Patent Application
20180078963
The present invention relates to the field of monitoring changes in the performance of spray nozzles feeding a feed material to be spray dried into a spray drying chamber or feeding a feed material to be spray cooled into a spray cooling chamber.
Spray drying and spray cooling is a well-known process for turning a liquid product into a powder product. The drying or cooling takes place in a drying or cooling chamber, in which the drying air or cooling air is brought into contact with the feed, i.e. the liquid being atomized into droplets, by one or more nozzles or other forms of atomizer, such as a rotary atomizer. The drying or cooling air enters the drying or cooling chamber via an air disperser positioned at or near the top of the chamber. For spray drying systems, further drying means may be provided in such drying system, for instance a fluidized bed.
In particular when atomizing the liquid into droplets by the use of nozzles, e.g. pressure nozzles or two-fluid nozzles, it is important that the nozzles do spray as intended, as otherwise a non-uniform process will be performed resulting in for example formation of deposits on the nozzle itself and on the wall inside of the chamber, and in the production of a product of less quality and a reduced capacity. As a consequence, lumps may occur in the product, and also, the risk of formation of deposits may lead to an exaggerated heating and in worst case fire in this deposit, as being in a heated zone for a longer time than anticipated.
It is desirable to ensure that leaks in nozzles in spray dryers or spray coolers that can disrupt the droplet dispersal within the chamber are detected, and to reduce the build-up of solid matter that can make the chamber overheat and, in extreme cases, may even cause fire and an explosion. Traditionally, fire or explosion detection and extinguishing or relieving devices are provided in such spray systems. Monitoring and surveillance of various processes are known in the art. This monitoring is traditionally carried out in order to ascertain correct operational temperatures of for instance the output of an outlet such as a nozzle, or to ensure that the item to be sprayed has the correct temperature. EP 2 504 090 discloses such a temperature monitoring system comprising one or more infrared cameras adapted for measuring the temperature within a predefined area in a spray dryer system.
As the final product produced must be according to specifications as to quality and particle size, and as sources for starting fires and/or dust explosions in spray systems must be eliminated as far as possible for obvious reasons, the spraying process must be monitored. Monitoring systems using built-in visual TV cameras have been developed for which the TV cameras are used to give the operator an actual picture of the nozzle spraying performance. This means that an operator must watch a monitor all the time during processing. As many spray plants are running 24 hours a day for periods, a manual monitoring all the time is time consuming in terms of man-power. This is a disadvantage.
Thus, there is a need for an image analyzing system, which automatically detects changes in nozzle spraying performance and automatically issues warning or notifications when such changes are detected.
It is an object of the present invention to provide a monitoring method and system, which brings a solution to the problem of automatically issuing warning or notifications when changes in nozzle spraying performance appear.
The foregoing and other objects are achieved by the features of the independent claims. Further implementation forms are apparent from the dependent claims, the description and figures.
According to a first aspect, there is provided a method for monitoring deposit build-up on spray nozzles in a spray drying or spray cooling chamber, said method comprising:
It is within an embodiment of the invention that a notification or an error notification may be issued based on the detected differences. It is within an embodiment of the invention that the notification or error notification is issued when a predetermined minimum difference is detected or determined for each of a predetermined number of recorded process images. Here, the predetermined number of recorded images may be obtained at different points in time. The predetermined number of recorded images may preferably be obtained at subsequent points in time or at consecutive points in time.
According to an embodiment of the first aspect, then the step of detecting differences within the first interest area between the reference image and each of a number of process images being compared to the reference image comprises:
According to an embodiment of the present invention, a process image, for which the determined number of error pixels within the same area of interest reaches or exceeds a predetermined minimum number of error pixels may be classified as an error image.
According to an embodiment of the present invention, a warning or notification may be issued based on one more process images being classified as error images with respect to the same image area of interest. It is also within an embodiment of the invention that a warning or notification may be issued based on a determined number of process images being classified as error images for the same image area of interest.
Thus, the issuance of a notification based on the detected differences may comprise:
It should be understood that the present invention covers embodiments for which the camera is a digital still image camera and embodiments for which the camera is a digital video camera taking a stream of images, where only a part of the streamed images are recorded and compared with the reference image.
