal state model and reconstructing it back to the original dimensional space to obtain a reconstructed feature vector; taking the difference between the multidimensional feature vector and the reconstructed feature vector as the drift vector; and taking the Euclidean distance between the multidimensional feature vector and the reconstructed feature vector as the degradation drift distance.

[0011] This invention, by constructing a normal state model and analyzing deviation trajectories, can more deeply monitor the degradation process of equipment, thereby improving the realism of the diagnostic model.

[0012] Preferably, the degenerate feature signature satisfies the expression: ;in, Let represent the degradation signature of the i-th printing task, N be the length of the time window, and p be the task index within the time window. For the task The drift vector, To prevent extremely small positive numbers with a denominator of 0.

[0013] Preferably, the degradation trend index satisfies the expression: ;in, Let be the degradation trend index for the i-th printing task. This is the arithmetic mean of all degradation drift distances within the time window. The degradation drift distance sequence within the time window. The slope of the linear regression of the degradation drift distance sequence within the time window. This is the trend amplification factor. It is the hyperbolic tangent function.

[0014] Preferably, determining the occurrence probability of each preset fault mode includes: matching the degraded feature signature with a preset fault signature knowledge base, wherein the fault signature knowledge base stores Gaussian mixture models for multiple known fault modes; calculating the likelihood probability of the degraded feature signature generated by the Gaussian mixture model for each fault mode, and combining it with Bayes' theorem to calculate the posterior probability of each preset fault mode as the occurrence probability.

[0015] Preferably, the predictive maintenance score satisfies the expression: ;in, For the predictive maintenance score of the i-th printing task, Let be the degradation trend index for the i-th printing task. This represents the degenerate feature signature of the i-th printing task. This indicates the first fault signature in the preset fault signature knowledge base. Types of failure modes For the first The probability of occurrence of each failure mode. For the first Risk cost coefficient for each failure mode.

[0016] This invention introduces a fault diagnosis based on a probability model and a scoring mechanism that combines risk and cost, making the final early warning decision more accurate.

[0017] Preferably, the step of issuing an early warning based on the predictive maintenance score includes: normalizing the predictive maintenance score to obtain a normalized predictive maintenance score; and comparing the normalized predictive maintenance score with a preset attention threshold and a critical threshold to issue a graded early warning.

[0018] This invention normalizes the predictive maintenance score, thereby fixing and standardizing the early warning threshold, avoiding decision-making difficulties caused by the scale of the original data, and making the early warning behavior more stable and reliable.

[0019] Preferably, the normalized predictive maintenance score satisfies the following relation: ;in, To maintain the normalized predictive score for the i-th printing task, Let be the upper limit of the normalization interval, and k be the kurtosis parameter. The midpoint parameter is used to transform the center position for determining the normalized center position.

[0020] Secondly, the present invention provides an operation monitoring system based on a medical printer, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned operation monitoring method based on a medical printer is implemented.

[0021] By adopting the above technical solution, a computer program for the operation monitoring method based on a medical printer is generated and stored in a memory so that it can be loaded and executed by a processor. This allows for the creation of a terminal device based on the memory and processor, making it convenient to use.

[0022] This invention overcomes the shortcomings of existing technologies that can only detect anomalies in isolated single tasks. By constructing a normal state model and analyzing the evolution trajectory of feature vectors from continuous tasks, it can extract degradation feature signatures that characterize the dominant direction of equipment degradation. This achieves a leap from judging whether there is an anomaly to diagnosing where the anomaly is, providing a direct technical basis for accurately locating and eliminating potential fault roots.

