Automatic analysis device

Abstract

This automated analysis device: dispenses pure water into a reaction vessel 1 and performs water blank measurement prior to dispensing a sample into the reaction vessel and performing sample measurement; adjusts light absorbency data relating to a reaction solution of a reagent and the sample measured by means of the sample measurement, by means of water blank light absorbency data measured by means of the water blank measurement; and, concurrent with execution of analysis sequencing to analyze the sample on the basis of the adjusted light absorbency data, determines a water quality abnormality of reaction tank circulating water 112 on the basis of the water blank light absorbency data measured by means of the water blank measurement. In this way, a water quality abnormality of the reaction tank circulating water can be detected without adding a dedicated mechanism or a verification operation.

Description

[Technical field]

[0001] The present invention relates to an analyzer for clinical testing that performs qualitative and quantitative analysis of biological samples such as blood and urine, and in particular to an automatic analyzer that is provided with a reaction chamber circulation flow path. [Background technology]

[0002] For example, in an automatic biochemical analyzer, in order to perform component analysis of biological samples such as serum or urine, the test sample is reacted with a reagent in a reaction vessel, and the resulting changes in color and turbidity are optically measured at regular intervals using a photometric unit such as a spectrophotometer. To prevent the rate of chemical reaction between the sample and the reagent in the reaction vessel from varying due to temperature fluctuations, the reaction vessel is immersed in constant-temperature water with a stable temperature. This constant-temperature water is circulated via a heater and a cooler to keep the reaction vessel at a stable temperature.

[0003] However, if air bubbles are mixed into the constant temperature water and cross the optical axis during the measurement of absorbance data of the sample in the reaction vessel, the light will be scattered or absorbed by the air bubbles, resulting in a different value for the measured absorbance data from that of a normal chemical reaction. For this reason, automatic analyzers are equipped with a degassing unit in the circulation flow path to suppress the generation of air bubbles, but if the degassing performance of the degassing unit deteriorates, air bubbles will be generated in the circulating water.

[0004] However, the presence of air bubbles does not necessarily cause the analysis results to be abnormal. If the test item involves obtaining absorbance at two wavelengths and calculating the concentration using the difference, even if the measured absorbance is affected by air bubbles, the effect of air bubbles can be suppressed by taking the difference, so it may not be possible to notice an increase in air bubbles from the analysis results.

[0005] Patent Document 1 discloses a method for determining the presence or absence of disturbances related to the photometer of a reaction vessel by repeatedly measuring light of a wavelength that is not absorbed by the reaction liquid. Examples of disturbances include scratches, dirt, and air bubbles on the reaction vessel.

[0006] Patent Document 2 discloses that absorbance data in the same reaction vessel is measured at multiple points at specified time intervals in a chronological order, and if a spike-like change in absorbance is detected, alarm information is added to the measurement data, and based on the frequency of alarm occurrence, it is determined that this is due to an increase in foreign matter such as air bubbles or scale in the thermostatic water.

[0007] Patent document 3 discloses that during operation of an automatic analyzer, reaction tank water is dispensed into a reaction container, the absorbance of the reaction tank water is measured to determine the water quality of the reaction tank water, and the reaction tank water is supplied or discharged depending on the water quality determination. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Japanese Patent Application Publication No. 08-005562 [Patent Document 2] JP 2000-039400 A [Patent Document 3] JP 2020-159804 A Summary of the Invention [Problem to be solved by the invention]

[0009] The main cause of air bubbles being mixed into the constant temperature water is a decrease in the degassing performance of the degassing unit for the constant temperature water. When the concentration of dissolved oxygen in the reaction tank circulating water increases due to a decrease in the degassing performance of the degassing unit, it leads to a state where air bubbles are likely to be generated. However, the degassing unit is usually located inside the device and is difficult to access frequently, and furthermore, since the degassing unit is connected to a flow path, it is difficult to check the internal condition of the degassing unit.

[0010] Patent Document 1 and Patent Document 2 are for determining abnormality from the change in absorbance at wavelengths without absorption in a reaction solution obtained by chemically reacting a sample with a reagent. Figure 14 shows the reaction process of the sub-wavelength of creatinine. Even if the analysis item is not a photometric target and a wavelength that is supposed to have no absorption is measured, a change in absorbance may occur depending on the components contained in the specimen. There is no problem if a stable value is measured without absorption, as in graph 501 (△), but depending on the components contained in the specimen, a change in absorbance may occur as in graph 502 (◆). In this way, in the reaction solution between the sample and the reagent, a change in absorbance may occur due to an unintended side reaction. For this reason, it is difficult to accurately determine abnormality.

