WO2011135606A1 - 時系列データ診断圧縮方法 - Google Patents
時系列データ診断圧縮方法 Download PDFInfo
- Publication number
- WO2011135606A1 WO2011135606A1 PCT/JP2010/002968 JP2010002968W WO2011135606A1 WO 2011135606 A1 WO2011135606 A1 WO 2011135606A1 JP 2010002968 W JP2010002968 W JP 2010002968W WO 2011135606 A1 WO2011135606 A1 WO 2011135606A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- data
- compression
- time
- diagnosis
- unit
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
- G05B23/0264—Control of logging system, e.g. decision on which data to store; time-stamping measurements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/23—Updating
- G06F16/2365—Ensuring data consistency and integrity
Definitions
- the present invention is for a state-based maintenance service that collects various time-series sensing information such as logging data, process data, and event data from equipment installed in a plant, manufacturing equipment in a factory, heavy equipment, large equipment, etc.
- the present invention relates to a time-series data diagnostic compression method and a time-series data collection / storage device.
- the plant collects tens of thousands of operating data per second per site. If all the operation data measured in the plant is collected and stored in the storage on the server, several gigabytes of data is accumulated in the storage every day.
- operation data measured by hundreds of sensors installed in the equipment is collected at a high-speed cycle such as 50 milliseconds, and the amount of data collected is A device reaches the order of gigabytes per day.
- Patent Document 1 discloses that a data compression rate (allowable error) is not fixed to an initial setting value at the time of system design, but is generated by a plant or equipment. By dynamically setting in response to events such as abnormalities (alarms) and operations, it is possible to maintain high compression during normal conditions, while reducing the compression rate to low or no compression during abnormal conditions. A compression method for reducing the data storage amount while avoiding loss is described.
- Patent Document 3 describes abnormality sign diagnosis using vector quantization clustering.
- the vector quantization technique is described in detail in Non-Patent Document 3
- the clustering cluster analysis
- the tolerance of the operation data of the collected equipment is not fixed to the initial setting value at the time of system design, but dynamically according to the event such as an abnormality (alarm) or operation generated by the plant or equipment.
- the data storage amount can be reduced while avoiding the loss of data necessary for the maintenance service by setting the compression rate to high during normal operation and reducing the compression rate to low or no compression during abnormal operation.
- an object of the present invention is to solve at least one of the above problems.
- the present invention provides means for performing an abnormal sign diagnosis and setting an allowable error using the diagnosis result.
- the clustering technology based on vector quantization is used to diagnose abnormal signs, and machine input data is mechanically learned using multivariate analysis, and the degree of deviation from the input data at the time of diagnosis.
- data loss during this period can be prevented by disabling data compression during the period when the abnormality sign is detected by the abnormality sign diagnosis or by setting the tolerance to 0 (see (2)). Solution), and the validity of the tolerance can be confirmed (solution of the problem (3)).
- the compression of the time series data collected from the device based on the diagnosis result by the general-purpose abnormality sign diagnosis technology independent of the device is managed, knowledge of the device is not required, and various devices can be used. Implementation of data compression / storage becomes easy.
- the data storage amount can be reduced efficiently without losing the data and information necessary for the state-based maintenance service of the target device.
- the block diagram of a data collection storage device The block diagram of an original data storage buffer.
- FIG. 1 is a configuration diagram of a compression strategy history storage unit.
- FIG. 2 is a configuration diagram of a compression strategy history storage unit.
- the block diagram of the data collection storage device which guarantees the abnormal sign diagnostic accuracy of Example 2.
- FIG. FIG. 10 is a configuration diagram of an abnormality sign diagnosis execution history storage unit according to the second embodiment.
- the block diagram of the data collection storage device which compresses after the storage of Example 3.
- the data collection / storage device 100 includes a data collection diagnosis compression unit 120 and a storage unit 130.
- the data collection diagnostic compression unit 120 includes a data collection unit 110, a diagnostic compression management unit 1201, and a data compression unit 125.
- the data collection unit 110 has an original data storage buffer 111.
- the diagnostic compression management unit 1201 includes a failure sign diagnosis unit 121 and a compression strategy management unit 124.
- the data compression unit 125 has a compressed data storage buffer 1251.
- the storage unit 130 includes a failure sign diagnosis execution history storage unit 131, a compression strategy history storage unit 132, and a compressed data storage unit 133.
- the storage unit 130 includes a database or file on a hard disk drive or nonvolatile memory.
- the device 200 is provided with sensors 1, sensors 2,..., Sensor n (201, 202,..., 20n, where n is a positive integer).
- the information group becomes input data of the data collection storage device 100.
- the data collection unit 110 collects time-series input data from the sensors 1, 2,..., Sensor n and stores the collected data in the original data storage buffer 111.
- the input data stored in the original data storage buffer 111 is sequentially transferred to the failure sign diagnosis unit 121 and the data compression unit 125.
- the failure sign diagnosis unit 121 performs an abnormality sign diagnosis using the received input data, and outputs the diagnosis result to the compression strategy management unit 124.
- the compression strategy management unit 124 sets the compression strategy of the data compression unit 125 based on the received diagnosis result.
- the data compression unit 125 compresses the input data passed from the data collection unit 110 according to the compression strategy of the compression strategy management unit 124, and temporarily stores the compressed data in the compressed data storage buffer 1251.
- Data is stored in the compressed data storage unit 133 according to the compression strategy of the compression strategy management unit 124, either compressed data stored in the compressed data storage buffer 1251 or uncompressed original data stored in the original data storage buffer 111. Store the data. Details will be described later.
- diagnosis result output from the failure sign diagnosis unit 121 and the allowable error output from the compression strategy management unit 124 are stored in the failure sign diagnosis execution history storage unit 131 and the compression strategy history storage unit 132, respectively.
- the input data of the data collection and storage device 100 shows sensor information 201, 202,..., 20n installed in a single device 200, but the device is composed of a plurality of devices. May be. Alternatively, it may be various types of logging data such as information other than the device main body, for example, sensing information of peripheral devices and piping parts, or sensing information of surrounding environment information. Further, it may be process data or event data. Furthermore, it is not limited to sensor information, but may be output information from other monitoring control devices.
- the logging data refers to information necessary for device operation control, such as the number of revolutions and position, status information during operation of the device, such as water temperature and vibration, or temperature and humidity. It refers to various types of sensor information collected periodically, such as environmental information of surroundings where various devices are installed.
- Process data refers to a data set that is added through a series of business processes that are added (or modified in some cases) as each task is completed.
- Event data refers to information such as device operations and alarms.
- Sensor 1, sensor 2,..., Sensor n (201, 202,..., 20n, where n is a positive integer) installed in the device 200, such as logging data, process data, and event data.
- Time series sensing information (input data) is registered as a set with the measurement time.
- This buffer is composed of, for example, a ring buffer, and when input data registered in the buffer is full, new input data is overwritten in order from old input data.
- FIG. 3 is a diagram for explaining normal cluster formation using vector quantization in the abnormality sign diagnosis unit.
- FIG. 3 shows a situation where the space in which the normal input data vector is distributed is classified into k hyperspherical clusters 1, clusters 2,..., Cluster k by clustering the learning data.
- the center of gravity of each cluster is indicated by c1, c2,.
- the set of barycentric points c1, c2,..., Ck of each cluster formed by the normal input data vector group is managed by the failure predictor diagnosis unit 121 as a code book.
- FIG. 4 is a diagram for explaining an abnormality detection method using clusters in the abnormality sign diagnosis unit.
- the failure sign diagnosis unit 121 calculates the minimum distance among the distances between the input data vector and the center of gravity of each cluster, and outputs this as the device abnormality degree.
- xa and xb be input data vectors at certain times Ta and Tb at the time of abnormality sign diagnosis.
- the input data vector xa at time Ta is included in the hypersphere formed by the cluster k and is normal.
- the input data vector xb at time Tb is not included in any cluster, and in this case, there is a sign of abnormality.
- the device abnormality degree of xb is dkb.
- each sensing data is preferably normalized to its standard deviation to make the cluster hyperspherical by performing scaling conversion such as Mahalanobis distance.
- the device abnormality degree can be treated as abnormal sign detection when a value of 1 is a hypersphere boundary, and within 1 is normal and exceeds 1.
- each value of the input data vector (v1i, v2i,..., Vni) is treated as normalized.
- FIG. 5 is a diagram for explaining a method of calculating the degree of abnormality using a cluster and a method of calculating the degree of abnormality contribution of each sensor in the abnormality sign diagnosis unit.