It is within an embodiment of the invention that
It should be understood that the distance from a camera to different arranged spray nozzles will vary, thus the number of image pixels taken up in an image differs with the distance between the camera and the spray nozzle.
The present invention therefore also covers an embodiment, wherein
According to an embodiment of the invention the detection of a pixel difference between similar addressed pixels in the reference image and the process image is at least partly based on a difference in pixel intensity. It is preferred that the predetermined minimum pixel difference is defined as a minimum difference in pixel intensity between similar addressed pixels in the reference image and the process image. The minimum difference in pixel intensity may be set to at least 25%, such as at least 30%, such as at least 35%, or such as at least 39% of the maximum intensity when measuring the pixel intensity on a scale from 0% to 100%. The images may be represented as gray scale images when measuring the pixel intensity, and the pixel intensity may be given as a binary bit number, such as a 7 or 8-bit number, and the minimum difference in pixel intensity is set to a number representing a percentage of the maximum binary bit number corresponding to the predetermined minimum percentage difference in pixel intensity.
It is also within one or more embodiments of the invention that the detection of a pixel difference between similar addressed pixels in the reference image and the process image is at least partly based on a difference in pixel color. The images may be represented as RGB color images, and the predetermined minimum pixel difference may be defined as a minimum difference in pixel intensity of one of the RGB colors for similar addressed pixels in the reference image and the process image. The minimum difference in pixel intensity for one color may be set to at least 25%, such as at least 30%, such as at least 35%, or such as at least 39% of the maximum intensity when measuring the pixel intensity on a scale from 0% to 100%.
The present invention also covers one or more embodiments wherein for an image area of interest covering at least a part of a spray nozzle, the predetermined minimum number of error pixels for said area of interest of the reference image is determined based on the number of pixels representing a characteristic measure of the nozzle. Here, the characteristic measure of the nozzle may be the width of the nozzle. It is preferred that for an area of interest covering part of a spray nozzle the predetermined minimum number of error pixels is larger than the number of pixels representing the characteristic measure of the nozzle. The predetermined minimum number of error pixels may be at least 1.5 times larger, such as at least 2 times larger, such as at least 2.5 times larger, or such as at least 3 times larger than the number of pixels representing the characteristic measure of the nozzle.
According to an embodiment of the invention, the detection of a difference between the reference image and a process image is at least partly based on spectral analysis. Here, a frequency spectrum may be determined for both the reference image and the process image, and the obtained frequency spectra may be compared. A predetermined change in the number of detected frequencies and/or a predetermined change in the power spectrum may be defined as representing a predetermined minimum difference for a process image to be an error image. A notification may be issued when the predetermined minimum difference is determined for each of a predetermined number of recorded process images. The minimum difference in the number of detected frequencies may be set to just a single frequency.
The predetermined number of obtained and/or recorded process images may be selected within a wide range of numbers, depending on the spray nozzle setup. The invention covers embodiments, wherein the predetermined number is at least one, such as at least two, such as at least three, such as at least five, or such as at least ten. However, it is preferred that it is in the range of 1 to 50, such as 1, 2, 3, 5, 10, 20, or 50.
It is within an embodiment of the invention that a single digital optical camera is used for obtaining the reference and process images. However, the invention also covers embodiments where several digital optical cameras are used, and where a reference image and corresponding process images are obtained for each camera. In a preferred embodiment 8 or up to 8 cameras are used, each covering a different angular view of an air inlet zone.
In order to provide sufficient light for obtaining images, it is preferred that a number of lamps providing light in the visible frequency spectrum is arranged in connection with or connected to at least one of the digital optical cameras. The present invention covers one or more embodiments for which the feed material is fed into the spray drying chamber through several spray nozzles, such as up to 28 nozzles. It is preferred that for at least one spray nozzle for which a part is covered by an area of interest, an output part for outputting a spray of feed material is covered by said area of interest.
It is within one or more embodiments of the present invention that the method comprises a pre-filtering step being performed on detected pixel differences within an area of interest for an obtained process image before determining the number of error pixels, said pre-filtering comprising filtering out detected pixel differences recognized as one or more predetermined shapes in the process image.
