CN113695064B - Intelligent crushing method with condenser - Google Patents
Intelligent crushing method with condenser Download PDFInfo
- Publication number
- CN113695064B CN113695064B CN202111258679.0A CN202111258679A CN113695064B CN 113695064 B CN113695064 B CN 113695064B CN 202111258679 A CN202111258679 A CN 202111258679A CN 113695064 B CN113695064 B CN 113695064B
- Authority
- CN
- China
- Prior art keywords
- sequence
- sample
- condenser
- temperature
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B02—CRUSHING, PULVERISING, OR DISINTEGRATING; PREPARATORY TREATMENT OF GRAIN FOR MILLING
- B02C—CRUSHING, PULVERISING, OR DISINTEGRATING IN GENERAL; MILLING GRAIN
- B02C25/00—Control arrangements specially adapted for crushing or disintegrating
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Landscapes
- Engineering & Computer Science (AREA)
- Food Science & Technology (AREA)
- Control Of Electric Motors In General (AREA)
Abstract
The invention relates to an intelligent crushing method with a condenser, which comprises the following steps: recording the motor load and the condenser temperature of a prototype, and evaluating the crushing adhesion degree based on the images in the cabin of the crushing system; and calculating the reliability of adhesion and estimating a refrigeration interval. The problem that the coupling between the condenser and the crushing equipment is too low to adjust the target value of the condenser based on the feedback of the condenser and the crushing equipment is solved, the precision is improved, the stability is improved, the allowance is reduced, and the cost is reduced.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to an intelligent crushing method with a condenser.
Background
If the temperature of the condenser is too low in the crushing process of the object with the lower melting point, the refrigeration can be failed, so that the raw material is melted, and if the temperature allowance is not enough, the raw material is melted. Therefore, a method for automatically correcting the refrigeration target value based on the self working condition of the equipment is needed, and the reliability of the equipment is improved.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention adopts the following technical scheme:
an intelligent smashing method with a condenser comprises the following steps:
the method comprises the following steps: acquiring a motor load sequence, a condenser temperature sequence and an image change sequence of a prototype; calculating according to the image change sequence to obtain a texture response sequence and a granularity change sequence;
step two: calculating the reliability of adhesion according to the motor load sequence, the condenser temperature sequence, the texture response sequence and the granularity change sequence, and determining membership degrees of different levels according to the reliability;
step three: according to the membership degree, the equipment is automatically switched to the refrigeration parameter of the maximum confidence coefficient sample of a cluster corresponding to the maximum membership degree, and intelligent crushing control is carried out;
and (3) calculating the reliability of adhesion:
wherein A is a sample of the current work, B is a sample B in the history samples,
for the motor load sequence of sample a,
for the motor load sequence of sample B,
for the texture response sequence of sample a,
for the texture response sequence of sample B,
is a sequence of particle size changes for sample a,
is a sequence of particle size changes for sample B,
is the temperature of the ith sample in the temperature sequence of sample a,
for the temperature of the ith sample in the temperature sequence of sample B, the Pearson distance has a value in the range of [0,2]]；
Calculating and summing the credibility of all historical samples B based on the current sample A to obtain the current confidence C,
。
further, the first step is specifically as follows: when the motor reaches a target rotating speed, acquiring the load characteristics of the motor and the temperature of the condenser, and further acquiring a motor load change sequence and a condenser temperature sequence in the whole crushing process;
collecting at least two frames of images in the whole crushing process, performing convolution processing on each image to obtain transverse edge strength and longitudinal edge strength, selecting the maximum value of the transverse edge strength and the longitudinal edge strength as the edge strength of the image, and further obtaining an image change sequence corresponding to the image;
summing all the edge intensities of each image to obtain texture response and obtain texture response sequences of all the images;
and taking the difference sequence of the image change sequences of any two adjacent frames as the granularity change, and obtaining the granularity change sequences of all the images.
Further, the membership degree obtaining process is as follows:
1) establishing an inter-class distance function, and clustering by adopting a Kmeans mode and designating K as 5;
the inter-class distance function is:
wherein Trunc is a 0 truncation function, and the value less than 0 is made to be 0;
2) calculating confidence of each class's sub-cluster
Wherein N is the number of samples in the sub-cluster;
3) carrying out range standardization on the confidence coefficient to enable a value range to be located at [0,1 ];
4) and (3) performing the same treatment on the rest clusters, performing membership degree treatment based on the current working sample A, and calculating the membership degree Q:
，
where B is a cluster of samples.