- the failure level calculation unit 121 uses the cluster to calculate the degree of abnormality and the method for calculating the degree of abnormal contribution of each sensor.
- the values obtained by dividing each component of the distance d by the distance d and normalized are r1 and r2, respectively, these values can be used as indicators of how much they contribute to the device abnormality.
- r1 and r2 are referred to as abnormality contributions, and may be output from the failure sign diagnosis unit 121 together with the device abnormality degree, and is expressed by Expression 2.
- r1 d1 / d
- r2 d2 / d
- This method detects abnormal signs by mechanically learning the input data when the device is normal using multivariate analysis, so it does not require device knowledge as in the prior art. It is possible to avoid omissions in the detection of signs due to shortage, and it is possible to deal with various devices.
- the computational load of the abnormal sign diagnosis process depends on the dimensions of the input data vector and the total number of clusters created during learning. It is possible to make the system compact, and it is possible to realize the data collection and storage device 100 on which the failure sign diagnosis unit 121 is mounted with an embedded device having restrictions on CPU power and memory size.
- the failure sign diagnosis unit 121 a cluster analysis method using vector quantization is used to calculate the degree of divergence from the cluster formed using normal input data and calculate it.
- the embodiment in which the device abnormality level is output has been described.
- the cluster analysis method used by the failure sign diagnosis unit 121 is not limited to vector quantization. Furthermore, it is not limited to cluster analysis, but other multivariate analysis methods (for example, MT method) may be applied. Alternatively, if there is a simulator for the device, the degree of abnormality of the device may be output using the simulation result.
- the data compression unit 125 performs irreversible compression such as dead band compression or change rate compression (Non-Patent Document 1), swing door compression (Non-Patent Document 2), and compression using a virtual straight line (Patent Document 2). Is assumed to be implemented.
- the compression strategy management unit 124 changes the degree of data compression performed by the data compression unit 125 by setting an allowable error (margin) according to the degree of device abnormality output by the failure sign diagnosis unit 121.
- FIG. 6 shows a setting example 1 of an allowable error in the compression strategy management unit.
- the first embodiment in the compression strategy management unit 124 will be described with reference to FIG.
- the upper graph shows the time lapse of the device abnormality degree output from the failure sign diagnosis unit 1211
- the lower graph shows the compression strategy management unit 124 sends the data compression unit 125 to the data compression unit 125 according to the device abnormality degree.
- the allowable error is set as follows.
- tolerance m
- tolerance 0 (no compression) That is, as a result of diagnosing the input data, the input data is compressed with the allowable error m in the normal interval, and the allowable error is 0 or not compressed in the abnormal sign detection interval.
- the allowable error m is set to m% (for example, 1%) of the maximum value that the sensor can take. Or you may set the value of m separately for every sensor.
- the abnormality degree of an apparatus is as follows on the basis of the time te. (1)
- FIG. 7 shows a setting example 2 of the allowable error in the compression strategy management unit. A second embodiment of the compression strategy management unit 124 will be described with reference to FIG.
- the input data compressed based on the allowable error may be data from all types of sensors collected by the data collection unit 110. Alternatively, it may be data of only a diagnosis target sensor set in advance. Alternatively, it may be data of only a sensor having a high abnormality contribution value (greater than the threshold value rth) (rth value is, for example, 0.4. Since the abnormality contribution degree is normalized, it takes between 0 and 1).
- the data compression unit 125 has been described assuming lossy compression, but the data compression unit 125 performs lossless compression such as zip compression or gz compression. You may implement.
- the compression strategy management unit 124 sets the compression level instead of the allowable error according to the device abnormality degree output from the failure sign diagnosis unit 121.
- the device abnormality degree is 1 or less
- An operation for setting the compression level to 0 (no compression) may be considered when a compression level for performing the compression is set and an abnormality sign with a device abnormality degree exceeding 1 is detected.
- the data compression unit 125 may perform multi-stage compression such as performing lossless compression after performing lossy compression.
- ⁇ Compressed data storage unit> An embodiment of a compressed data storage unit 133 that stores data after compression by the data compression unit 125 will be described with reference to FIGS. 8, 9, 10, and 11.
- the configuration of the compressed data storage buffer 1251 in the data compression unit 125 is basically the same as that of the compressed data storage unit 133.
- FIG. 8 shows Example 1 of the compressed data storage unit.
- FIG. 8 shows an example of the compressed data storage unit 133 by the compressed data storage unit 133 when the permissible error of each sensing data in the period from time to tm is m (fixed).
- the portion indicated by null is the sensing data thinned out by the data compression unit 125.
- the sensor 1 has four sensing data from time t2 to t5
- the sensor 2 has four sensing data from time t1 to t4
- the sensor n has four sensing data from time t4, t5, t7, and t8. Each is thinned out.
- FIG. 9 shows Example 2 of the compressed data storage unit.
- FIG. 9 shows a compressed data storage unit 133 when the compression strategy management unit of FIG. 6 is applied to the storage example of FIG.
- the time te in FIG. 6 corresponds to t5 in FIG. 9, and after the time t5, the allowable error becomes 0, and compression is not performed.
- the sensing data of sensor 1 at time t5, sensor n at times t5, t7, and t8 are thinned out (denoted by null in the figure), but in FIG. 9, the sensing data is stored.
- FIG. 10 shows an example 3 of the compressed data storage unit.
- FIG. 10 shows a compressed data storage unit 133 when the compression strategy management unit of FIG. 7 is applied to the storage example of FIG.
- the time te in FIG. 7 corresponds to t5 in FIG. 9, and the time te-tp corresponds to t3. Therefore, after time t3, the allowable error is 0, and compression is not performed.
- the sensing data at time t3 to t5 of sensor 1, times t3 and t4 of sensor 2, and times t3, t5, t7 and t8 of sensor n are thinned out (denoted by null in the figure). In 9, each sensing data is stored.
- FIG. 11 shows Example 4 of the compressed data storage unit.
- FIG. 11 shows a compressed data storage unit 133 in the case where the sensors to be diagnosed by the compression strategy management unit 124 are only the sensors 1 and 2 for the storage example of FIG.
- the sensing data of sensor 1 and sensor 2 is not thinned out after time t5 as in FIG. However, since the sensor n is not a diagnosis target, data thinning is performed similarly to the sensor n of FIG. 6 even if an abnormality sign is detected.
- FIG. 12 is a block diagram of the abnormality sign diagnosis execution history storage unit, and an embodiment of the failure sign diagnosis execution history storage unit 131 will be described with reference to FIG.
- the device abnormality degree obtained as a result of diagnosing the time-series sensing information (input data) by the failure sign diagnosis unit 121 and the abnormality contribution degree of each sensor are registered as a set together with the measurement time.
- This buffer is composed of, for example, a ring buffer, and when the device abnormality degree and abnormality contribution degree registered in the buffer are full, new input data is overwritten in order from old input data.
- failure sign diagnosis execution history storage unit 131 may be a set of only the time and the device abnormality level.
- FIG. 13 shows the configuration diagram 1 of the compression strategy history storage unit, and the first embodiment of the compression strategy history storage unit 132 will be described with reference to FIG.
- an allowable error which is a compression parameter used when data compression is performed on time-series sensing information (input data) is registered together with a measurement time. That is, a set of (measurement time ti, allowable error mi, where i is an integer from 0 to m) is registered in the compression strategy history storage unit 132 in time series.
- This buffer is composed of, for example, a ring buffer, and when the device abnormality degree and abnormality contribution degree registered in the buffer are full, new input data is overwritten in order from old input data.
- FIG. 14 shows a configuration diagram 2 of the compression strategy history storage unit, and a second embodiment of the compression strategy history storage unit 132 will be described with reference to FIG.
- an allowable error that is a compression parameter used when compressing the sensing data is registered as a set with the measurement time. That is, a set of (measurement time ti, allowable error m1i of sensor 1, allowable error m2i of sensor 2,..., Allowable error mni of sensor n, where i is an integer from 0 to m) is compressed. Registered in the strategy history storage unit 132 in time series.
- This buffer is composed of, for example, a ring buffer, and when the device abnormality degree and abnormality contribution degree registered in the buffer are full, new input data is overwritten in order from old input data.
- FIG. 15 is a detailed configuration diagram of the data collection and storage device that guarantees the accuracy of abnormality sign diagnosis according to the second embodiment. Note that FIG. 16 is used in the middle of the description.
- FIG. 15 shows a second embodiment of the data collection / storage device 100.