[0011] Patent Document 3 determines whether the water quality of the reaction tank circulating water is abnormal by measuring the reaction tank water (or, as an alternative, a standard sample with known absorbance). However, during operation, a dedicated measurement must be performed to determine the water quality of the reaction tank circulating water, which may reduce the testing efficiency of the automatic analyzer.

[0012] The object of the present invention is to realize an automatic analyzer that can detect water quality abnormalities that tend to cause air bubbles to form in the reaction tank circulating water by using absorbance values ​​obtained during operation, without the need for additional dedicated mechanisms or measurements. [Means for solving the problem]

[0013] An automatic analyzer according to one embodiment of the present invention comprises a reaction vessel, a reaction disk that holds reaction vessels immersed in the reaction vessel circulating water in the reaction vessel, a circulation flow path that adjusts the temperature of the reaction vessel circulating water from the reaction vessel and returns it to the reaction vessel, a degassing device provided in the circulation flow path that degasses the reaction vessel circulating water, a light source lamp, a photometer having a light detector that detects light irradiated from the light source lamp and transmitted through the reaction vessel, and a control unit that controls the analysis sequence of the sample and analyzes the sample, and the control unit performs a water blank measurement by dispensing pure water into the reaction vessel prior to dispensing the sample into the reaction vessel and performing sample measurement, corrects the absorbance data of the reaction solution between the sample and the reagent measured by the sample measurement with the water blank absorbance data measured by the water blank measurement, and analyzes the sample based on the corrected absorbance data, and the control unit judges water quality abnormalities in the reaction vessel circulating water based on the water blank absorbance data measured by the water blank measurement. Effect of the Invention

[0014] According to the present invention, it is possible to detect an abnormality in the water quality of the circulating water in the reaction tank without adding a dedicated mechanism or confirmation operation. Problems, configurations, and effects other than those described above will become apparent from the description of the following embodiments. [Brief description of the drawings]

[0015] [Figure 1] FIG. 1 is a diagram illustrating the configuration of an automatic analyzer. [Figure 2A] 2 is an example of a reaction water circulation flow path. [Figure 2B] 1 is a configuration example of a degassing device. [Diagram 3] This is a characteristic diagram of the water blank absorbance range (normal). [Figure 4] This is a characteristic diagram of the water blank absorbance range (abnormal). [Diagram 5] FIG. 13 is a graph showing the correlation between the dissolved oxygen concentration and the average value of the water blank absorbance range. [Figure 6] FIG. 1 is a graph showing the correlation between dissolved oxygen concentration and water blank absorbance range standard deviation. [Figure 7]13 is a flowchart for determining water quality abnormality. [Figure 8] FIG. 1 is a functional block diagram for creating / updating a learning model. [Figure 9A] 1 is an example of maintenance history data. [Figure 9B] 4 is an example of alarm information data. [Figure 10] 1 is an example of a learning model using a neural network. [Figure 11] This is an example of a GUI. [Figure 12] This is an example of a GUI. [Figure 13] This is an example of a GUI. [Figure 14] FIG. 1 is a diagram for explaining a problem with the conventional technology. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0016] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

[0017] The operation of the automatic analyzer will be described based on the configuration shown in Fig. 1. The control of the analysis sequence in the automatic analyzer and the analysis process in response to a detection signal from a photometer (to be described later) are carried out by a control unit 22.

[0018] First, the reaction vessels 1 are washed by the washing mechanism 17 and the pure water washing pump 15. After that, pure water is dispensed into each reaction vessel 1, and measurement is performed with a photometer consisting of a light source lamp 3, a spectroscopic diffraction grating 4, and a photodetector 5 to obtain photometric data (water blank value). After that, the pure water is removed from the reaction vessels 1.

[0019] Next, the sample in the sample cup 7 on the sample disk 8 is dispensed into the reaction vessel 1 on the reaction disk 2 using the sample dispensing mechanism 11 and the sample dispensing pump 14. Similarly, the reagent in the reagent bottle 9 on the reagent disk 10 is dispensed into the reaction vessel 1 by the reagent dispensing mechanism 12 and the reagent dispensing pump 16, and both are stirred using the stirring mechanism 13. The color of the reaction liquid in the reaction vessel 1 due to the chemical reaction is measured by a photometer at regular time intervals determined according to the rotation of the reaction disk 2. To promote a stable chemical reaction, the temperature of the reaction vessel 1 is kept constant by the constant temperature water supply device 6. After the measurement is completed, the reaction vessel 1 is washed by the washing mechanism 17 and the pure water pump 15 for washing, and is used for the measurement of the next sample.

[0020] The configuration of the photometer will be explained. Light emitted from a light source lamp 3 passes through a reaction vessel 1 and is separated into each wavelength component by a spectroscopic diffraction grating 4. At the destination of the separated light, the required number of photodetectors 5 are installed at positions corresponding to the wavelengths required for measurement. The photodetectors 5 photoelectrically convert the incident light and generate a current proportional to the amount of light, i.e., the number of photons of light. This current is called a photocurrent. To obtain an electrical signal proportional to the concentration of the component to be measured contained in the sample, the digital value is converted by an AD conversion unit 20, stored in a storage device 23 via an interface bus 27, and analyzed and processed by a control unit 22.

[0021] The absorbance data corrected by the water blank value is used to calculate the concentration of the sample. Because absorbance fluctuates due to scratches or dirt on the sample container and changes in the light intensity of the light source lamp, the water blank value is generally obtained immediately before the sample measurement.

[0022] In this way, in the automatic analyzer, the reaction vessel 1 is repeatedly used in the order of cleaning, water blank measurement, and sample measurement. The water blank measurement is performed when blank water (purified water, pure water) is dispensed into the reaction vessel 1 and the reaction vessel 1 passes in front of the photometer due to the rotation of the reaction disk 2. Purified water has low wettability, so air bubbles are easily attached to the wall of the reaction vessel 1. If air bubbles are attached to the inner wall of the reaction vessel, the correct absorbance cannot be obtained, so in the water blank measurement, water is discharged into the reaction vessel 1 to prevent air bubbles from being generated, and no stirring operation is performed. In addition, the reaction vessel 1 is kept at a constant temperature of 37°C by circulating reaction tank circulating water. When the temperature of the blank water rises, dissolved oxygen becomes air bubbles and is easily generated, so the measurement is completed in a short time. However, it is common to perform multiple measurements in consideration of the variation in the measured values. In the water blank absorbance measurement, stable measured values ​​can be obtained because there are no factors in the reaction vessel that cause the absorbance to fluctuate.

[0023] FIG. 2A shows an example of a circulation flow path of the reaction tank circulating water. The reaction vessel 1 is arranged on a circumference in the reaction vessel 18. The reaction vessel 18 is filled with the reaction tank circulating water 112. The reaction tank circulating water 112 is circulated by the circulation pump 105 through the reaction vessel 18, the degassing device 102, the cooling device 104, the circulation pump 105, and the heating device 106 in this order. In order to stably carry out the chemical reaction of the reaction liquid 111 in the reaction vessel 1, the temperature of the reaction tank circulating water 112 is kept constant by the action of the cooling device 104 and the heating device 106. The water level of the reaction tank circulating water 112 drops due to evaporation, overflow from the reaction vessel 18, etc. In that case, water is supplied from the water supply tank 107 by the water supply pump 108.

[0024] Air bubbles in the reaction tank circulating water 112 disturb the absorbance measurement. For this reason, in the circulation flow path of Fig. 2A, the reaction tank circulating water 112 is constantly degassed by the degassing device 102 to prevent an increase in the concentration of dissolved oxygen in the circulating water. Furthermore, assuming that the generated air bubbles will adhere to the reaction vessel 1 or the optical axis window, a surfactant is generally added to the reaction tank circulating water 112.

[0025] 2B shows an example of the configuration of the degassing device 102. In the degassing device 102, a container 121 is provided to surround a flow path 120 of the reaction vessel circulating water 112 containing hollow fibers, and the container 121 is connected to a vacuum pump 103. The reaction vessel circulating water 112 in the flow path 120 is degassed by creating a negative pressure inside the container 121 with the vacuum pump 103. In the degassing device 102, for example, if the hollow fibers are cut or the flow path 120 is clogged, the reaction vessel circulating water 112 may pass through the degassing device 102 without being properly degassed. Alternatively, if the container 121 is cracked and its airtightness is lost, the reaction vessel circulating water 112 may not be sufficiently degassed.

[0026] When the dissolved oxygen concentration of the reaction tank circulating water 112 is kept at 5.0 mg / L or less, the absorbance fluctuation due to bubbles is not observed. The effect of the dissolved oxygen concentration on the water blank absorbance range will be described with reference to Figures 3 and 4. Figures 3 and 4 show the results of measuring the water blank value three times and plotting the range (maximum value - minimum value). If any of the three photometric data is affected by bubbles in the reaction tank circulating water 112, the absorbance range will increase. This measurement was performed 2400 times for each. In both cases, the first 1200 measurements included measurements using light of six wavelengths (340, 415, 480, 546, 600, 700 nm), and the latter 1200 measurements were only measurements using light of 340 nm. In both cases, the first and second measurements showed the same tendency, and no effect due to the selection of wavelength was observed.

[0027] Figure 3 shows an example of a measurement when the degassing mechanism is working properly, and the dissolved oxygen concentration in the reactor circulating water was 4.7 mg / L. The water blank absorbance range (ABS.) was generally below 0.001, and even the largest ranges were within about 0.002. The average value and standard deviation of the water blank absorbance range in this case were 6.9 mg / L and 2.7 mg / L, respectively.