[0023] Furthermore, the intelligent early warning decision-making mechanism proposed in this invention not only assesses the severity of the degradation trend, but also uses the probability of failure to weight the risk costs of different failures, and normalizes the final score, making the early warning decision more accurate and reliable. Attached Figure Description

[0024] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent upon reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein:

[0025] Figure 1 This is a flowchart illustrating an operational monitoring method based on a medical printer according to the present invention;

[0026] Figure 2 This is a schematic diagram illustrating the evolution trend of monitoring indicators. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0029] This invention discloses a method for monitoring the operation of a medical printer, referring to... Figure 1 This includes steps S1-S4:

[0030] S1. Collect multi-source sensor data during a single printing task of the medical printer and extract multi-dimensional feature vectors from it.

[0031] In an alternative embodiment, high-precision sensors can be deployed at key locations on the medical printer. For example, a Hall effect current sensor can be deployed on the main power line to monitor power consumption; a sheathed thermistor can be deployed near the heating roller or printhead to monitor core temperature; and a triaxial MEMS accelerometer can be deployed at key structural points on the casing to monitor device vibration.

[0032] It is worth noting that the control system must ensure that all data streams are collected synchronously at the same preset sampling frequency, for example, 100Hz. The monitoring system can obtain the context data of each print job in real time from the printer driver or host computer system through the interface, including the unique job ID, the type of print file, print quality settings, etc.

[0033] In this optional embodiment, the acquired raw data stream is first preprocessed, for example, by using a Kalman filter for online filtering. Then, the system precisely segments the continuous data stream according to the task start / end timestamps to extract independent data segments that strictly correspond to a single printing task.

[0034] Furthermore, for each segmented print task event data, the system can extract a multi-dimensional feature vector. This vector integrates three dimensions of state information: global resource consumption, dynamic operation mode, and cross-domain physical quantity correlation of the print task.

[0035] Specifically, global resource consumption describes the overall resource consumption of a task, including total task duration, peak power, average power, total power consumption, and average operating temperature; dynamic operating mode is used to capture the microscopic shape and dynamic characteristics inside the energy consumption curve. Discrete wavelet transform can be used to decompose the power time series into 4 layers, and then calculate the energy proportion of detail coefficients and approximation coefficients of each layer.

[0036] Cross-domain physical quantity correlation is used to describe the coupling relationship between different physical quantities. The corresponding state information includes power consumption-temperature rise time delay information and power consumption-vibration correlation information.

[0037] Among them, the power consumption-temperature rise lag information measures the time difference between when the printer starts consuming a large amount of power, for example, when the heating roller starts working, and when the temperature of its core components rises accordingly. In this case, if the heating module efficiency decreases or the heat dissipation system becomes blocked, the device needs more time to reach the target operating temperature, which will cause the power consumption-temperature rise lag value to increase systematically. Therefore, monitoring the trend of the lag can effectively provide early warning of faults related to thermal efficiency. The power consumption-vibration correlation information measures the degree of correlation between the printer's electrical fluctuations and mechanical vibrations, reflecting the coupling between the electrical and mechanical systems of the device. In this case, if the device ages, the parts become loose, or the electrical components are unstable, it may cause electrical noise to trigger abnormal mechanical resonance. In this case, the machine's vibration will change with the power fluctuations, resulting in a significant increase in the correlation between power consumption and vibration.

[0038] Specifically, the power consumption-temperature rise time delay information can be obtained by calculating the cross-correlation function between the power consumption time series and the temperature change time series, and taking the time delay corresponding to the maximum correlation between the two as the time delay value; the power consumption-vibration correlation can be obtained by extracting the high-frequency components from the power series and vibration series respectively, and then calculating the Pearson correlation coefficient between the two high-frequency components.

[0039] In this way, by extracting multi-dimensional feature information, comprehensive and in-depth data support can be provided for subsequent equipment status assessment and fault diagnosis.

[0040] S2. Based on the pre-built normal state model, evaluate the degree to which the multidimensional feature vector deviates from the normal state model to obtain the drift vector representing the direction of deviation and the degradation drift distance representing the magnitude of deviation.