- the data collection / storage device 100 includes a data collection diagnosis compression unit 120 and a storage unit 130.
- the data collection diagnostic compression unit 120 includes a data collection unit 110, a diagnostic compression management unit 1201, a data compression unit 125, and a data expansion unit 1252.
- the data collection unit 110 has an original data storage buffer 111.
- the diagnostic compression management unit 1201 includes a failure sign diagnosis unit 121 and a compression strategy management unit 124.
- the data compression unit 125 has a compressed data storage buffer 1251.
- the storage unit 130 includes a failure sign diagnosis execution history storage unit 131, a compression strategy history storage unit 132, and a compressed data storage unit 133.
- the device 200 is provided with sensors 1, sensors 2,..., Sensor n (201, 202,..., 20n, where n is a positive integer).
- the information group becomes input data of the data collection storage device 100.
- the data collection unit 110 collects time-series input data from the sensors 1, 2,..., Sensor n and stores the collected data in the original data storage buffer 111.
- the input data stored in the original data storage buffer 111 is sequentially transferred to the failure sign diagnosis unit 121 and the data compression unit 125.
- the failure sign diagnosis unit 121 performs an abnormality sign diagnosis using the received input data, and outputs the diagnosis result to the compression strategy management unit 124.
- the compression strategy management unit 124 sets the compression strategy of the data compression unit 125 based on the received diagnosis result.
- the data compression unit 125 compresses the input data passed from the data collection unit 110 according to the compression strategy of the compression strategy management unit 124, and temporarily stores the compressed data in the compressed data storage buffer 1251.
- Data is stored in the compressed data storage unit 133 according to the compression strategy of the compression strategy management unit 124, either compressed data stored in the compressed data storage buffer 1251 or uncompressed original data stored in the original data storage buffer 111. Store the data.
- FIG. 16 shows an operation flow of the diagnostic compression management unit of the second embodiment.
- the processing shown in FIG. 16 is performed among the data compression unit 125, the data expansion unit 1252, and the abnormality sign diagnosis unit 12.
- the compressed data stored in the compressed data storage buffer 1251 is expanded by the data expansion unit 1252, and the expanded data is sequentially transferred to the failure sign diagnosis unit 121 (S101).
- Compressed data can be expanded by, for example, performing linear approximation using data at both ends of the thinned section.
- the failure sign diagnosis unit 121 performs an abnormality sign diagnosis using the developed data (S102).
- the diagnosis result is compared with the diagnosis result performed with the original data at the same time (S103).
- the abnormality sign diagnosis unit 12 instructs the data compression unit 125 to store the compressed data in the compressed data storage buffer 1251 in the compressed data storage unit 133 ( S105).
- the abnormality sign diagnosis unit 12 instructs the data compression unit 125 to store the original data at that time stored in the original data storage buffer 111 in the compressed data storage unit 133 (S106). .
- the output of the abnormality sign diagnosis unit is stored in the failure sign diagnosis execution history storage unit 131 (S107).
- the allowable error output from the compression strategy management unit 124 is stored in the compression strategy history storage unit 132.
- the input data of the data collection and storage device 100 shows sensor information 201, 202,..., 20n installed in a single device 200, but the device is composed of a plurality of devices. May be. Alternatively, it may be various types of logging data such as information other than the device main body, for example, sensing information of peripheral devices and piping parts, or sensing information of surrounding environment information. Further, it may be process data or event data. Furthermore, it is not limited to sensor information, but may be output information from other monitoring control devices.
- FIG. 17 shows a configuration diagram of the abnormality sign diagnosis execution history storage unit of the second embodiment.
- a second embodiment of the failure sign diagnosis execution history storage unit 131 will be described with reference to FIG.
- the failure sign diagnosis execution history storage unit 131 the time series sensing information (input data) is developed by the failure sign diagnosis unit 121, and the device abnormality degree and the abnormality contribution degree of each sensor are developed by the data development unit 1252. Between the device abnormality degree obtained by diagnosing the detected data by the failure sign diagnosis unit 121, the abnormality contribution degree of each sensor, the device abnormality degree using the input data, and the equipment abnormality degree using the developed data Is registered as a set with the measurement time.
- This buffer is composed of, for example, a ring buffer, and when the device abnormality degree and abnormality contribution degree registered in the buffer are full, new input data is overwritten in order from old input data.
- the failure sign diagnosis execution history storage unit 131 stores the compressed data when the error is within the threshold value, and stores the original data (that is, no diagnostic error) when the error exceeds the threshold value. Thus, it is possible to ensure that the error of the abnormality sign diagnosis is equal to or less than the threshold value for all the data stored in the failure sign diagnosis execution history storage unit 131.
- failure sign diagnosis execution history storage unit 131 may be a set of only the time, the error of the device abnormality level, and the device abnormality level.
- FIG. 18 shows a configuration diagram of a data collection and storage device that performs compression after storage in the third embodiment.
- a third embodiment of the data collection and storage device 100 will be described with reference to FIG.
- the data collection / storage device 100 includes a data collection diagnosis compression unit 120 and a storage unit 130.
- the data collection diagnostic compression unit 120 includes a data collection unit 110, a diagnostic compression management unit 1201, and a data compression unit 125.
- the diagnostic compression management unit 1201 includes a failure sign diagnosis unit 121 and a compression strategy management unit 124.
- the storage unit 130 includes a failure sign diagnosis execution history storage unit 131, a compression strategy history storage unit 132, a compressed data storage unit 133, and an original data storage unit 134.
- the device 200 is provided with sensors 1, sensors 2,..., Sensor n (201, 202,..., 20n, where n is a positive integer).
- the information group becomes input data of the data collection storage device 100.
- the data collection unit 110 collects time-series input data from the above-described sensors 1, 2,..., Sensor n and sequentially passes them to the failure sign diagnosis unit 121 and the original data storage unit 134.
- the failure sign diagnosis unit 121 performs an abnormality sign diagnosis using the received input data, and outputs the diagnosis result to the compression strategy management unit 124.
- the compression strategy management unit 124 sets the compression strategy of the data compression unit 125 based on the received diagnosis result.
- the data compression unit 125 sequentially extracts the input data stored in the original data storage unit 134 according to the compression strategy of the compression strategy management unit 124, and sets the compression strategy at the collection time of the extracted input data to the compression strategy management unit 124. Then, the data is compressed according to the compression strategy, and the compressed data or the uncompressed original data is stored in the compressed data storage unit 133.
- diagnosis result output from the failure sign diagnosis unit 121 and the allowable error output from the compression strategy management unit 124 are stored in the failure sign diagnosis execution history storage unit 131 and the compression strategy history storage unit 132, respectively.
- the timing at which the data compression unit 125 retrieves the input data stored in the original data storage unit 134 may be a point in time when the input data stored in the original data storage unit 134 has passed for a certain period. Alternatively, it may be a point in time when the free capacity of the original data storage unit 134 becomes a certain threshold value or less. Or it is good also as the time of performing the detailed diagnosis as shown in below-mentioned FIG. 19, FIG.
- the data stored in the original data storage unit 134 may be compressed by performing lossless compression when storing the input data in the original data storage unit 134 and performing lossless expansion when extracting the input data.
- the input data of the data collection and storage device 100 shows sensor information 201, 202,..., 20n installed in a single device 200, but the device is composed of a plurality of devices. It may be. Alternatively, it may be various types of logging data such as information other than the device main body, for example, sensing information of peripheral devices and piping parts, or sensing information of surrounding environment information. Further, it may be process data or event data. Furthermore, it is not limited to sensor information, but may be output information from other monitoring control devices.
- the data collection diagnosis compression unit in the present embodiment that performs compression after storing the collected data may be one that guarantees the abnormality sign diagnosis accuracy described in the second embodiment.
- FIG. 19 shows an application mode 1 of the data collection and storage device of the fourth embodiment to the system, and a first application mode of the data collection and storage device to the system will be described with reference to FIG.
- the device 200 is provided with sensors 1, sensors 2,..., Sensor n (201, 202,..., 20n, where n is a positive integer). Connected via a network 31. Further, the data transmission apparatus 1001 is connected to the server computer 1002 via the network 32.
- the data transmission device 1001 includes a data transmission / reception unit 141.
- the server computer 1002 includes a data transmission / reception unit 142, a data collection diagnosis compression unit 120, a data compression / decompression unit 150, a storage unit 130, a detailed diagnosis unit 160, and an operation / display unit 170.
- a time-series sensing information group by sensors installed in the device 200 becomes input data and is transmitted to the data transmission device 1001 via the network 31.