The pre-filtering may be a shape filtering whereby pixel differences due to error shapes which have no connection with deposit build-up may be filtered out in order not to generate false error images, where an error image is an image for which the detected number of error pixels has reached the predetermined minimum number of error pixels. The movement of a nozzle may result in pixel differences identified as a substantially rectangular shape in a process image when compared to the reference image, and these types of shapes may be accepted for the process images, and may therefore be filtered out in order not to generate error images. The size of the substantially rectangular shape may have a height close to the width of the nozzle width, and a width equal to the movement of the nozzle, which may be in the range of 10-30% of the nozzle width, such as about 20% of the nozzle width.
A light source may create a pixel difference between compared images, where the pixel difference can be identified as an elongated error shape due to a curved surface, and such error shapes may also be generalized as a substantially rectangular shape, and these types of shapes may be accepted for the process images, and may therefore be filtered out in the pre-filtering step in order not to generate error images. Shapes in pixel differences such as trapezoid or isosceles trapezoid may reflects the jet or spray and may also be accepted and filtered out in the pre-filtering step.
Thus, for the pre-filtering step it is preferred that the one or more predetermined shapes comprise a substantially rectangular shape and/or a trapezoid or isosceles trapezoid shape. Here, the predetermined shaped may be a substantially rectangular shape having a height substantially equal to the width of the nozzle width, and a width substantially equal to 10-30% of the nozzle width, such as about 20% of the nozzle width.
For the present invention it is preferred that a computer control system is provided, and that the computer control system is configured to control or perform the steps of image recording, image comparison, image difference detection and error notification according to any one of the above described embodiments of the method of the invention. The computer control system may also be configured to control the step of pre-filtering. Thus, it is within one or more embodiments of the invention that a computer control system is provided and configured to store or record images obtained by the digital camera(s), the computer control system further being configured to perform the steps of:
According to the present invention there is also provided a spray dryer or spry cooling system comprising:
It is preferred that the one or more cameras are configured for obtaining images in the visible frequency spectrum. It is also preferred that the system of the invention further comprises a number of light emitting lamps, which lamps may provide light in the visible frequency spectrum. The one or more lamps may be arranged for providing light to an area comprising a spray nozzle output. It is preferred that one or more lamps are arranged in connection with or connected to at least one of the digital optical cameras.
For the system of the invention it is preferred that the system comprises several spray nozzles each having an output for feeding a feed material to be spray dried or spray cooled into the chamber. The system may comprise a single digital optical camera being arranged for obtaining the images, but it also within an embodiment of the invention that several digital optical cameras are arranged for obtaining the images. Here, the cameras may cover different angular views.
It is within one or more embodiments of the system of the invention that the image area obtained by the one or more cameras comprises one or more areas of interest, whereby at least a part of each nozzle is covered by at least one area of interest. It is preferred that for at least one spray nozzle for which a part is covered by an area of interest, an output part for outputting a spray of feed material is covered by the area of interest.
The system of the invention also covers one or more embodiments wherein the computer control system is configured to control image recording, image comparison, image difference detection and error notification as described in connection with any one of the methods of the present invention.
The system 100 of
The system 100 further comprises a number of digital optical cameras 108a,b,c,d for obtaining images in the visible frequency spectrum, and arranged at different positions in connection with the chamber 101. Each camera 108a,b,c,d is arranged with a visual view for obtaining images having an image area comprising at least a first image area of interest covering a part of the spray nozzle 104. It is preferred that the first image area covers the spray nozzle output 104 and an area adjacent the spray nozzle output 104. However, it is also within embodiments of the invention that one or more of the cameras 108a,b,c,d have an image area with a first image area of interest not covering the spray nozzle but covering the area adjacent the spray nozzle output 104.
The chamber 101 may also be provided with several spray nozzles 104, which each may be arranged at the end of a rod or lance 103. When having several spray nozzles 104, it is preferred that at least a part of each spray nozzle 104, such as the output of the spray nozzle 104, is covered by one or more image areas of interest within the image area obtained by at least one camera. It is within a preferred embodiment of the invention that at least a part of each spray nozzle 104 is covered by an image area of interest within the image area covered by images obtained by one camera. According to an embodiment of the invention, the chamber 101 may be provided with up to 28 spray nozzles or even more if needed.