[0041] In an optional embodiment, for each task type, a model can be constructed using principal component analysis (PCA) based on its large number of historical normal feature vectors, thereby obtaining its f principal components. These f principal components define an f-dimensional hyperplane, which can be used as an approximation of the normal state manifold. In this scheme, this hyperplane defined by the principal components is used as the preset normal state model. Any new printing task is considered normal if its feature vector falls on or is close to this hyperplane; conversely, if it deviates significantly, it is considered to have experienced abnormal degradation.

[0042] For a new task i, its feature vector is projected onto the hyperplane formed by the principal components of the normal state model and reconstructed back into the original dimensional space to obtain the reconstructed feature vector. In this scheme, the difference between the multidimensional feature vector and the reconstructed feature vector is used as the drift vector; the Euclidean distance between the multidimensional feature vector and the reconstructed feature vector is used as the degradation drift distance.

[0043] For example, suppose the feature vector of the current task i is [10, 5, 8], and the corresponding reconstruction vector calculated by the PCA model is [10.2, 5.1, 7.5], then the drift vector is [-0.2, -0.1, 0.5]; the corresponding degradation drift distance is... .

[0044] In this way, by comparing the current state with the normal model, we can obtain a deviation index that includes information on direction and magnitude, laying the foundation for subsequent trend analysis.

[0045] S3. Within a time window consisting of multiple consecutive printing tasks, calculate the degradation feature signature based on the drift vector of each task, and calculate the degradation trend index based on the degradation drift distance of each task.

[0046] In an optional embodiment, within a time window consisting of multiple consecutive print jobs, a degradation feature signature can be calculated based on the drift vector of each job to characterize the recent dominant degradation direction of the printing device. The degradation feature signature satisfies the expression:

[0047]

[0048] in, This represents the degradation feature signature of the i-th printing task, where N is the length of the time window. For task indexing within the time window, For the task The drift vector, To prevent extremely small positive numbers with a denominator of 0, for example, the value is taken as... . It has the same dimensions as the drift vector, and the magnitude of each component represents the dominant direction of recent printing device state degradation in the corresponding feature dimension.

[0049] For example, suppose the time window length N is 3, then The values ​​are 0, 1, and 2, and the drift vectors for the three most recent printing tasks are as follows: The range is [-0.2, -0.1, 0.5]. The values ​​are [-0.22, -0.09, 0.55]. The corresponding [-0.18, -0.12, 0.48] values ​​are: The degenerate drift distance values ​​are 0.548, 0.602 and 0.531, respectively. After normalizing the three drift vectors, we get [-0.365,-0.183,0.912], [-0.365,-0.149,0.914] and [-0.339,-0.226,0.904], respectively. The average value of these three normalized vectors is [-0.356,-0.186,0.910].

[0050] Next, a degradation trend index can be calculated based on the degradation drift distance of each task to characterize the recent degradation severity of the equipment. The degradation trend index satisfies the following expression:

[0051]

[0052] in, Let be the degradation trend index for the i-th printing task. This is the arithmetic mean of all degradation drift distances within the time window. The degradation drift distance sequence within the time window. The slope of the linear regression of the degradation drift distance sequence within the time window. This is the trend amplification factor; for example, it is set to 2. It is the hyperbolic tangent function.

[0053] For example, if the degradation drift distance sequence for the three most recent tasks is {0.531, 0.602, 0.548}, then the arithmetic mean of all degradation drift distances within the time window is 0.560. Performing linear regression on the sequence {0.531, 0.602, 0.548} yields the slope... The value is 0.0085, therefore the corresponding for As can be seen, the final results not only reflect the current trend of degradation, but are also amplified by the presence of a slight positive growth trend.

[0054] Thus, by analyzing the data within the time window, it is possible to extract diagnostically significant degradation direction signatures and trend indices characterizing the rate of deterioration from isolated single-task deviations.

[0055] S4. Based on the degradation feature signature, determine the occurrence probability of each preset failure mode, use the occurrence probability to weight the risk cost of each failure mode, and combine it with the degradation trend index to calculate the predictive maintenance score. Finally, issue an early warning based on the predictive maintenance score.