- the data transmission device 1001 transmits the transmitted input data to the server computer 1002 via the network 32.
- the network 31 is, for example, a field network used for monitoring and controlling devices, and is configured by a wired network such as CAN or Lonworks, a wireless sensor network such as ZigBee or ISA100.11a, or a mixture of these networks. However, other networks may be used.
- the individual sensors and the data transmission device 1001 may be connected on a one-to-one basis instead of a network.
- the network 32 is the above-described network, a high-speed network such as Ethernet or optical fiber, a WiFi-based wireless network such as IEEE802.11a / g / b / n, or a wireless mesh network such as IEEE802.11s.
- the network is composed of the Internet, a private line, a public wireless network, a mobile phone network, a fixed telephone network, or a mixture of these networks, but may be composed of other networks.
- the data transmission / reception units 141 and 142 are used in the data transmission device 1001 and the server computer 1002.
- the data transmission / reception units 141 and 142 include a modem and a network module.
- the server computer 1002 passes the input data received by the data transmission / reception unit 142 to the data collection diagnosis compression unit 120.
- the data collection diagnosis compression unit 120 performs input data collection / abnormality sign diagnosis and data compression by performing the operations described in the first, second, or third embodiment.
- the data compressed by the data collection diagnosis compression unit 120 is further compressed by the data compression / decompression unit 150 and stored in the storage unit 130.
- the detailed diagnosis unit 160 receives the data stored in the storage unit 130 via the data compression / decompression unit, and performs a detailed diagnosis.
- the detailed diagnosis is performed by, for example, clustering by vector quantization and setting the dimension of the input data vector and the total number of clusters created during learning to be large.
- a method other than vector quantization, another multivariate analysis method (for example, MT method), or a diagnostic method using a simulation result of the device may be used.
- the data compression / decompression unit 150 performs lossless compression / decompression. Thereby, the amount of data stored in the storage unit 130 can be further reduced.
- the data compression / decompression unit 150 may be omitted from the server computer 1002.
- the detailed diagnosis unit 160 may also be omitted from the server computer 1002.
- User 4 confirms the data diagnosis result and changes various setting parameters via the operation / display unit. That is, the diagnosis result of the data collection diagnostic compression unit 120, the compression strategy, and the compressed data are passed to the operation / display unit 170, so that the user 4 confirms or confirms the result. Update learning data and change compression strategy (change tolerance).
- diagnosis result of the detailed diagnosis unit 160 is also passed to the operation / display unit 170 so that the user 4 can check or check the learning data for the diagnosis of the detailed diagnosis unit 160 as a result of checking or the diagnosis algorithm to be executed. Switch.
- FIG. For example, if the detailed diagnosis result is normal, data compression is performed, and if an abnormal sign is detected, no compression is performed.
- the user 4 may check the diagnosis result of the data collection diagnosis compression unit 120 or the diagnosis result of the detailed diagnosis unit 160 via the operation / display unit 170, and the user 4 may instruct data compression.
- the input data of the data collection storage device 100 shows sensor information 201, 202,..., 20n installed in a single device 200, but the device is composed of a plurality of devices. It may be. Alternatively, it may be various types of logging data such as information other than the device main body, for example, sensing information of peripheral devices and piping parts, or sensing information of surrounding environment information. Further, it may be process data or event data. Furthermore, it is not limited to sensor information, but may be output information from other monitoring control devices.
- FIG. 20 shows a second mode of application of the data collection and storage device of the fifth embodiment to the system, and a second mode of application of the data collection and storage device to the system will be described with reference to FIG.
- the device 200 is provided with sensors 1, sensors 2,..., Sensor n (201, 202,..., 20n, where n is a positive integer). And a network 31.
- the data collection / compression device 10011 is connected to the server computer 1002 via the network 32.
- the data collection and compression apparatus 10011 includes a data collection diagnosis compression unit 120 and a data transmission / reception unit 141.
- the server computer 1002 includes a data transmission / reception unit 142, a data compression / decompression unit 150, a storage unit 130, a detailed diagnosis unit 160, and an operation / display unit 170.
- a time-series sensing information group by sensors installed in the device 200 becomes input data and is transmitted to the data collection / compression device 10011 via the network 31.
- the data collection diagnosis compression unit 120 performs the operation described in the first, second, or third embodiment on the transmitted input data, thereby Collect / predict abnormalities and compress data.
- the data compressed by the data collection diagnostic compression unit 120 is sent to the data transmission / reception unit 141 and transmitted to the server computer 1002 via the network 32.
- the network 31 is, for example, a field network used for monitoring and controlling devices, and is configured by a wired network such as CAN or Lonworks, a wireless sensor network such as ZigBee or ISA100.11a, or a mixture of these networks. However, other networks may be used.
- the individual sensors and the data transmission device 1001 may be connected on a one-to-one basis instead of a network.
- the network 32 is the above-described network, a high-speed network such as Ethernet (registered trademark) or an optical fiber, a WiFi-based wireless network such as IEEE802.11a / g / b / n, or an IEEE802.11s.
- a wireless mesh network, the Internet, a dedicated line, a public wireless network, a mobile phone network, a fixed telephone network, or a mixture of these networks may be used, but other networks may be used.
- the data transmission / reception units 141 and 142 are used in the data transmission device 1001 and the server computer 1002.
- the data transmission / reception units 141 and 142 are configured by a modem or a network module.
- the compressed data received by the data transmission / reception unit 142 is transferred to the data compression / decompression unit 150, further compressed, and stored in the storage unit 130.
- the detailed diagnosis unit 160 receives the data stored in the storage unit 130 via the data compression / decompression unit, and performs a detailed diagnosis.
- the detailed diagnosis is performed by, for example, clustering by vector quantization and setting the dimension of the input data vector and the total number of clusters created during learning to be large.
- a method other than vector quantization, another multivariate analysis method (for example, MT method), or a diagnostic method using a simulation result of the device may be used.
- the data compression / decompression unit 150 performs lossless compression / decompression. Thereby, the amount of data stored in the storage unit 130 can be further reduced.
- the data compression / decompression unit 150 may be omitted from the server computer 1002.
- the detailed diagnosis unit 160 may also be omitted from the server computer 1002.
- the diagnosis result of the data collection diagnosis compression unit 120, the compression strategy, and the compressed data are transferred to the operation / display unit 170 via the network 32 and the data transmission / reception unit 142, so that the user 4 confirms or confirms the data
- the learning data for diagnosis of the collected diagnosis compression unit 120 is updated, and the compression strategy is changed (change in allowable error).
- diagnosis result of the detailed diagnosis unit 160 is also passed to the operation / display unit 170 so that the user 4 can check or check the learning data for the diagnosis of the detailed diagnosis unit 160 as a result of checking or the diagnosis algorithm to be executed. Switch.
- FIG. For example, if the detailed diagnosis result is normal, data compression is performed, and if an abnormal sign is detected, no compression is performed.
- the user 4 may check the diagnosis result of the data collection diagnosis compression unit 120 or the diagnosis result of the detailed diagnosis unit 160 via the operation / display unit 170, and the user 4 may instruct data compression.
- the amount of data flowing through the network 32 is smaller than that in the fourth embodiment, so that the transmission speed is not high, for example, because the network 32 is a wireless network or a telephone line. It is effective in the case.
- the input data of the data collection storage device 100 shows sensor information 201, 202,..., 20n installed in a single device 200, but the device is composed of a plurality of devices. It may be. Alternatively, it may be various types of logging data such as information other than the device main body, for example, sensing information of peripheral devices and piping parts, or sensing information of surrounding environment information. Further, it may be process data or event data. Furthermore, it is not limited to sensor information, but may be output information from other monitoring control devices.
- FIG. 21 shows an application form of the data collection / storage device of the sixth embodiment to a heavy machine.
- FIG. 21 shows an application form of the data collection / storage device 100 to a heavy machine.
- the figure shows the form of application to a large dump truck.
- an engine speed measurement sensor, a coolant temperature measurement sensor, an exhaust pipe temperature measurement sensor, and a battery voltage measurement sensor (201, 202, 203, and 204, respectively) are installed in a large dump 2001.
- a network 31 a network 31.
- the data collection and storage device 100 notifies the operator 41 by displaying the diagnosis result on the display unit 180. Upon receiving the notification, the operator 41 can confirm the display contents and determine whether to continue or stop the operation of the device.
- FIG. 22 shows a screen configuration diagram of the display unit.
- FIG. 22 shows an application form to the large dump shown in FIG. 21.