It is preferred that all nozzles 104 in the system 100 are of known size, and that all nozzles 104 have the same size. The spray nozzles 104 may move during operation due to air impact, and the nozzles 104 may move approximately 0-30% of the diameter of the nozzle lance 103.
In order to optimize monitoring of the performance of the spray nozzles 104, it is preferred that the system 100 is provided with several cameras 108a,b,c,d. These cameras 108a,b,c,d may then be arranged to cover different angular views of the spray nozzles 104. It is preferred that for at least one spray nozzle 104, the cameras 108a,b,c,d are arranged so as in combination to cover a full viewing angle of 360° of the nozzle 104.
The system 100 of
The system 100 further comprises a computer control system 109, which is configured to store or record the images obtained by the digital cameras 108a,b,c,d. The computer system 109 is further configured to compare images obtained by the same camera 108a at different points in times, and to detect differences between the compared images and issue a notification based on such detected differences. The computer system 109 may be configured to control image recording, image comparison, image difference detection and notification as described in connection with the flow charts of
In order for the cameras 108a,b,c,d to obtain images of a useful quality, there should be sufficient light on the objects for which the images are obtained. For the system 100 the objects are the spray nozzle(s) 104. Although not shown in
The cameras 108a,b,c,d may be mounted outside the chamber 101, and windows may be provided in the chamber 101 for allowing images to be obtained and for allowing light into the chamber 101. However, it is also within embodiments that openings are provided in the chamber 101 for receiving a camera 108, 201, 203, 206 and lamps 202, 204, 205, 207-212 connected to the camera 201, 203, 206. The camera and lamps may be mounted in a house, which may be provided with a window, where the camera housing may fit to an opening of the chamber 101. The camera house may be mounted on a flange, in order to be detachable connected to the chamber 101. Cleaning of the windows may be performed by use of Cleaning in Place, CIP.
A change in the performance of a spray nozzle will occur when deposit is build-up on the spray nozzle during operation, and a warning or notification should be issued when such a deposit build-up is detected.
Methods according to one or more embodiments of the present invention for monitoring deposit build-up 303 on spray nozzles 301b in a spray drying or spray cooling chamber 101 will be described in the following. The methods may be used when performing a spray drying or spray cooling process using the system 100 as described here above.
Step 402 is followed by a step 403 of process image recording. Here, one or more process images are repetitively obtained and recorded by use of the one or more digital optical cameras, where each process image covers the same image area as a corresponding reference image. If several cameras are used, then a reference image is obtained for each camera in step 402, and process images are also obtained for each camera in step 403, whereby process images obtained by a camera have a corresponding reference image. When a process image has been recorded in step 403, the next step is a step of image comparison, step 404, in which the obtained process image is compared to its reference image.
Step 404 is followed by a step of image difference detection, step 405. Step 405 is illustrated in further details if
If in step 503 the minimum number of error pixels is reached, then the process image is stored as an error image, step 506, and it is determined in step 507 if the number of stored error images has reached a predetermined minimum number of error images, and if not, the process returns to step 403, as indicated by arrows 508 and 406. At step 403 a next process image is recorded and subjected to the image comparison of step 404 followed by the difference detection of step 405. If a first error image has been recorded and stored, then there is a great possibility that the following obtained process images with the same reference image will be stored as error images, and when in step 507 it is determined that the number of error images has reached the predetermined minimum number of error images, a warning or notification is issued, step 408, as indicated by arrow 407 of
Although in step 404 it is preferred to compare all recorded process images with a reference image, it is also within embodiments of the methods of the present invention that only part of the obtained process images are compared with the reference image. Also for step 405 it is preferred that a number of error pixels is determined for each process image being compared to a reference image, but it is within embodiments of the methods of the present invention that the number or error pixels is only determined for a part of the process images being compared with the reference image.
When in step 408 a warning or notification is issued, action may be taken by the operators running the spray drying or spray cooling process. The action may be to have a look at the spray nozzles to see, if any of the nozzles need to be cleaned and if so, then the nozzles are cleaned. If there is no need for cleaning a nozzle, then the monitoring process may be initialized and a new reference image may be recorded before proceeding with the process.
The type of warning or notification is optional, but may be visual, audible or numerical in order to notify the operators. The type of notification may be an error notification.