[0056] In an optional embodiment, a fault signature knowledge base can be preset, which stores Gaussian mixture models for various known fault modes. When a new degradation feature signature is calculated, the system matches it with the Gaussian mixture model in the knowledge base to obtain the likelihood probability, and calculates the posterior probability of each preset fault mode using Bayes' theorem. This probability is the occurrence probability.

[0057] Furthermore, a risk-cost coefficient can be defined for each failure mode. The risk cost of each failure mode is weighted using the probability of occurrence, and a predictive maintenance score is calculated by combining the degradation trend index. The predictive maintenance score satisfies the expression:

[0058]

[0059] in, For the predictive maintenance score of the i-th printing task, Let be the degradation trend index for the i-th printing task. This represents the degenerate feature signature of the i-th printing task. This indicates the first fault signature in the preset fault signature knowledge base. Types of failure modes For the first The probability of occurrence of each failure mode. For the first Risk cost coefficient for each failure mode.

[0060] For example, the risk cost factor for heating module aging is 200, and the risk cost factor for motor wear is 500, due to the calculated... Given a probability of 0.57, and assuming the probability of heating module aging is 0.85 and the probability of motor wear is 0.1, the weighted risk cost is: Predictive maintenance score is .

[0061] Furthermore, to facilitate the establishment of fixed and intuitive early warning thresholds, the unbounded predictive maintenance score is mapped to a bounded normalized predictive maintenance score, for example, in the interval [0, 20]. The predictive maintenance score can be normalized, and the normalized predictive maintenance score satisfies the following relationship:

[0062]

[0063] in, To maintain the normalized predictive score for the i-th printing task, is the upper limit of the normalization interval, for example, a value of 20; k is the kurtosis parameter, for example, 0.05. The midpoint parameter is used to determine the normalized center position; for example, it is set to 100.

[0064] In this optional embodiment, a graded warning can be issued based on the final predicted maintenance score. Specifically, a attention threshold of 5 and a critical threshold of 10 can be preset, and the normalized predicted maintenance score, attention threshold, and critical threshold can be compared respectively. When the final normalized predicted maintenance score exceeds the attention threshold, staff can be reminded to pay attention to data anomalies in a timely manner. When the final normalized predicted maintenance score exceeds the critical threshold, the system will issue an alarm message to remind staff to perform a shutdown inspection in a timely manner.

[0065] like Figure 2 The diagram illustrates the evolution trend of monitoring indicators. The horizontal axis represents the chronological order of printing tasks, and the vertical axis represents the normalized predictive maintenance score, which is normalized to between 0 and 20. The solid red line represents the normalized predictive maintenance score, the dashed blue line represents the degradation drift distance, the dashed orange line represents the degradation trend index, the dashed green line is the attention threshold, and the dashed dark red line is the critical threshold. It can be seen that after the device degradation begins in the 30th task, the solid red line rises smoothly and rapidly, crossing the green attention threshold and the dark red critical threshold, demonstrating the tiered warning process. The degradation drift distance shows a slow increasing trend, while the degradation trend index, being a non-linear method, has a steeper growth slope.

[0066] In this way, by introducing risk costs and normalizing the final score, stable, intuitive, and actionable intelligent early warning information can be output.

[0067] This invention also discloses an operation monitoring system based on a medical printer, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement an operation monitoring method based on a medical printer according to the present invention.

