Back to results
US 10,496,056 B2Granted
Measured value analysis apparatus and measured value analysis method
AssigneesIHI Corporation, Diesel United, Ltd.
InventorTakashi Fujii
Life of this patent
Estimating the term…
Abstract
A measured value analysis apparatus includes a controller and a memory, in which the controller: selects one examination object item, which is an examination object, and a plurality of related items which are directly or indirectly related to the examination object item; creates a correlation model between measured values from past measured values of the examination object item and the plurality of related items; and calculates a predicted value of the examination object item by applying the measured values of the related items to the correlation model and calculates divergence between an actual measured value of the examination object item and the predicted value.
Claims7
A measured value analysis apparatus for analyzing a vessel on the basis of a measured value of the vessel, comprising:
- a controller and a memory, the controller executing a program stored in the memory to:
- select a specified related item from among a plurality of related items, each of which is related to an examination object item, wherein the examination object item is a physical quantity influenced by a structure of the vessel and an unmeasured environmental factor other than the plurality of related items, each of the plurality of related items is an item which influences an operating source of the vessel, and the plurality of related items includes a fuel input amount, a supercharging pressure, a fuel heating value, a scavenging pressure, a supercharger rotational speed, a load index, an air temperature, and a supercharger exhaust gas temperature;
- read respective past measured values of the examination object item, the operating source, and the specified related item from the memory and create a correlation model between the examination object item, the operating source, and the specified related item on the basis of the past measured values;
- calculate a predicted value of the operating source by applying the measured values of the specified related item to the correlation model and calculate divergence between an actual measured value of the operating source and the predicted value;
- compare the calculated divergence with a specified threshold value;
- finalize the correlation model when the calculated divergence is equal to or less than the specified threshold value as a result of the comparison;
- re-select the specified related item, then create the correlation model, and further execute the calculation and the comparison of the divergence when the calculated divergence exceeds the specified threshold value as a result of the comparison;
- calculate a predicted value of the physical quantity by applying the respective measured values of the operating source and the specified related item to the finalized correlation model;
- compare an actual measured value of the physical quantity and the predicted value of the physical quantity;
- set a result of the comparison of the actual measured value of the physical quantity and the predicted value of the physical quantity as an influence of the unmeasured environmental factor;
- and
- exclude the influence of the unmeasured environmental factor from a predicted value of the examination object.
9 more elements
A measured value analysis method for analyzing measured values of a plurality of measurement items, comprising:
- selecting a specified related item from among a plurality of related items, each of which is related to an examination object item, wherein the examination object item is a physical quantity influenced by a structure of the vessel and an unmeasured environmental factor, each of the plurality of related items is an item which influences an operating source of the vessel, and the plurality of related items includes a fuel input amount, a supercharging pressure, a fuel heating value, a scavenging pressure, a supercharger rotational speed, a load index, an air temperature, and a supercharger exhaust gas temperature;
- reading respective past measured values of the examination object item, the operating source, and the specified related item from the memory and creating a correlation model between the examination object item, the operating source, and the specified related item on the basis of the past measured values;
- calculating a predicted value of the operating source by applying the measured values of the specified related item to the correlation model and calculating divergence between an actual measured value of the operating source and the predicted value;
- comparing the calculated divergence with a specified threshold value;
- finalizing the correlation model when the calculated divergence is equal to or less than the specified threshold value as a result of the comparison;
- re-selecting the specified related item, then creating the correlation model, and further executing the calculation and the comparison of the divergence when the calculated divergence exceeds the specified threshold value as a result of the comparison;
- calculating a predicted value of the physical quantity by applying the respective measured values of the operating source and the specified related item to the finalized correlation model;
- comparing an actual measured value of the physical quantity and the predicted value of the physical quantity;
- setting a result of the comparison of the actual measured value of the physical quantity and the predicted value of the physical quantity as an influence of the unmeasured environmental factor;
- and
- excluding the influence of the unmeasured environmental factor from a predicted value of the examination object.
8 more elements
Dependencies are inferred from the claim text.