[0028] Figure 4 shows the measurement results when the dissolved oxygen concentration was 6.5 mg / L. Bubbles were generated in the circulating water of the reaction tank, and the generated bubbles randomly crossed the light path of the photometer. The floating bubbles vary in size, and the amount of change in absorbance varies depending on their size and amount, so the distribution width of the water blank absorbance range becomes wider. In this result, the maximum value of the water blank absorbance range (ABS.) is approximately 0.01, but there are cases where the range exceeds this. The average value and standard deviation of this water blank absorbance range were 11.5 mg / L and 7.7 mg / L, respectively.

[0029] FIG. 5 shows the average value of the water blank absorbance range (ABS.) calculated from the results obtained by changing the dissolved oxygen concentration of the reaction tank circulating water from 4.7 to 6.5 mg / L, and plotted. FIG. 6 also shows the standard deviation of the water blank absorbance range (ABS.) plotted. Both are correlated with the dissolved oxygen concentration. Thus, it can be seen that the state of the dissolved oxygen concentration can be detected by monitoring the water blank absorbance range of the reaction tank circulating water 112. In this detection, since the water blank absorbance range shows the same tendency independent of the measurement wavelength, photometric data of different wavelengths may be mixed in the monitoring. Since the water blank measurement is performed with light of the wavelength used for the test, the water blank measurement is measured with a different wavelength depending on the requested test content. For this reason, it is an advantageous feature that the water blank absorbance range does not depend on the measurement wavelength. Of course, the water quality monitoring of the reaction tank circulating water 112 may be performed only with data of a specific wavelength.

[0030] As an index of the dissolved oxygen concentration, the average value (Figure 5) or standard deviation (Figure 6) of the water blank absorbance range may be used, or an index derived from these may be used. For example, the frequency of exceeding the set water blank absorbance range (average value or standard deviation) may be used as an index.

[0031] Next, the process by which the automatic analyzer judges whether there is an abnormality in the water quality of the reaction tank circulating water based on the index of the water blank absorbance range will be described in detail. A flow chart for judging whether there is an abnormality in the water quality is shown in Fig. 7. This process is executed by the control unit 22 starting the reaction tank circulating water monitoring program and coordinating the operation of the hardware of the automatic analyzer.

[0032] First, the control unit 22 measures the water blank absorbance (S01) and stores it in the storage device 23 (S02). From the obtained photometric data, the water blank absorbance range for each water blank measurement is calculated (S03), and stored in the storage device 23 (S04). This water blank absorbance measurement is performed as the water blank measurement included in the analysis sequence of the automatic analyzer described with reference to Fig. 1. Therefore, by the control unit 22 executing the reaction tank circulating water monitoring function, no new measurement for this purpose is required.

[0033] FIG. 11 shows an example of a GUI (Graphical User Interface) for setting the water blank absorbance measurement used in the reaction tank circulating water monitoring function. The measurement condition setting screen 300 includes a measurement condition setting section 301 and an exclusion condition setting section 302. The measurement condition setting section 301 sets which wavelength of the water blank absorbance measurement to be used for the calculation of the index, among the wavelengths that can be measured by the photometer of the device. In this example, the wavelengths that can be measured by the photometer are listed and can be selected by the user. The exclusion condition 302 sets the water blank absorbance measurement to be excluded from the calculation of the index. For example, immediately after the reaction tank circulating water is replaced, degassing by the degassing device 102 may be insufficient, and the dissolved oxygen concentration of the reaction tank circulating water 112 may be higher than the steady state. For this reason, it is possible to set the photometric data measured within a predetermined time after the replacement of the circulating water not to be used for the calculation of the index. Alternatively, immediately after the light source lamp 3 starts to be turned on, the amount of light from the light source lamp 3 may be unstable. For this reason, it is possible to set so that photometric data measured within a predetermined time after the light source lamp 3 is turned on is not used in calculating the index.

[0034] It is determined whether the necessary number of data points of the water blank absorbance range have been stored (S05), and the measurement and storage of the water blank absorbance and the calculation and storage of the water blank absorbance range are repeated until the necessary number of data points are accumulated. When the necessary number of data points are accumulated, the control unit 22 calculates an index value of the reaction tank circulating water based on the water blank absorbance range stored in the storage unit 23, and stores it in the storage unit 23 (S06). As the index, the index of the dissolved oxygen concentration as described above is used. The control unit 22 compares the index value with a threshold value (S07), and if the index value exceeds the threshold value, notifies the user of an abnormality in the water quality of the reaction tank circulating water (S08). The threshold value for abnormality determination may be set to a fixed value, or a distribution of normal data in the device may be obtained and used as a reference. For example, the standard deviation SD of the water blank absorbance range is calculated from a plurality of measurement data points collected when the device is in a normal state after maintenance of the reaction tank, and the standard deviation SD is multiplied by a set coefficient C to obtain the threshold value (C×SD). The threshold value is updated every time maintenance is performed on the reaction tank.