[0068] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0069] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

[0070] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A method for monitoring the operation of a medical printer, characterized in that, include: Multi-source sensor data is collected during a single printing task of a medical printer, and multi-dimensional feature vectors are extracted from them. The multi-dimensional feature vectors integrate the global resource consumption, dynamic operation mode, and state information of cross-domain physical quantity correlation of the printing task. Based on a pre-built normal state model, the degree to which the multidimensional feature vector deviates from the normal state model is evaluated to obtain a drift vector representing the direction of deviation and a degradation drift distance representing the magnitude of deviation. Within a time window consisting of multiple consecutive printing tasks, a degradation feature signature is calculated based on the drift vector of each task, and a degradation trend index is calculated based on the degradation drift distance of each task. The degradation feature signature represents the dominant degradation direction of the device in the near future, and the degradation trend index represents the severity of degradation of the device in the near future. Based on the degradation feature signature, the probability of occurrence of each preset failure mode is determined. The risk cost of each failure mode is weighted using the probability of occurrence, and a predictive maintenance score is calculated by combining the degradation trend index. Finally, an early warning is issued based on the predictive maintenance score.

2. The method for monitoring the operation of a medical printer according to claim 1, characterized in that, The global resource consumption status information includes total task duration, peak power, average power, total power consumption, and average operating temperature; the dynamic operation status information is obtained by calculating the energy ratio through discrete wavelet transform of the power time series; the cross-domain physical quantity correlation status information includes power consumption-temperature rise time delay and power consumption-vibration correlation.

3. The method for monitoring the operation of a medical printer according to claim 1, characterized in that, The normal state model is constructed based on principal component analysis. The steps for evaluating the degree to which the multidimensional eigenvectors deviate from the normal state model include: The multidimensional feature vector is projected onto a hyperplane composed of the principal components of the normal state model and reconstructed back into the original dimensional space to obtain the reconstructed feature vector. The difference between the multidimensional feature vector and the reconstructed feature vector is used as the drift vector; The Euclidean distance between the multidimensional feature vector and the reconstructed feature vector is taken as the degradation drift distance.

4. The method for monitoring the operation of a medical printer according to claim 1, characterized in that, The degenerate feature signature satisfies the expression: in, This represents the degradation feature signature of the i-th printing task, where N is the length of the time window. For task indexing within the time window, For the task The drift vector, To prevent extremely small positive numbers with a denominator of 0.

5. The method for monitoring the operation of a medical printer according to claim 1, characterized in that, The degradation trend index satisfies the expression: in, Let be the degradation trend index for the i-th printing task. This is the arithmetic mean of all degradation drift distances within the time window. The degradation drift distance sequence within the time window. The slope of the linear regression of the degradation drift distance sequence within the time window. This is the trend amplification factor. It is the hyperbolic tangent function.

6. The method for monitoring the operation of a medical printer according to claim 1, characterized in that, The determination of the occurrence probability of each preset fault mode includes: The degradation feature signature is matched with a preset fault signature knowledge base, which stores Gaussian mixture models for various known fault modes. The likelihood probability of the degradation feature signature generated by the Gaussian mixture model for each fault mode is calculated, and the posterior probability of each preset fault mode is calculated as the occurrence probability by combining the Bayesian formula.

7. A method for monitoring the operation of a medical printer according to claim 1 or 6, characterized in that, The predictive maintenance score satisfies the expression: in, For the predictive maintenance score of the i-th printing task, Let be the degradation trend index for the i-th printing task. This represents the degenerate feature signature of the i-th printing task. This indicates the first fault signature in the preset fault signature knowledge base. Types of failure modes For the first The probability of occurrence of each failure mode. For the first Risk cost coefficient for each failure mode.

8. The method for monitoring the operation of a medical printer according to claim 7, characterized in that, The provision of early warnings based on the predictive maintenance score includes: The predictive maintenance score is normalized to obtain the normalized predictive maintenance score. The normalized predictive maintenance score is compared with the preset attention threshold and critical threshold to perform graded early warning.

9. The method for monitoring the operation of a medical printer according to claim 8, characterized in that, The normalized predictive maintenance score satisfies the following relation: in, To maintain the normalized predictive score for the i-th printing task, Let be the upper limit of the normalization interval, and k be the kurtosis parameter. The midpoint parameter is used to transform the center position for determining the normalized center position.

10. A medical printer-based operation monitoring system, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement a method for monitoring the operation of a medical printer according to any one of claims 1-9.

Images