- the entire screen 1801 is composed of a time transition graph display screen 1802, an abnormal sensor display screen 1803, and a warning message log display screen 1804.
- the time transition graph display screen 1802 displays the time transition of the device abnormality degree, allowable error, data compression ratio, and sensor abnormality contribution degree.
- the time transition of the abnormal contribution degree of the engine speed measurement sensor 201, the cooling water temperature measurement sensor 202, the exhaust pipe temperature measurement sensor 203, and the battery voltage measurement sensor 204 is displayed.
- the device abnormality degree and the abnormality contribution degree of various sensors are obtained by graphically displaying the table stored in the abnormality sign execution history storage unit in FIG.
- the data compression ratio C is obtained by the following calculation formula, for example.
- the device abnormality degree exceeds 1, and an example in which the abnormality contribution degree of the cooling water temperature is dominant is illustrated.
- the abnormality contribution degree of the sensor to be displayed may be such that the sensor that the user wants to display can be selected with a check box, for example.
- the top n items (n is a positive integer. For example, 3) having the highest degree of abnormal contribution may be displayed. Further, not only the abnormal contribution degree of the sensor but also the observation value of the sensor may be displayed in a graph.
- the abnormal sensor display screen 1803 highlights a sensor having a high degree of abnormal contribution when an abnormality occurs, such as color display or blinking.
- an abnormal contribution degree of the cooling water temperature is highlighted when an abnormality occurs.
- the warning message log display screen 1803 displays a message indicating that an abnormality has occurred.
- a warning message is displayed indicating that an abnormality has occurred at time te and the abnormal contribution of the cooling water temperature is dominant.
- FIG. 23 shows an application form of the data collection and compression apparatus of the seventh embodiment to heavy equipment.
- FIG. 23 shows an application form of the de overnight collecting and compressing apparatus 10011 to a heavy machine.
- FIG. 23 shows an application form of the data collection and storage device of the fifth embodiment shown in FIG.
- a large dump 2001 is provided with an engine speed measurement sensor, a cooling water temperature measurement sensor, an exhaust pipe temperature measurement sensor, and a battery voltage measurement sensor (201, 202, 203, and 204, respectively).
- 10011 is connected to the network 31.
- the data collection / compression apparatus 10011 notifies the operator 41 by displaying the diagnosis result on the display unit 180, and further transmits it to the server computer 1002 via the network 32.
- the server computer 1002 performs a detailed diagnosis on the data received from the data collection / compression device 10011 and displays the result on the operator 41 via the data collection / compression device 10011 to the display unit 180. As a result, the operator 41 can be notified of the diagnosis result in the server computer 1002.
- Data collection storage device 110 Data collection unit 111 Original data storage buffer 120 Data collection diagnosis compression unit 121, 1211 Failure sign diagnosis unit 124 Compression strategy management unit 125 Data compression unit 130 Storage unit 131 Failure sign Diagnosis execution history storage unit 132 Compression strategy history storage unit 133 Compressed data storage unit 134 Original data storage units 141 and 142 Data transmission / reception unit 150 Data compression / decompression unit 160 Detailed diagnosis unit 170 Operation / display unit 180 Display unit 200 Device 201 Engine rotation Number measurement sensor 202 Cooling water temperature measurement sensor 203 Exhaust pipe temperature measurement sensor 204 Battery voltage measurement sensor 1001 Data transmission device 1002 Server computer 1201 Diagnostic compression management unit 1251 Compressed data storage buffer 1252 Data expansion 1801 the entire screen 1802 hours transition graph display screen, 1803 Abnormal sensor display screen 1804 Alarm message log display screen 2001 Large dump 10011 Data collection and compression device 110 Data collection unit 111 Original data storage buffer 120 Data collection diagnosis compression unit 121, 1211 Failure sign diagnosis unit 124 Compression strategy management unit 125 Data compression unit 130 Storage unit 131 Failure sign Diagno
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Computer Security & Cryptography (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
Description
(1)多様な機器に対応可能となる、機器の知識を必要としない時系列データ圧縮のための許容誤差の設定。
(2)異常発生時のみならず、異常予兆が検知された期間のデータの抜け防止。
(3)設定した許容誤差の妥当性の確認。
データ収集格納装置の第1の実施形態の基本構成を図1に示す。
元データ格納バッファ111の実施の形態を図2に示す。
故障予兆診断部121の動作を図3,図4,図5を用いて説明する。
〔式1〕
d=√((x1-c1)^2+(x2-c2)^2)=√(d1^2
+d2^2)
(ただし、d1=|x1-c1|,d2=|x2-c2|)
この距離dが機器異常度であり、図5では、クラスタcの領域を逸脱している(即ち、d>1)ので、異常予兆が検知されたことになる。
〔式2〕
r1=d1/d、r2=d2/d
上記の通り、ベクトル量子化によるクラスタリングを用いることにより、正常時の入力データの分布領域からの逸脱の度合いを機器異常度として得ることが出来るので、機器の稼働時の時系列入力データを逐次異常予兆診断部に通すことにより、異常の予兆を検知することが可能になる。
圧縮戦略管理部124の動作を図6,図7を用いて説明する。
機器異常度>1(異常予兆検知区間)では、許容誤差=0(無圧縮)
即ち、入力データを診断した結果、正常区間では入力データを許容誤差mで圧縮し、異常予兆検知区間では許容誤差0あるいは圧縮しない。
(1)T≦teでは機器異常度≦1であるので、許容誤差=m
(2)te<Tでは機器異常度>1であるので、許容誤差=0
図7は、圧縮戦略管理部における許容誤差の設定例2を示したものである。圧縮戦略管理部124における、第2の実施の形態を、図7を用いて説明する。
(1)機器異常度≦1では、許容誤差=m
(2)しかし、機器異常度>1になった時点をteとすると、その時点から過去にtpの時間だけ遡った時点、即ちte-tp以降の許容誤差=0
従って、本図の例では、圧縮戦略管理部124が機器異常度に応じてデータ圧縮部125に設定する許容誤差は、以下の通りとなる。
(1)T≦te-tpで、許容誤差=m
(2)te-tp<Tで、許容誤差=0
上記は、機器異常度から許容誤差への変換は、許容誤差=1を境にして、許容誤差がmまたは0の何れかを取る例であるが、機器異常度の値に応じて許容誤差を連続的に変動させても構わない。