[0035] FIG. 12 shows an example of a GUI for setting a judgment condition in step S07. In step S07, one judgment condition may be set and judgment may be made based on whether or not the condition is satisfied, or multiple judgment conditions may be set and judgment may be made based on a combination of the conditions (AND condition, OR condition, etc.). For this reason, the GUI is made to be capable of setting multiple judgment conditions. In the validation setting unit 311, it is set whether or not the condition is to be used for judgment in step S07. In the data number setting unit 312, the number of data required for judgment in step S05 is set. Note that, in this GUI, the acquisition time setting unit 313 can set the number of data by the acquisition time of the photometric data, rather than the number of data itself. When both the data number setting unit 312 and the acquisition time setting unit 313 are set as in judgment condition 1, it is determined in step S05 that the required number of data in the water blank absorbance range has been accumulated when either one of them is satisfied. In the index setting unit 314, an index used for judgment in step S07 is set, and in the threshold setting unit 315, a threshold used for judgment in step S07 is set. When the frequency of exceeding the threshold is used as the judgment condition, the threshold for the frequency of the threshold exceeding can be set in the threshold exceeding frequency setting unit 316. As a result, for example, in judgment condition 2, a judgment condition is set such that an abnormality is judged to be present if the frequency of deviation from the average value exceeding 300 occurs three or more times within 12 hours.

[0036] If an abnormality in the water quality of the reaction tank circulating water is detected, it is possible to output the information to the display 24 or printer 25 to warn the user. In addition, by adding and recording the alarm information together with the measurement data in the storage device 23 and providing a function for later referring to the stored data, it is possible to analyze the cause of the abnormality from the measurement data. This can improve the reliability of the measurement. Furthermore, the stored data can be transferred via a communication line to a service center, where the status of the automatic analyzer can be monitored. This can enable measures such as issuing maintenance instructions or replacing parts at an early stage, improving the reliability of the device.

[0037] Since the water quality abnormality of the reaction tank circulating water is judged based on the water blank absorbance, the cause of the abnormality may be a lack of surfactant concentration in the reaction tank circulating water in addition to a failure of the degassing device. When the reaction tank circulating water 112 overflows from the reaction tank 18, the surfactant concentration of the reaction tank circulating water 112 gradually decreases as water is replenished. When the surfactant concentration decreases, air bubbles are likely to form on the reaction vessel or the window of the reaction tank. In this case, the water blank absorbance range is expected to be improved by adding a surfactant. Therefore, when an abnormality is judged in step S07, a step of adding a surfactant to the reaction tank circulating water 112 may be provided. The surfactant is added, the water blank absorbance is measured again, the index value is calculated, and an abnormality judgment is made (S01 to S07). When the index value still indicates an abnormality, an alarm is generated and the user is notified (S08).

[0038] In addition, since the amount of the residue increases if there is any floating matter such as dirt in the reaction tank circulating water 112, this may be improved by performing maintenance such as changing the water in the reaction tank, cleaning the reaction tank, etc. Therefore, by storing and monitoring the maintenance status together with the water blank absorbance range, it becomes possible to more appropriately determine whether the water quality is abnormal.

[0039] That is, in this embodiment, the water quality abnormality of the reaction tank circulating water 112 is judged based on the water blank absorbance range, but there are several possible causes for the deterioration of the water blank absorbance range, and the method of resolving the abnormality differs for each cause. Therefore, the water quality may be judged based on the past operation history of the automatic analyzer, and an alarm may be issued. This makes it possible to narrow down the cases where there is a high possibility that an abnormality has occurred in the degassing device, for example, and issue an alarm. For this purpose, it is considered that the water quality abnormality of the reaction tank circulating water 112 is judged using an abnormality judgment algorithm based on machine learning.

[0040] 8 is a functional block diagram of creating / updating a learning model using a learning device 210. The learning device 210 is realized by a processing device (processor) executing a program stored in the storage device. The physical configuration of the learning device 210 may be a single server or a PC (Personal Computer), or any part of the processing device or storage device may be configured as multiple servers connected via a network. In addition, functions equivalent to all or part of the functions configured by software may be realized by hardware such as an FPGA (Field Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit).