The distance from a camera to different arranged spray nozzles will vary, and the number of image pixels taken up in an image may differ with the distance between the camera and the spray nozzle.
The method of the invention therefore covers embodiments, for which the image area of the recoded reference image comprises several areas of interest. Here, the detection in step 501 of a difference between similar addressed pixels in the reference image and a process image is performed for at least two areas of or for all defined areas of interest, and for step 502 a number of error pixels is determined for each of these at least two areas of interest for the process images being compared to the reference image. It is noticed that each of said at least two areas of interest has a corresponding predetermined minimum number of error pixels, where the minimum number of error pixels may differ between the defined image areas of interest. A warning or notification is issued when the determined number of error pixels reaches the corresponding predetermined minimum number of error pixels within at least one of the at least two areas of interest for each of the predetermined number of obtained process images, where the predetermined number of obtained images may be obtained at different points in time, and preferably obtained at subsequent points in time or at consecutive points in time.
Thus, the method of the invention also covers an embodiment, for which the image area of the recorded reference image, step 402, comprises both a first and a second area of interest, with the first area of interest covering parts of a first spray nozzle and the second area of interest covering parts of a second spray nozzle. Here, corresponding parts of the first nozzle and the second nozzle may be larger in the reference image for the first nozzle than for the second nozzle, and the corresponding predetermined minimum number of error pixels may be larger for the first area of interest covering the first nozzle than for the second area of interest covering the second nozzle.
It is within one or more embodiments of the invention that for an image area of interest covering at least a part of a spray nozzle, the predetermined minimum number of error pixels for this area of interest of the reference image is determined based on the number of pixels representing a characteristic measure of the nozzle, where the characteristic measure of the nozzle may be the width of the nozzle. It is preferred that for an area of interest covering part of a spray nozzle the predetermined minimum number of error pixels is larger than the number of pixels representing the characteristic measure of the nozzle. The predetermined minimum number of error pixels may differ depending on the number of pixels representing the characteristic measure of the nozzle, and may decrease with an increasing number of pixels representing the characteristic measure of the nozzle. The predetermined minimum number of error pixels may be at least 1.5 times larger, such as at least 2 times larger, such as at least 2.5 times larger, or such as at least 3 times larger than the number of pixels representing the characteristic measure of the nozzle.
The present invention covers embodiments, for which the detection of a pixel difference between similar addressed pixels in the reference image and a corresponding process image is at least partly based on a difference in pixel intensity. Thus, the predetermined minimum pixel difference may be defined as a minimum difference in pixel intensity between similar addressed pixels in the reference image and the process image. The predetermined minimum difference in pixel intensity may be set to at least 25%, such as at least 30%, such as at least 35%, or such as at least 39%.
The detection of pixel differences may be based on gray scale images, where the pixel intensity may be given as a binary bit number, such as a 7 or 8-bit number, and the minimum difference in pixel intensity may be set to a number representing a percentage of the maximum binary bit number corresponding to the predetermined minimum percentage difference in pixel intensity.
The present invention also covers embodiments, for which the detection of a pixel difference between similar addressed pixels in the reference image and the process image is at least partly based on a difference in pixel color. Here, the images may be RGB color images, and the predetermined minimum pixel difference may be defined as a minimum difference in pixel intensity of one of the RGB colors for similar addressed pixels in the reference image and the process image. The predetermined minimum difference in pixel intensity for one color may be set to at least 25%, such as at least 30%, such as at least 35%, or such as at least 39%.
It is noticed that a filtering or kind of noise filtering may be obtained when selecting the predetermined minimum number for the obtained error images and the predetermined minimum number of error pixels. A low value may be determined for the number of error images and a high value may be determined for the number of error pixels, or a high value may be determined for the number of error images and a low value may be determined for the number of error pixels. Also, relatively low values or relatively high values may be determined for both numbers.
The present invention also covers embodiments, for which the detection of a difference within an interest area between the reference image and the process image is at least partly based on spectral analysis. Here, a frequency spectrum may be determined for both the reference image and the process image, and the obtained frequency spectra may be compared. The frequency spectra may be obtained by use of Discrete Fourier Transform, DFT, for transforming the digital images from the spatial domain to the frequency domain. A predetermined change in the number of detected frequencies and/or a predetermined change in the power spectrum may be defined as representing a predetermined minimum difference for a process image to be an error image. A notification may be issued when the predetermined minimum difference is detected for each of a predetermined number of recorded process images. The minimum difference in the number of detected frequencies may be set to just one frequency.