データ圧縮部125による圧縮後のデータを格納する圧縮データ格納部133の実施の形態を図8,図9,図10,図11を用いて説明する。なお、データ圧縮部125内の圧縮データ格納バッファ1251の構成も圧縮データ格納部133と基本的に同じである。
図12は、異常予兆診断実施履歴格納部の構成図を示すものであり、故障予兆診断実施履歴格納部131の実施の形態を、図12を用いて説明する。
故障予兆診断実施履歴格納部131の各時刻の機器異常度を参照することにより、機器が正常かあるいは異常予兆が検知されているか確認することができ、機器200のメンテナンスエンジニアは、異常予兆が検知されている時間帯を中心により詳細な診断を行うことができる。
図13は、圧縮戦略履歴格納部の構成図1を示すものであり、圧縮戦略履歴格納部132の第1の実施の形態を、図13を用いて説明する。
上記にて、故障予兆診断実施履歴格納部131には、誤差が閾値内の時刻は圧縮データが格納され、誤差が閾値を越える時刻は元データ(即ち、診断誤差無し)が格納されることになり、故障予兆診断実施履歴格納部131に格納されている全てのデータに対し、異常予兆診断の誤差は閾値以下になることを保証することができる。
機器異常度≦1(正常区間)で許容誤差=m
機器異常度>1(異常予兆検知区間)では、許容誤差=0
としている。
/Σ(センサのデータサイズ×過去丁区間の時間内のデータ数)
過去T区間の時間は、観測周期のm倍(mは正の整数。例えば1000)。
31,32 ネットワーク
41 オペレータ
100 データ収集格納装置
110 データ収集部
111 元データ格納バッファ
120 データ収集診断圧縮部
121,1211 故障予兆診断部
124 圧縮戦略管理部
125 データ圧縮部
130 格納部
131 故障予兆診断実施履歴格納部
132 圧縮戦略履歴格納部
133 圧縮データ格納部
134 元データ格納部
141,142 データ送受信部
150 データ圧縮・展開部
160 詳細診断部
170 操作・表示部
180 表示部
200 機器
201 エンジン回転数計測センサ
202 冷却水温度計測センサ
203 排気管温度計測センサ
204 バッテリ電圧計測センサ
1001 データ伝送装置
1002 サーバ計算機
1201 診断圧縮管理部
1251 圧縮データ格納バッファ
1252 データ展開部
1801 全体画面
1802 時間推移グラフ表示画面、
1803 異常センサ表示画面
1804 警報メッセージのログ表示画面
2001 大型ダンプ
10011 データ収集圧縮装置
Claims (14)
- 機器に設置した1個またはそれ以上のセンサから収集した時系列データに対して異常予兆診断を行い、その診断結果を用いて該データの圧縮戦略を管理することにより、該データをデータ圧縮手段により圧縮し、ストレージに格納することを特徴とする時系列データ診断圧縮方法。
- 前記異常予兆診断は、収集した時系列データに対し、予め正常時に機器から収集した時系列データ群を用いてベクトル量子化することにより分類したクラスタとの乖離の度合いを診断結果とすることを特徴とする請求項1記載の時系列データ診断圧縮方法。
- 前記データの圧縮戦略を管理することは、前記異常予兆診断が出力する診断結果より、データの許容誤差を算出することを特徴とする請求項1記載の時系列データ診断圧縮方法。
- 前記データ圧縮手段は、指定されたデータの許容誤差範囲のデータを間引き、許容誤差を逸脱したデータのみ残す非可逆データ圧縮を実施することを特徴とする請求項1記載の時系列データ診断圧縮方法。
- 前記データ圧縮手段は、前記非可逆データ圧縮実施後、さらに可逆圧縮を実施する多段の圧縮であることを特徴とする請求項4記載の時系列データ診断圧縮方法。
- 前記データの圧縮戦略を管理することは、算出するデータの許容誤差が、時系列データを構成する1個またはそれ以上のセンサデータ全てに対し、同一の許容誤差が設定されることであることを特徴とする請求項3記載の時系列データ診断圧縮方法。
- 前記算出するデータの許容誤差は、時系列データを構成する1個またはそれ以上のセンサデータに対し、個々のセンサ毎に独立した許容誤差が設定されることを特徴とする請求項6記載の時系列データ診断圧縮方法。
- 前記時系列データに対して異常予兆診断を行い、その診断結果を用いて該データの圧縮戦略を管理することにより、該データをデータ圧縮手段により圧縮した後、該圧縮データを展開し、展開したデータに対して異常予兆診断を行い、その診断結果と先に圧縮前のデータに対して実施した異常予兆診断の診断結果とを比較して得られる両者の誤差が、基準値以下であれば圧縮データを、または、基準値を超えていれば圧縮前のデータを、ストレージに格納することを特徴とする請求項1記載の時系列データ診断圧縮方法。
- 前記異常予兆診断が出力する診断結果をストレージに格納することを特徴とする請求項2に記載の時系列データ診断圧縮方法。
- 前記データの圧縮戦略を管理することが、出力するデータの許容誤差をストレージに格納することであることを特徴とする請求項3記載の時系列データ診断圧縮方法。
- 機器に設置した1個またはそれ以上のセンサから収集した時系列データをストレージに格納するのと同時に異常予兆診断を行い、その診断結果を用いてデータの許容誤差を算出し、算出した許容誤差をストレージに格納し、先にストレージに格納した該データと許容誤差を取り出し、取り出した該データをデータ圧縮手段により圧縮し、該圧縮したデータをストレージに格納することを特徴とする時系列データ診断圧縮方法。
- 機器に設置した1個またはそれ以上のセンサから時系列データを収集する手段と、収集した該データに対して異常予兆診断を実施する手段と、異常予兆診断の診断結果を用いて該データの圧縮戦略を管理する手段と、収集した該データを圧縮するデータ圧縮手段と、圧縮したデータを保管する格納手段と、を具備したことを特徴とする時系列データ収集格納装置。
- 機器に設置した1個またはそれ以上のセンサから時系列データを第1の伝送路より受信し、第2の伝送路に伝送するデータ送受信手段を有する第1の構成手段であるデータ伝送手段と、該データ伝送手段が伝送するデータを該第2の伝送路を介して受信するデータ送受信手段と、受信したデータより診断,圧縮するデータ収集圧縮手段と、該データ収集圧縮手段が出力するデータを圧縮・展開するデータを圧縮・展開手段と、該データの圧縮・展開する手段が圧縮したデータを格納するデータ格納手段と、該データ格納手段に格納された圧縮データを該データ圧縮・展開手段を介して得られる展開されたデータより詳細な診断を実施する詳細診断手段と、該データ収集圧縮手段あるいは、該詳細診断手段の出力を表示し、該手段に対して設定情報の操作を行う操作・表示部より構成される第2の構成手段であるサーバ計算機と、を有することを特徴とする時系列データ収集診断格納システム。
- 機器に設置した1個またはそれ以上のセンサから時系列データを第1の伝送路より受信し、受信したデータより診断,圧縮するデータ収集圧縮手段を介して圧縮したデータを第2の伝送路に伝送するデータ送受信手段を有する第1の構成手段であるデータ収集圧縮手段と、該データ収集圧縮手段が伝送する圧縮データを該第2の伝送路を介して受信するデータ送受信手段と、受信した該圧縮データを圧縮・展開するデータを圧縮・展開手段と、該データの圧縮・展開する手段が圧縮したデータを格納するデータ格納手段と、該データ格納手段に格納された圧縮データを該データ圧縮・展開手段を介して得られる展開されたデータより詳細な診断を実施する詳細診断手段と、該データ収集圧縮手段あるいは、該詳細診断手段の出力を表示し、該手段に対して設定情報の操作を行う操作・表示部より構成される第2の構成手段であるサーバ計算機と、を有することを特徴とする時系列データ収集診断格納システム。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2012512524A JP5435126B2 (ja) | 2010-04-26 | 2010-04-26 | 時系列データ診断圧縮方法 |
US13/643,588 US9189485B2 (en) | 2010-04-26 | 2010-04-26 | Time-series data diagnosing/compressing method |
PCT/JP2010/002968 WO2011135606A1 (ja) | 2010-04-26 | 2010-04-26 | 時系列データ診断圧縮方法 |
CN201080066461.9A CN102859457B (zh) | 2010-04-26 | 2010-04-26 | 时序数据诊断压缩方法 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2010/002968 WO2011135606A1 (ja) | 2010-04-26 | 2010-04-26 | 時系列データ診断圧縮方法 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2011135606A1 true WO2011135606A1 (ja) | 2011-11-03 |
Family
ID=44860961
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2010/002968 WO2011135606A1 (ja) | 2010-04-26 | 2010-04-26 | 時系列データ診断圧縮方法 |
Country Status (4)
Country | Link |
---|---|
US (1) | US9189485B2 (ja) |
JP (1) | JP5435126B2 (ja) |
CN (1) | CN102859457B (ja) |
WO (1) | WO2011135606A1 (ja) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2017084106A (ja) * | 2015-10-28 | 2017-05-18 | 株式会社 日立産業制御ソリューションズ | 気付き情報提供装置及び気付き情報提供方法 |
WO2018052015A1 (ja) * | 2016-09-14 | 2018-03-22 | 日本電気株式会社 | システムの分析支援装置、システムの分析支援方法及びプログラム |
JP2018101369A (ja) * | 2016-12-21 | 2018-06-28 | 千代田化工建設株式会社 | プラント設備診断システム |
JP2018173840A (ja) * | 2017-03-31 | 2018-11-08 | アズビル株式会社 | 時系列データ記録方法および装置 |
JP2018173772A (ja) * | 2017-03-31 | 2018-11-08 | アズビル株式会社 | 時系列データ記録方法および装置 |
JPWO2017217069A1 (ja) * | 2016-06-13 | 2019-03-14 | 株式会社日立製作所 | 回転機診断装置、回転機診断方法及び回転機診断プログラム |
JP2019046386A (ja) * | 2017-09-06 | 2019-03-22 | 日本電信電話株式会社 | 故障検知モデル構築装置、故障検知モデル構築方法及びプログラム |
JP2019067139A (ja) * | 2017-09-29 | 2019-04-25 | エヌ・ティ・ティ・コミュニケーションズ株式会社 | 監視装置、監視方法、監視プログラム、表示装置、表示方法および表示プログラム |
WO2019123510A1 (ja) * | 2017-12-18 | 2019-06-27 | 三菱電機株式会社 | 表示制御装置、表示システム、表示装置、表示方法および表示プログラム |
JP2019199817A (ja) * | 2018-05-15 | 2019-11-21 | 三菱重工業株式会社 | 故障診断システム及び故障診断方法 |
CN111370122A (zh) * | 2020-02-27 | 2020-07-03 | 西安交通大学 | 一种基于知识指导的时序数据风险预测方法、***及其应用 |
WO2022191164A1 (ja) * | 2021-03-12 | 2022-09-15 | 住友重機械工業株式会社 | 表示装置、表示方法、制御装置及びコンピュータプログラム |
JP7484261B2 (ja) | 2020-03-17 | 2024-05-16 | 富士電機株式会社 | 情報処理装置、情報処理方法及び情報処理プログラム |
JP7500980B2 (ja) | 2020-02-04 | 2024-06-18 | 富士電機株式会社 | 情報処理装置、情報処理方法及び情報処理プログラム |
Families Citing this family (53)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP6178591B2 (ja) * | 2013-03-06 | 2017-08-09 | 富士機械製造株式会社 | 工具異常判別システム |
GB201313205D0 (en) * | 2013-07-24 | 2013-09-04 | Ent Ltd | Method and system for condition monitoring |