[0041] The learning device 210 acquires information related to the water quality of the reaction tank circulating water (collectively referred to as water quality related data), such as water blank absorbance range data 201, maintenance history data 202, and alarm information data 203, from the storage device 23 of the automatic analyzer. In addition, the water quality related data may include water blank absorbance range data and index data calculated based on the water blank absorbance range data. FIG. 9A shows an example of the maintenance history data 202. The maintenance history data 202 includes the name 231 of the performed maintenance, the date 232 and the time 233 when the maintenance was performed. FIG. 9B shows an example of the alarm information data 203. The alarm information data 203 includes the date 241, the time 242, the alarm code 243, and the alarm name 244 when the automatic analyzer issued an alarm.

[0042] The input unit 211 acquires these data directly or via a network. The training data generating unit 212 generates training data from the data acquired by the input unit 211. FIG. 13 shows an example of a GUI that the training data generating unit 212 displays on the display unit of the learning device 210. In the training data import screen 400, a graph 402 indicating the transition of the index (standard deviation of the water blank absorbance range) of the judgment condition 1 shown in FIG. 12 is displayed below a graph 401 indicating the transition of the water blank absorbance range. The display positions of the two graphs 401 and 402 can be moved in conjunction with each other by a slider 403. The operator cuts out training data appropriate for learning while observing the transition of the water blank absorbance range and the index. Specifically, by specifying an import start position 404 and an import end position 405 on the graph and pressing an abnormality pattern import button 406, the water quality related data of this section can be imported as a pattern indicating water quality abnormality. Conversely, by pressing the exclusion pattern import button 407, the water quality related data for this section can be imported as a pattern that does not indicate water quality abnormality. This allows the operator's findings to be reflected in the learning of the learning model.

[0043] The training data input unit 213 inputs the training data generated by the training data generation unit 212 into a learning model for determining water quality abnormality, and the learning model update unit 214 updates the learning model. The updated learning model 220 is transferred from the output unit 217 to the automatic analyzer, and the control unit 22 uses the learning model 220 to determine the performance of the degassing device. The prototype of the learning model used to determine the performance of the degassing device is stored in the algorithm database 215. In order to improve the accuracy of the learning model, it is necessary to learn using as much training data as possible. For this reason, similar data is collected from other automatic analyzers and stored in the other device data database 216. The learning model stored in the algorithm database 215 is created based on the training data from the other device data database 216. The learning model update unit 214 updates the learning model, and when a change occurs in the algorithm of the learning model, it feeds back change information to the training data generation unit 212 and updates the training data generation conditions of the training data generation unit 212.

[0044] The learning model 220 may be any learning device such as a neural network, a regression tree, or a Bayesian classifier. FIG. 10 shows an example of a learning model using a neural network. In the learning model 90, information input to an input layer 91 is propagated to an intermediate layer 92 and then to an output layer 93, and an inference result based on the information input from the output layer 93 to the input layer 91 is output. The input layer 91, the intermediate layer 92, and the output layer 93 each have a plurality of input units, intermediate units, and output units indicated by circles. Although the intermediate layer of a neural network usually has a plurality of layers, the intermediate layer 92 is shown as a representative in the figure. Information input to each input unit of the input layer 91 is weighted by a coupling coefficient between the input unit and the intermediate unit, and is input to each intermediate unit. The value of the intermediate unit in the intermediate layer 92 is calculated by adding values ​​from the input units. In addition, the output from each intermediate unit in the intermediate layer 92 is weighted by a coupling coefficient between the intermediate unit and the output unit, and is input to each output unit. The values ​​from the hidden units are added to calculate the values ​​of the output units in the output layer 93. In this way, the processing in the hidden layer 92 corresponds to nonlinearly transforming the values ​​of the input data input to the input layer 91 and outputting them as output data from the output layer 93.

[0045] The learning model 90 is a learning model related to the output data "degassing performance judgment", and is shown together with examples of input and output data. Note that both input and output data are examples. Water quality-related data (or processed data thereof, hereinafter referred to as water quality-related data without distinction) is input to the input layer 91, and the calculation result according to the input is output from the output layer 93. The data input to the input layer 91 includes measurement result information 94, such as the water blank absorbance, water blank absorbance range (maximum value-minimum value), and the frequency at which the value of the water blank absorbance range becomes an abnormal value, defined for each learning model, and device-related information 95, including the status of maintenance performed by the user and information on the history of alarms that have occurred so far. Maintenance includes operations such as water replacement in the reaction tank, cleaning of the reaction tank, and replacement of the degassing device, which may affect the degassing performance. In addition, the alarm history information includes alarms of various levels, and there are cases where maintenance is required to cancel the alarm, and cases where the alarm is at a level of caution that does not require immediate maintenance but requires observation.

[0046] The output data from the output layer 93 is the inference result for the degassing performance judgment result 97 and the cause of abnormality 98, and the necessity of action for the degassing device 102 can be determined based on this result. The output values ​​of the output data of the degassing performance judgment result 97, such as "deterioration of degassing performance (alarm level)", "caution required for degassing performance", and "normal degassing performance", are probability values, and a determination is made based on whether the output probability value is above or below a predetermined threshold value. The cause of abnormality 98 suggests the occurrence of an abnormality due to a factor other than the degassing device. For example, it is possible to infer the occurrence of an abnormal value due to an abnormality cause other than degassing performance, such as dirt in the reaction tank, clogging due to debris, or the light source lamp. [Explanation of symbols]

[0047] 1: reaction vessel, 2: reaction disk, 3: light source lamp, 4: spectroscopic diffraction grating, 5: photodetector, 6: constant temperature water supply device, 7: sample cup, 8: sample disk, 9: reagent bottle, 10: reagent disk, 11: sample dispensing mechanism, 12: reagent dispensing mechanism, 13: stirring mechanism, 14: sample dispensing pump, 15: pure water pump for cleaning, 16: reagent dispensing pump, 17: cleaning mechanism, 18: reaction tank, 20: AD conversion unit, 22: control unit, 23: storage device, 24: display, 25: printer, 26: keyboard ,27: Interface bus, 90: Learning model, 91: Input layer, 92: Middle layer, 93: Output layer, 94: Measurement result information, 95: Equipment related information, 97: Degassing performance judgment result, 98: Cause of abnormality, 102: Degassing device, 103: Vacuum pump, 104: Cooling device, 105: Circulation pump, 106: Heating device, 107: Water supply tank, 108: Water supply pump, 109: Water supply solenoid valve, 110: Drain solenoid valve, 111: Reaction liquid, 112: Reaction tank circulating water, 120: Flow path, 121: Container, 201: Water brand absorbance range data, 202: maintenance history data, 203: alarm information data, 210: learning device, 211: input section, 212: training data generation section, 213: training data input section, 214: learning model update section, 215: algorithm database, 216: other device data database, 217: output section, 220: learning model, 231: maintenance name, 232, 241: date, 233, 242: time, 243: alarm code, 244: alarm name, 300: measurement conditions Setting screen, 301: measurement condition setting section, 302: exclusion condition setting section, 310: judgment condition setting section, 311: activation setting section, 312: number of data setting section, 313: acquisition time setting section, 314: index setting section, 315: threshold setting section, 316: threshold over occurrence frequency setting section, 400: training data import screen, 401, 402: graph, 403: slider, 404: import start position, 405: import end position, 406: abnormal pattern import button, 407: exclusion pattern import button, 501, 502: graph.

Claims

1. A reaction vessel; a reaction disk for holding a reaction vessel in a state where the reaction vessel is immersed in the reaction vessel circulating water in the reaction vessel; a circulation flow path for adjusting the temperature of the reaction tank circulation water from the reaction tank and returning the reaction tank circulation water to the reaction tank; a degassing device provided in the circulation flow path and configured to degas the reaction vessel circulating water; a photometer including a light source lamp and a photodetector for detecting light irradiated from the light source lamp and transmitted through the reaction vessel; a control unit that controls an analysis sequence of the sample and performs the analysis of the sample; the control unit, prior to dispensing a sample into the reaction vessel and performing a sample measurement, dispenses pure water into the reaction vessel to perform a water blank measurement, corrects absorbance data of a reaction solution of the sample and a reagent measured in the sample measurement using the water blank absorbance data measured in the water blank measurement, and analyzes the sample based on the corrected absorbance data; The control unit performs multiple measurements on the reaction vessel into which pure water has been dispensed in one water blank measurement, and accumulates water blank absorbance range data, which is the difference between the maximum and minimum values ​​of the water blank absorbance data in the multiple measurements, in a storage device.The automatic analyzer determines whether there is an abnormality in the water quality of the reaction tank circulating water based on an index calculated by statistically processing the accumulated water blank absorbance range data after a predetermined number of data or a predetermined data acquisition time has elapsed.

2. (delete)

3. In claim 1, The control unit of the automatic analyzer adds alarm information regarding the abnormality in the water quality of the reaction tank circulating water to the analysis results of the sample when an abnormality in the water quality of the reaction tank circulating water is determined.

4. In claim 1, An automatic analyzer, wherein the index is the average value or standard deviation of the water blank absorbance range data.

5. In claim 1, An automatic analyzer that determines whether the water quality of the reaction tank circulating water is abnormal based on the frequency with which the index exceeds a threshold value.

6. In claim 1, The control unit is an automatic analyzer that adds a surfactant to the reaction tank circulating water when it determines that the water quality of the reaction tank circulating water is abnormal.

7. (delete)

8. In claim 1, The control unit statistically processes a predetermined number of the water blank absorbance range data to calculate a first index and a second index, and determines whether there is an abnormality in the water quality of the reaction tank circulating water by combining a first judgment condition using the first index or a second judgment condition using the second index.