Further, the third step is specifically: and obtaining the refrigeration parameters of the corresponding condenser according to the cluster corresponding to the maximum membership degree.
The invention has the beneficial effects that:
the problem that the coupling between the condenser and the crushing equipment is too low to adjust the target value of the condenser based on the feedback of the condenser and the crushing equipment is solved, the precision is improved, the stability is improved, the allowance is reduced, and the cost is reduced.
Drawings
FIG. 1 is a schematic of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and examples.
As shown in fig. 1, the pulverizing container is provided with a condenser, wherein the condenser is used in the pulverizing system to cool the air medium of the pulverizing container, take away the heat in the pulverizing process, and cool the container at the same time, so as to ensure that the temperature of the pulverized material is in a lower temperature environment.
Based on the device, the intelligent crushing method with the condenser provided by the invention comprises the following steps:
the method comprises the following steps: acquiring a motor load sequence, a condenser temperature sequence and an image change sequence of a prototype; calculating according to the image change sequence to obtain a texture response sequence and a granularity change sequence;
firstly, the time of one starting process is specified, T =3s in the embodiment, and when the temperature of the heat exchanger is higher than 2 ℃, because the heat exchange cannot be caused in time due to rapid crushing, the blocking is easy to occur.
In the invention, the heat exchange mechanism and the air duct structure of the container are not in the protection range, and the embodiment only provides a simple and understandable structure for the convenience of an implementer, and does not represent a better structure.
The invention features a data processing method and a control method for images of the interior of a condenser and a grinding container.
In the embodiment, the motor load and the surface temperature of the condenser are recorded when the motor reaches the target rotating speed in the starting process of the crushing equipment. Most motor control systems currently operate on the basis of PID to control the speed, and the practitioner can determine the time to reach the target speed during start-up from the speed reading of the governor, which in the present embodiment is assumed to be 3 seconds, as previously mentioned.
The PID controller of the motor can read out the current driver duty cycle and therefore record the motor load based on the duty cycle of the motor, in this embodiment using a read frequency of 50 Hz. Then during start-up (three seconds) 150 PWM samples W are obtained. The remaining samples S after start-up (after three seconds) are then recorded, the number of samples being an indefinite sequence.
In addition to recording the PWM samples of the motor, the operating temperature of the condenser needs to be recorded. The condenser temperature is a slowly varying value, and in this example 1Hz is used to record the temperature from the start of the click to the end of the whole milling process, resulting in a temperature T. The working temperature of the condenser is measured by contacting the fins with a temperature probe, and the method is a common temperature measurement method.
Therefore, a starting load sequence W of the motor and a condenser temperature change value T corresponding to the load S in the crushing process can be obtained in the primary crushing process.
The process of obtaining the texture response sequence and the granularity change sequence in this embodiment is as follows:
1) collecting at least two frames of images in the whole crushing process: in particular, the acquisition of the image is performed by a camera, which is placed vertically downwards, resulting in a top view. In this embodiment, the image information in the container is collected at a shutter speed of 1/125, and the operator should adjust the shutter speed according to the actual illumination and camera characteristics, and the shutter speed should not be too fast or too slow, preferably between 1/250 and 1/50, in order to observe the texture information during the shredding process. In this embodiment, the frame rate is 5 FPS.
The camera required by the invention has lower cost, the conventional consumption-level camera can complete the acquisition task, and an image set X can be obtained in each crushing process.
2) Preprocessing each frame of image based on a Sobel operator: in the predetermined image coordinate system, X is an abscissa and Y is an ordinate. Carrying out convolution processing of Sobel operator on the image, extracting transverse edge intensity and obtaining a transverse edge intensity image
And longitudinal edge intensity image
. Will eventually be
The final image edge intensity is obtained by the following processing:
i.e. are pairs
Such that G corresponds to pixels that are their maximum, G being a single channel.
3) Summing G to obtain texture response
The response size in the G image can be represented, and then texture response sequences of all the images are obtained; meanwhile, the difference sequence of the image change sequences of any two adjacent frames is used as the granularity change, and the granularity change sequences of all the images are obtained.
The invention is characterized in that the crushed material gradually loses edges along with the reduction of the particle size, so the sequence is processed as follows: because the texture is random, the texture change rule in the crushing process cannot be intuitively and stably described by using a mode based on deep learning, frequency domain analysis and the like, and because the image sequence in the crushing process has a process from roughness to fineness, the following rule exists: when the fragmentation is sufficiently fine, the texture between the images of the previous and subsequent frames is similar, so the difference between the previous and subsequent frames is used to describe the fragmentation process. When the raw material is rough, the difference between the edges of the previous and next frames is large, i.e., the frame difference is large. The frame difference is small along with the gradual fineness of the raw materials, and the frame difference is large if the raw materials are melted and agglomerated again due to the fault of the condenser in the crushing process. The following process is therefore performed for the edge between two frames to describe the texture change of the fragmentation process.