US10373259B1 (en) | 2014-05-20 | 2019-08-06 | State Farm Mutual Automobile Insurance Company | Fully autonomous vehicle insurance pricing |
US10089693B1 (en) | 2014-05-20 | 2018-10-02 | State Farm Mutual Automobile Insurance Company | Fully autonomous vehicle insurance pricing |
US9972054B1 (en) | 2014-05-20 | 2018-05-15 | State Farm Mutual Automobile Insurance Company | Accident fault determination for autonomous vehicles |
US10599155B1 (en) | 2014-05-20 | 2020-03-24 | State Farm Mutual Automobile Insurance Company | Autonomous vehicle operation feature monitoring and evaluation of effectiveness |
US11669090B2 (en) | 2014-05-20 | 2023-06-06 | State Farm Mutual Automobile Insurance Company | Autonomous vehicle operation feature monitoring and evaluation of effectiveness |
US10540723B1 (en) | 2014-07-21 | 2020-01-21 | State Farm Mutual Automobile Insurance Company | Methods of providing insurance savings based upon telematics and usage-based insurance |
US10241509B1 (en) | 2014-11-13 | 2019-03-26 | State Farm Mutual Automobile Insurance Company | Autonomous vehicle control assessment and selection |
US20160155098A1 (en) * | 2014-12-01 | 2016-06-02 | Uptake, LLC | Historical Health Metrics |
US20160188631A1 (en) * | 2014-12-30 | 2016-06-30 | Alcatel-Lucent Usa Inc. | Locating call measurement data |
IN2015CH01270A (ja) * | 2015-03-13 | 2015-04-10 | Wipro Ltd | |
JP6474894B2 (ja) * | 2015-06-09 | 2019-02-27 | 株式会社日立製作所 | データ収集システムおよび方法、計測データ量の削減方法 |
US20210166323A1 (en) | 2015-08-28 | 2021-06-03 | State Farm Mutual Automobile Insurance Company | Determination of driver or vehicle discounts and risk profiles based upon vehicular travel environment |
TWI584232B (zh) * | 2016-01-19 | 2017-05-21 | 台達電子工業股份有限公司 | 區域異常偵測系統及區域異常偵測方法 |
US10295363B1 (en) | 2016-01-22 | 2019-05-21 | State Farm Mutual Automobile Insurance Company | Autonomous operation suitability assessment and mapping |
US11441916B1 (en) | 2016-01-22 | 2022-09-13 | State Farm Mutual Automobile Insurance Company | Autonomous vehicle trip routing |
US11242051B1 (en) | 2016-01-22 | 2022-02-08 | State Farm Mutual Automobile Insurance Company | Autonomous vehicle action communications |
US10395332B1 (en) | 2016-01-22 | 2019-08-27 | State Farm Mutual Automobile Insurance Company | Coordinated autonomous vehicle automatic area scanning |
US11719545B2 (en) | 2016-01-22 | 2023-08-08 | Hyundai Motor Company | Autonomous vehicle component damage and salvage assessment |
US10324463B1 (en) | 2016-01-22 | 2019-06-18 | State Farm Mutual Automobile Insurance Company | Autonomous vehicle operation adjustment based upon route |
US10134278B1 (en) | 2016-01-22 | 2018-11-20 | State Farm Mutual Automobile Insurance Company | Autonomous vehicle application |
WO2017172635A1 (en) * | 2016-03-31 | 2017-10-05 | Osisoft, Llc | Event based data collection, integration, and presentation |
CN106341842B (zh) * | 2016-08-23 | 2020-05-01 | 北京邮电大学 | 一种无线传感器网络数据的传输方法及装置 |
JP6842299B2 (ja) * | 2016-12-28 | 2021-03-17 | 三菱パワー株式会社 | 診断装置、診断方法及びプログラム |
TWI711911B (zh) * | 2018-03-20 | 2020-12-01 | 日商住友重機械工業股份有限公司 | 異常監視裝置及異常監視方法 |
CN108594788B (zh) * | 2018-03-27 | 2020-09-22 | 西北工业大学 | 一种基于深度随机森林算法的飞机作动器故障检测与诊断方法 |
WO2019199729A1 (en) * | 2018-04-09 | 2019-10-17 | The Government Of The United States Of America, As Represented By The Secretary Of The Navy | Reducing error in data compression |
JP6887407B2 (ja) * | 2018-08-28 | 2021-06-16 | 横河電機株式会社 | 装置、プログラム、プログラム記録媒体、および方法 |
JP6615963B1 (ja) * | 2018-08-31 | 2019-12-04 | 株式会社日立パワーソリューションズ | 異常予兆診断装置及び異常予兆診断方法 |
JP7406915B2 (ja) * | 2018-11-02 | 2023-12-28 | 三菱重工業株式会社 | 単位空間更新装置、単位空間更新方法、及びプログラム |
US10805150B2 (en) | 2018-12-04 | 2020-10-13 | Nokia Solutions And Networks Oy | Regenerative telemetry method for resource reduction |
JP7229861B2 (ja) * | 2019-06-20 | 2023-02-28 | 株式会社日立製作所 | 故障予兆診断装置およびその方法 |
US11748626B2 (en) | 2019-08-12 | 2023-09-05 | Micron Technology, Inc. | Storage devices with neural network accelerators for automotive predictive maintenance |
US11586943B2 (en) | 2019-08-12 | 2023-02-21 | Micron Technology, Inc. | Storage and access of neural network inputs in automotive predictive maintenance |
US11635893B2 (en) | 2019-08-12 | 2023-04-25 | Micron Technology, Inc. | Communications between processors and storage devices in automotive predictive maintenance implemented via artificial neural networks |
US11853863B2 (en) * | 2019-08-12 | 2023-12-26 | Micron Technology, Inc. | Predictive maintenance of automotive tires |
US11775816B2 (en) | 2019-08-12 | 2023-10-03 | Micron Technology, Inc. | Storage and access of neural network outputs in automotive predictive maintenance |
US11586194B2 (en) | 2019-08-12 | 2023-02-21 | Micron Technology, Inc. | Storage and access of neural network models of automotive predictive maintenance |
US11702086B2 (en) | 2019-08-21 | 2023-07-18 | Micron Technology, Inc. | Intelligent recording of errant vehicle behaviors |
US11650746B2 (en) | 2019-09-05 | 2023-05-16 | Micron Technology, Inc. | Intelligent write-amplification reduction for data storage devices configured on autonomous vehicles |
US11693562B2 (en) | 2019-09-05 | 2023-07-04 | Micron Technology, Inc. | Bandwidth optimization for different types of operations scheduled in a data storage device |
US10840943B1 (en) * | 2019-10-15 | 2020-11-17 | EMC IP Holding Company LLC | System and method of data compression between backup server and storage |
US11250648B2 (en) | 2019-12-18 | 2022-02-15 | Micron Technology, Inc. | Predictive maintenance of automotive transmission |
US11531339B2 (en) | 2020-02-14 | 2022-12-20 | Micron Technology, Inc. | Monitoring of drive by wire sensors in vehicles |
US11709625B2 (en) | 2020-02-14 | 2023-07-25 | Micron Technology, Inc. | Optimization of power usage of data storage devices |
CN111311427B (zh) * | 2020-02-18 | 2021-03-26 | 广东大橘果业有限公司 | 基于互联网的种植智能管理*** |