9. A reaction vessel; a reaction disk for holding a reaction vessel in a state where the reaction vessel is immersed in the reaction vessel circulating water in the reaction vessel; a circulation flow path for adjusting the temperature of the reaction tank circulation water from the reaction tank and returning the reaction tank circulation water to the reaction tank; a degassing device provided in the circulation flow path and configured to degas the reaction vessel circulating water; a photometer including a light source lamp and a photodetector for detecting light irradiated from the light source lamp and transmitted through the reaction vessel; a control unit that controls an analysis sequence of the sample and performs the analysis of the sample; the control unit, prior to dispensing a sample into the reaction vessel and performing a sample measurement, dispenses pure water into the reaction vessel to perform a water blank measurement, corrects absorbance data of a reaction solution of the sample and a reagent measured in the sample measurement using the water blank absorbance data measured in the water blank measurement, and analyzes the sample based on the corrected absorbance data; The control unit performs multiple measurements on the reaction vessel into which pure water has been dispensed in one water blank measurement, and determines the performance of the degassing device using a learning model that inputs the water blank absorbance data, water blank absorbance range data which is the difference between the maximum and minimum values ​​of the water blank absorbance data in the multiple measurements, maintenance history data related to the water quality of the reaction tank circulating water, and alarm information data related to the water quality of the reaction tank circulating water.

10. In claim 9, The learning model is an automatic analyzer that includes as input an index calculated by statistically processing a predetermined number of the water blank absorbance range data.

11. In claim 9, The learning model of the automatic analyzer is updated using the water blank absorbance data extracted based on an index calculated by statistically processing a predetermined number of the water blank absorbance range data, the water blank absorbance range data, the maintenance history data, and the alarm information data as training data.

12. A reaction vessel; a reaction disk for holding a reaction vessel in a state where the reaction vessel is immersed in the reaction vessel circulating water in the reaction vessel; a circulation flow path for adjusting the temperature of the reaction tank circulation water from the reaction tank and returning the reaction tank circulation water to the reaction tank; a degassing device provided in the circulation flow path and configured to degas the reaction vessel circulating water; a photometer including a light source lamp, a photodetector that detects light irradiated from the light source lamp and transmitted through the reaction vessel, and a diffraction grating that separates the light transmitted through the reaction vessel into wavelength components; a control unit that controls an analysis sequence of the sample and performs the analysis of the sample; the control unit, prior to dispensing a sample into the reaction vessel and performing a sample measurement, dispenses pure water into the reaction vessel to perform a water blank measurement, corrects absorbance data of a reaction solution of the sample and a reagent measured in the sample measurement using the water blank absorbance data measured in the water blank measurement, and analyzes the sample based on the corrected absorbance data; The control unit performs a plurality of measurements on the reaction vessel into which pure water has been dispensed in one water blank measurement, stores water blank absorbance range data, which is the difference between the maximum value and the minimum value of the water blank absorbance data in the plurality of measurements, in a storage device, and determines a water quality abnormality of the reaction tank circulating water based on an index calculated by statistically processing a predetermined number of the water blank absorbance range data, The control unit is an automatic analyzer capable of selecting the water blank absorbance data to be used in calculating the index based on the wavelength of light used in the water blank measurement.

13. A reaction vessel; a reaction disk for holding a reaction vessel in a state where the reaction vessel is immersed in the reaction vessel circulating water in the reaction vessel; a circulation flow path for adjusting the temperature of the reaction tank circulation water from the reaction tank and returning the reaction tank circulation water to the reaction tank; a degassing device provided in the circulation flow path and configured to degas the reaction vessel circulating water; a photometer including a light source lamp and a photodetector for detecting light irradiated from the light source lamp and transmitted through the reaction vessel; a control unit that controls an analysis sequence of the sample and performs the analysis of the sample; the control unit, prior to dispensing a sample into the reaction vessel and performing a sample measurement, dispenses pure water into the reaction vessel to perform a water blank measurement, corrects absorbance data of a reaction solution of the sample and a reagent measured in the sample measurement using the water blank absorbance data measured in the water blank measurement, and analyzes the sample based on the corrected absorbance data; The control unit performs a plurality of measurements on the reaction vessel into which pure water has been dispensed in one water blank measurement, stores water blank absorbance range data, which is the difference between the maximum value and the minimum value of the water blank absorbance data in the plurality of measurements, in a storage device, and determines a water quality abnormality of the reaction tank circulating water based on an index calculated by statistically processing a predetermined number of the water blank absorbance range data, The control unit of the automatic analyzer is capable of excluding the water blank absorbance data used to calculate the index based on the elapsed time since the reaction tank circulating water was replaced or the elapsed time since the light source lamp was turned on.

Images