The granularity change is as follows:
then there is
And n is n frames collected in the crushing process.
Thus, a texture response sequence of the texture feature description information in the whole crushing process is obtained
And the sequence of particle size variations
。
Step two: calculating the reliability of adhesion according to the motor load sequence, the condenser temperature sequence, the texture response sequence and the granularity change sequence, and determining membership degrees of different levels according to the reliability;
the credibility of adhesion is calculated based on sample characteristics, broken materials gradually lose edges along with the reduction of particle sizes, but normally crushed materials cannot be melted, and the melted materials are agglomerated, so that edge response is not a normal decreasing relation, the credibility considers the problem that frosting occurs when the temperature of a condenser is too low, so that the heat exchange effect is reduced, and the crushing environment temperature is increased on the contrary
And calculating the adhesion reliability of each sample, wherein the core idea is that most normal samples and few abnormal samples are used, so that whether the characteristic sequence in the current working process meets the historical condition or not can be calculated under a large number of samples.
Specifically, the method comprises the following steps: for the sample set of the device, a cohesive confidence function is constructed:
wherein A is a sample of the current work, B is a sample B in the history samples,
for the motor load sequence of sample a,
for the motor load sequence of sample B,
for the texture response sequence of sample a,
for the texture response sequence of sample B,
is a sequence of particle size changes for sample a,
is a sequence of particle size changes for sample B,
is the temperature of the ith sample in the temperature sequence of sample a,
is the temperature of the ith sample in the temperature series of sample B.
Calculating and summing the credibility of all historical samples B based on the current sample A to obtain the current confidence C,
。
it should be noted that the numerator in the formula of the confidence function represents the surface temperature of the condenser and the euclidean distance of one sample, if the temperature difference between the current sample and the target sample is too large, the confidence is reduced, the denominator represents the pearson coefficients of the constrained texture response pearson coefficients and the particle size variation sequence, and converts them into the pearson distance, wherein the two pearson coefficients have the most relevant value range of [ -1,1], which means the same working state as a certain sample, whereas-1 means the complete negative correlation, which cannot be negative in the normal natural law, if the condenser cannot cool down the raw material due to a fault, and the agglomeration occurs even more serious result, the two pearson coefficients may be negative, and thus, the conversion is represented by the pearson distance, the pearson distance has the value range of [0,2], and the distance is lower, when the motor samples are similar, the molecular value is close to 1, otherwise, the value is smaller when abnormality occurs, and may be a negative value, when confidence is accumulated, if abnormal conditions such as raw material melting occur in the crushing process, the confidence is remarkably reduced because the later motor load is not linear load.
The membership degree obtaining process comprises the following steps:
1) establishing an inter-class distance function, and clustering according to the inter-class distance;
wherein the function of the distance between classes is
Wherein, Trunc is a 0 truncation function, the value less than 0 is 0, the implementer uses a Kmeans method to designate K as a value based on the distance function, the K recommended by the invention is 5-10, and the implementer should correspond the K value to the refrigeration gear or PID target value of the corresponding condenser, for example, 5 gear or PID target value is [ -2,1,0,1,2 ].
For convenience of understanding, K =5 is selected, and the purpose here is to cluster samples of different textures based on similar temperatures and motor loads, evaluate confidence degrees of the samples, obtain representative image sequence characteristics of the working condition (similar motor load and condenser temperature change), and use the representative image sequence characteristics to judge the working condition of the current equipment to obtain a comprehensive condenser temperature value.
2) Calculating the confidence of each sub-cluster of each category and obtaining the confidence of the sub-cluster
Wherein N is within a sub-clusterThe number of samples;
3) carrying out range standardization on the confidence coefficient to enable a value range to be located at [0,1 ];
4) and performing the same treatment on the rest clusters, performing membership treatment based on the current working sample A, calculating the membership Q, and calculating a refrigeration target value (or corresponding gear) to be adjusted.
Wherein B is a cluster of samples, so far, Q corresponding to A in each cluster is obtained, wherein Q represents the characteristic of the observed texture change of the crushed raw material.