EP3936957B1 (en) * | 2020-07-08 | 2023-08-23 | Bull SAS | Error-based method for calculating a remaining useful life of an apparatus |
US11864359B2 (en) * | 2020-08-27 | 2024-01-02 | Nvidia Corporation | Intelligent threshold leak remediaton of datacenter cooling systems |
JP2022094017A (ja) * | 2020-12-14 | 2022-06-24 | トヨタ自動車株式会社 | 車載システム |
JP7294361B2 (ja) * | 2021-03-23 | 2023-06-20 | 横河電機株式会社 | フィールド機器、測定方法、及びプログラム |
CN116910006B (zh) * | 2023-07-24 | 2024-03-29 | 深圳市盛弘新能源设备有限公司 | 基于新能源电池的数据压缩存储处理方法及*** |
CN117235557B (zh) * | 2023-11-14 | 2024-02-02 | 山东贺铭电气有限公司 | 基于大数据分析的电气设备故障快速诊断方法 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2000259222A (ja) * | 1999-03-04 | 2000-09-22 | Hitachi Ltd | 機器監視・予防保全システム |
JP2003015734A (ja) * | 2001-07-02 | 2003-01-17 | Toshiba Corp | 時系列データ圧縮方法および時系列データ格納装置およびプログラム |
JP2004030491A (ja) * | 2002-06-28 | 2004-01-29 | Hitachi Industries Co Ltd | 機器メンテナンス診断システム |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP3568984B2 (ja) * | 1994-06-20 | 2004-09-22 | 株式会社日立製作所 | 情報の再生方法、再生装置、出力方法及び出力装置 |
US6532440B1 (en) * | 1998-10-16 | 2003-03-11 | Fujitsu Limited | Multiple error and fault diagnosis based on Xlists |
US7107257B2 (en) * | 2001-11-05 | 2006-09-12 | Lenovo (Singapore) Pte. Ltd. | Consolidated monitoring system and method using the internet for diagnosis of an installed product set on a computing device |
CN1277181C (zh) * | 2003-06-25 | 2006-09-27 | 中国科学院计算技术研究所 | 一种单输出无反馈时序测试响应压缩电路 |
JP2006276924A (ja) | 2005-03-28 | 2006-10-12 | Hitachi Ltd | 設備機器診断装置及び設備機器診断プログラム |
JP4858031B2 (ja) * | 2006-09-13 | 2012-01-18 | コニカミノルタビジネステクノロジーズ株式会社 | 画像形成装置 |
JP4719667B2 (ja) | 2006-12-28 | 2011-07-06 | 日立オートモティブシステムズ株式会社 | 時系列データ圧縮方法 |
-
2010
- 2010-04-26 CN CN201080066461.9A patent/CN102859457B/zh not_active Expired - Fee Related
- 2010-04-26 WO PCT/JP2010/002968 patent/WO2011135606A1/ja active Application Filing
- 2010-04-26 JP JP2012512524A patent/JP5435126B2/ja not_active Expired - Fee Related
- 2010-04-26 US US13/643,588 patent/US9189485B2/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2000259222A (ja) * | 1999-03-04 | 2000-09-22 | Hitachi Ltd | 機器監視・予防保全システム |
JP2003015734A (ja) * | 2001-07-02 | 2003-01-17 | Toshiba Corp | 時系列データ圧縮方法および時系列データ格納装置およびプログラム |
JP2004030491A (ja) * | 2002-06-28 | 2004-01-29 | Hitachi Industries Co Ltd | 機器メンテナンス診断システム |
Cited By (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2017084106A (ja) * | 2015-10-28 | 2017-05-18 | 株式会社 日立産業制御ソリューションズ | 気付き情報提供装置及び気付き情報提供方法 |
JPWO2017217069A1 (ja) * | 2016-06-13 | 2019-03-14 | 株式会社日立製作所 | 回転機診断装置、回転機診断方法及び回転機診断プログラム |
WO2018052015A1 (ja) * | 2016-09-14 | 2018-03-22 | 日本電気株式会社 | システムの分析支援装置、システムの分析支援方法及びプログラム |
US11455203B2 (en) | 2016-09-14 | 2022-09-27 | Nec Corporation | Abnormality detection support device, abnormality detection support method, and program |
JP7183790B2 (ja) | 2016-09-14 | 2022-12-06 | 日本電気株式会社 | システムの分析支援装置、システムの分析支援方法及びプログラム |
JPWO2018052015A1 (ja) * | 2016-09-14 | 2019-06-24 | 日本電気株式会社 | システムの分析支援装置、システムの分析支援方法及びプログラム |
JP2018101369A (ja) * | 2016-12-21 | 2018-06-28 | 千代田化工建設株式会社 | プラント設備診断システム |
JP2018173840A (ja) * | 2017-03-31 | 2018-11-08 | アズビル株式会社 | 時系列データ記録方法および装置 |
JP2018173772A (ja) * | 2017-03-31 | 2018-11-08 | アズビル株式会社 | 時系列データ記録方法および装置 |
JP2019046386A (ja) * | 2017-09-06 | 2019-03-22 | 日本電信電話株式会社 | 故障検知モデル構築装置、故障検知モデル構築方法及びプログラム |
JP2019067139A (ja) * | 2017-09-29 | 2019-04-25 | エヌ・ティ・ティ・コミュニケーションズ株式会社 | 監視装置、監視方法、監視プログラム、表示装置、表示方法および表示プログラム |
JP7019364B2 (ja) | 2017-09-29 | 2022-02-15 | エヌ・ティ・ティ・コミュニケーションズ株式会社 | 監視装置、監視方法、監視プログラム、表示装置、表示方法および表示プログラム |
WO2019123510A1 (ja) * | 2017-12-18 | 2019-06-27 | 三菱電機株式会社 | 表示制御装置、表示システム、表示装置、表示方法および表示プログラム |
JPWO2019123510A1 (ja) * | 2017-12-18 | 2020-02-27 | 三菱電機株式会社 | 表示制御装置、表示システム、表示装置、表示方法および表示プログラム |
US11366627B2 (en) | 2017-12-18 | 2022-06-21 | Mitsubishi Electric Corporation | Display control device, display system, display device, and display method |
JP2019199817A (ja) * | 2018-05-15 | 2019-11-21 | 三菱重工業株式会社 | 故障診断システム及び故障診断方法 |
JP7160560B2 (ja) | 2018-05-15 | 2022-10-25 | 三菱重工業株式会社 | 故障診断システム及び故障診断方法 |
JP7500980B2 (ja) | 2020-02-04 | 2024-06-18 | 富士電機株式会社 | 情報処理装置、情報処理方法及び情報処理プログラム |
CN111370122A (zh) * | 2020-02-27 | 2020-07-03 | 西安交通大学 | 一种基于知识指导的时序数据风险预测方法、***及其应用 |
CN111370122B (zh) * | 2020-02-27 | 2023-12-19 | 西安交通大学 | 一种基于知识指导的时序数据风险预测方法、***及其应用 |
JP7484261B2 (ja) | 2020-03-17 | 2024-05-16 | 富士電機株式会社 | 情報処理装置、情報処理方法及び情報処理プログラム |
WO2022191164A1 (ja) * | 2021-03-12 | 2022-09-15 | 住友重機械工業株式会社 | 表示装置、表示方法、制御装置及びコンピュータプログラム |
Also Published As
Publication number | Publication date |
---|---|
CN102859457A (zh) | 2013-01-02 |
CN102859457B (zh) | 2015-11-25 |
US9189485B2 (en) | 2015-11-17 |
US20130097128A1 (en) | 2013-04-18 |
JPWO2011135606A1 (ja) | 2013-07-18 |
JP5435126B2 (ja) | 2014-03-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5435126B2 (ja) | 時系列データ診断圧縮方法 | |
US10754721B2 (en) | Computer system and method for defining and using a predictive model configured to predict asset failures | |
JP2018524704A (ja) | 予測モデルの動的な実行 | |
US20160371406A1 (en) | Dynamic Execution of Predictive Models & Workflows | |
US10254751B2 (en) | Local analytics at an asset | |
JP2015114778A (ja) | プラント制御システム、制御装置、管理装置、及びプラント情報処理方法 | |
JP2018519594A (ja) | 資産でのローカルな解析 | |
JPWO2016199210A1 (ja) | データ収集システムおよび方法、計測データ量の削減方法 | |
JP2023013797A (ja) | 画面生成方法、画面生成装置、およびプログラム | |
JP2023013798A (ja) | 画面生成方法、画面生成装置、およびプログラム |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 201080066461.9 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 10850630 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2012512524 Country of ref document: JP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13643588 Country of ref document: US |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 10850630 Country of ref document: EP Kind code of ref document: A1 |