It is within embodiments of the invention that the predetermined minimum number for the obtained error images has a value of just 1. But the number may be selected within—but not restricted to—the range from 1-50, or 1-100, such as 10, 20, 50 or 75.
The present invention covers embodiments, for which the predetermined minimum number of error pixels for an area of interest of the reference image is selected within the range of 1-1000 pixels, such as in the range of 50-500 pixels, or such as in the range of 50-200 pixels.
The one or more digital optical cameras used for obtaining the reference and process images may be digital still image cameras. The one or more cameras may also or alternatively be digital video cameras taking a stream of images, where all or only a part of the streamed images may recorded and compared with a reference image. The one or more cameras may be configured to obtain images comprising at least 2 MegaPixels, but the obtained images may comprise a smaller number of pixels, depending on the use of zooming in on the spray nozzles.
It is noted that when several digital optical cameras are used, then a reference image and corresponding process images should be obtained for each camera.
It is preferred that the minimum number of error pixels and the minimum number of error images are selected so that at least some deviations in the process image due to nozzle movement is accepted.
It is within one or more embodiments of the present invention that the method comprises a pre-filtering step being performed on detected pixel differences within an area of interest for an obtained process image before determining the number of error pixels. The pre-filtering may be a shape filtering whereby pixel differences due to error shapes which have no connection with deposit build-up may be filtered out in order not to generate false error images, where an error image is an image for which the detected number of error pixels has reached the predetermined minimum number of error pixels. The movement of a nozzle may result in pixel differences identified as a substantially rectangular shape in a process image when compared to the reference image, and these types of shapes may be accepted for the process images, and may therefore be filtered out in order not to generate error images. The size of the substantially rectangular shape may have a height close to the width of the nozzle width, and a width equal to the movement of the nozzle, which may be in the range of 10-30% of the nozzle width, such as about 20% of the nozzle width.
A light source may create a pixel difference between compared images, where the pixel difference can be identified as an elongated error shape due to a curved surface, and such error shapes may also be generalized as a substantially rectangular shape, and these types of shapes may be accepted for the process images, and may therefore be filtered out in the pre-filtering step in order not to generate error images. Shapes in pixel differences such as trapezoid or isosceles trapezoid may reflects the jet or spray and may also be accepted and filtered out in the pre-filtering step.
For simple system setups with a low number of spray nozzles, a reference image may be locked to a specific spray nozzle and set to follow movement of this nozzle, whereby errors may not be detected due to nozzle movement. Recognition of nozzle movement may be based on recognition of shape or contrast, and techniques such as morphology may be applied if needed.
For simple system setups with a low number of spray nozzles, e.g. where nozzles are not overlapping and contrast exists between nozzle and background, an auto detection of nozzles within the obtained images may be used based on recognition of nozzles as shapes or contrast objects. Auto detection of nozzles may be manually assisted, or solely from overlapping an image from e.g. CAD on which the nozzles are identified. Definition or identification of spray nozzles in each image area of interest can be obtained from user inputs or from automatic detection and may be logged for retrieval during error reporting. Individual nozzle identification may be assisted by marking of nozzles with color, bar code or other markings.
Although the present invention has been described with reference to specific features and embodiments thereof, it is evident that various modifications and embodiments can be made thereto without departing from the spirit and scope of the invention. The specification and drawings are, accordingly, to be regarded as an illustration of the invention as defined by the appended claims, and are contemplated to cover any and all modifications, variations, combinations or equivalents that fall within the scope of the present invention.
The term “comprising” as used in the appended claims does not exclude other elements or steps. The term “a” or “an” as used in the appended claims does not exclude a plurality.
Any method described herein and in the claims may be supplemented by any features of the apparatuses described herein and in the claims in terms of method features.
| Number | Date | Country | Kind |
|---|---|---|---|
| PA201500205 | Apr 2015 | DK | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2016/057080 | 3/31/2016 | WO | 00 |