As another example, in calculating the degree of membership, the practitioner typically sets an empirical threshold for C, which is determined by the number of samples and the stability of the plant, typically 80% of C under normal conditions.
Step three: and according to the membership degree, the equipment is automatically switched to the refrigeration parameter of the maximum confidence coefficient sample of the cluster corresponding to the maximum membership degree, and intelligent crushing control is carried out.
An implementer uses the cluster corresponding to the maximum Q and obtains a corresponding target working value of the condenser, so that the invention determines samples of different types of crushing processes based on the motor load of the equipment and the temperature change characteristics of the condenser, and evaluates normal samples and abnormal samples based on confidence, thereby avoiding the influence of abnormal samples in the subsequent process of determining the target value and improving the weight of the normal samples. And finally, a target value to be corrected is obtained, so that the condition that the normal crushing work is influenced because the condenser is frosted due to too much refrigeration allowance (too low temperature) of the equipment is avoided, and the stable work of the equipment is ensured.
The above embodiments are merely illustrative of the present invention, and should not be construed as limiting the scope of the present invention, and all designs identical or similar to the present invention are within the scope of the present invention.
Claims (4)
1. An intelligent crushing method with a condenser is characterized by comprising the following steps:
the method comprises the following steps: acquiring a motor load sequence, a condenser temperature sequence and an image change sequence of a prototype; calculating according to the image change sequence to obtain a texture response sequence and a granularity change sequence;
step two: calculating the reliability of adhesion according to the motor load sequence, the condenser temperature sequence, the texture response sequence and the granularity change sequence, and determining membership degrees of different levels according to the reliability;
step three: according to the membership degree, the equipment is automatically switched to the refrigeration parameter of the maximum confidence coefficient sample of a cluster corresponding to the maximum membership degree, and intelligent crushing control is carried out;
and (3) calculating the reliability of adhesion:
wherein A is a sample of the current work, B is a sample B in the history samples,
for the motor load sequence of sample a,
for the motor load sequence of sample B,
for the texture response sequence of sample a,
for the texture response sequence of sample B,
is a sequence of particle size changes for sample a,
is a sequence of particle size changes for sample B,
is the temperature of the ith sample in the temperature sequence of sample a,
for the temperature of the ith sample in the temperature sequence of sample B, the Pearson distance has a value in the range of [0,2]]；
Calculating and summing the credibility of all historical samples B based on the current sample A to obtain the current confidence C,
。
2. the intelligent crushing method with the condenser as claimed in claim 1, wherein the first step is specifically: when the motor reaches a target rotating speed, acquiring the load characteristics of the motor and the temperature of the condenser, and further acquiring a motor load change sequence and a condenser temperature sequence in the whole crushing process;
collecting at least two frames of images in the whole crushing process, performing convolution processing on each image to obtain transverse edge strength and longitudinal edge strength, selecting the maximum value of the transverse edge strength and the longitudinal edge strength as the edge strength of the image, and further obtaining an image change sequence corresponding to the image;
summing all the edge intensities of each image to obtain texture response and obtain texture response sequences of all the images;
and taking the difference sequence of the image change sequences of any two adjacent frames as the granularity change, and obtaining the granularity change sequences of all the images.
3. The intelligent crushing method with the condenser according to claim 1,
the membership degree obtaining process comprises the following steps:
1) establishing an inter-class distance function, and clustering by adopting a Kmeans mode and designating K as 5;
the inter-class distance function is:
wherein Trunc is a 0 truncation function, and the value less than 0 is made to be 0;
2) calculating confidence of each class's sub-cluster
Wherein N is the number of samples in the sub-cluster;
3) carrying out range standardization on the confidence coefficient to enable a value range to be located at [0,1 ];
4) and (3) performing the same treatment on the rest clusters, performing membership degree treatment based on the current working sample A, and calculating the membership degree Q:
where B is a cluster of samples.
4. The intelligent crushing method with the condenser as claimed in claim 3, wherein the third step is specifically: and obtaining the refrigeration parameters of the corresponding condenser according to the cluster corresponding to the maximum membership degree.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111258679.0A CN113695064B (en) | 2021-10-28 | 2021-10-28 | Intelligent crushing method with condenser |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111258679.0A CN113695064B (en) | 2021-10-28 | 2021-10-28 | Intelligent crushing method with condenser |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113695064A CN113695064A (en) | 2021-11-26 |
| CN113695064B true CN113695064B (en) | 2021-12-28 |
Family
ID=78647176
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111258679.0A Active CN113695064B (en) | 2021-10-28 | 2021-10-28 | Intelligent crushing method with condenser |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113695064B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101835158A (en) * | 2010-04-12 | 2010-09-15 | 北京航空航天大学 | Sensor network trust evaluation method based on node behaviors and D-S evidence theory |
| CN105404892A (en) * | 2015-10-23 | 2016-03-16 | 浙江工业大学 | Ordered fuzzy C mean value cluster method used for sequence data segmentation |
| CN106408513A (en) * | 2016-08-25 | 2017-02-15 | 天津大学 | Super-resolution reconstruction method of depth map |
| CN207076510U (en) * | 2017-08-03 | 2018-03-09 | 雅安远创陶瓷有限责任公司 | A kind of ceramic raw material grinder |
| JP2018153755A (en) * | 2017-03-17 | 2018-10-04 | 日本コークス工業株式会社 | Crushing system |
| CN109013032A (en) * | 2017-10-27 | 2018-12-18 | 江西理工大学 | A kind of method of source signal fusion forecasting ball mill filling rate, material ball ratio |
| CN111896430A (en) * | 2020-01-06 | 2020-11-06 | 上海恺擎智能科技有限公司 | Pollen monitoring method and device |
-
2021
- 2021-10-28 CN CN202111258679.0A patent/CN113695064B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101835158A (en) * | 2010-04-12 | 2010-09-15 | 北京航空航天大学 | Sensor network trust evaluation method based on node behaviors and D-S evidence theory |
| CN105404892A (en) * | 2015-10-23 | 2016-03-16 | 浙江工业大学 | Ordered fuzzy C mean value cluster method used for sequence data segmentation |
| CN106408513A (en) * | 2016-08-25 | 2017-02-15 | 天津大学 | Super-resolution reconstruction method of depth map |
| JP2018153755A (en) * | 2017-03-17 | 2018-10-04 | 日本コークス工業株式会社 | Crushing system |
| CN207076510U (en) * | 2017-08-03 | 2018-03-09 | 雅安远创陶瓷有限责任公司 | A kind of ceramic raw material grinder |
| CN109013032A (en) * | 2017-10-27 | 2018-12-18 | 江西理工大学 | A kind of method of source signal fusion forecasting ball mill filling rate, material ball ratio |
| CN111896430A (en) * | 2020-01-06 | 2020-11-06 | 上海恺擎智能科技有限公司 | Pollen monitoring method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113695064A (en) | 2021-11-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7536276B2 (en) | Method and apparatus for equipment health monitoring | |
| DE602006000840T2 (en) | Cooling device and control method for it | |
| US7876229B2 (en) | Flare monitoring | |
| CN116435634B (en) | Storage battery temperature state monitoring and management system | |
| EP2995875B1 (en) | Method for operating a clean-room and control device | |
| WO2006026267A2 (en) | Fault diagnostics and prognostics based on distance fault classifiers | |
| CN113609901A (en) | Power transmission and transformation equipment fault monitoring method and system | |
| CN116520915A (en) | Network center machine room temperature early warning control system based on thermal infrared image | |
| CN113695064B (en) | Intelligent crushing method with condenser | |
| JP2023502716A (en) | Heat treatment adjustment control method based on information collected in situ and its application | |
| EP1441274A2 (en) | Temperature control with infrared sensing | |
| CN107120794A (en) | Air conditioner operating condition adjusting method and air conditioner | |
| CN116300650A (en) | Monitoring system for preparing micro wax powder based on data analysis | |
| CN112748755B (en) | Temperature control device based on visual identification | |
| CN117290802A (en) | Host power supply operation monitoring method based on data processing | |
| CN110513823A (en) | A kind of temprature control method of intelligent power module, device and air conditioner | |
| CN110107986B (en) | Defrosting control method, device and system of temperature adjusting equipment and air conditioner | |
| DE102019216698A1 (en) | Method and device for cooling a motor vehicle battery | |
| DE102014212833A1 (en) | Method and device for cooling a battery | |
| CN110425711A (en) | A kind of temprature control method of intelligent power module, device and air conditioner | |
| CN115830106B (en) | Auxiliary positioning method for electrified cleaning of equipment in machine room | |
| DE102013208795A1 (en) | Cooling system with a cooling device for temperature control of a battery and method for controlling the temperature of a battery | |
| EP4148518A1 (en) | Kpca-based non-linear chiller fault detection method | |
| JPWO2019187155A1 (en) | Information processing equipment, information processing methods and programs | |
| CN110940514A (en) | Method for constructing and diagnosing early diagnosis model of fan gear box heat dissipation fault |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |