To be successful in a obtaining a uniform grind that is necessary to achieve a high percentage of recovery it is necessary to control the degree of fineness that the ore is reduced to. This is done by separating the fine material from the course and regrinding the coarse until it is fine enough for efficient mineral extraction.
To be able to obtain the necessary control over the amount of grinding required, a method of effective classification and separation by size must be available. For maximum effectiveness it should take place after every stage of grinding.
The types of equipment that are used to accomplish this are called CLASSIFIERS. There are three basic kinds used. The first two, the RAKE classifier and the SPIRAL or screw classifier work on the same principal, and are not often used any more. These two types were popular for many years. It wasnt until the development of the CYCLONE type classifier that their popularity faded. You may still find a few though, in the older mills and the mills that require a classification of the larger ore sizes that the cyclones are not very good at sizing.
Both the rake and spiral classifiers take advantage of the natural settling characteristics of ore. Any time that slurry is allowed to flow over a surface the tendency of the ore is to graduate itself into layers of different sized material. The larger sizes will be on the bottom, these are also the ones that are the slowest moving. As you come closer to the surface, the material will become smaller and faster until the very finest and the easiest to wash away is on top.
To understand how these two classifiers make use of this settling action, a description of them is required. First, to have the classification happen, the slurry must be able to flow. This means the classifiers must be inclined. The working portion of these two classifiers are the RAKES or SPIRAL/screw which are placed into the flow of ore. To separate the course material from the fine, the rakes and spiral make use of the same theory, but differ in its application. The theory is, as the slurry flows down the inclined bed of the classifier it will separate into different sizing. The larger ore that is on the bottom will not be flowing as fast as the light ore on top.
To separate the two, the rakes and the spiral will pull the all of the slurry back up the incline, then, let it go to flow back down towards the underflow or in this case the fine ore discharge point. The smaller, faster ore will be able to travel a longer distance than the large particles before the rakes or spiral will pull the ground material back towards the coarse ore discharge. If the Classifier is able to pull the course ore backwards further than it can travel forwards, then eventually the bigger particles will be pulled all the way to the top of the incline where they will be discharged. The smaller faster pieces of ground rock will end up at the bottom of the incline to be discharged as fine material that is ready for the next stage of processing.
This type of classifier will do away with the necessity of pumps. The length of the incline that is needed is long and steep enough to have the material lifted to the feed end of the mill. The flow of the finer ore will run down hill to the next piece of equipment.
The concentrator that used this type of classification was built on the side of a hill to make use of gravity to get the material from one stage of production to the next. It was because of this that this type of concentrator was referred to as a GRAVITY FLOW/MILL. I used the past tense in this paragraph because this design of mill is no longer in use.
I want to know what is the range of the % Solids content in overflow from screw/spiral classifier in Hematite Iron ore washing for efficient operation of classifier. I also want to know what is Auto dilution in thickener. Does Auto dilution has any effect on Pumping capacity of clarified water from thickener.
Each operation is different, but the good news is that you can simply determine the solids % wt. in the SOF, try 2 -3 times daily over one week, get a profile. If below 5% wt., you may not need auto dilution, also depending on ore and grind size. This is almost a clarifier regime, often workable without rakes in certain units. Above 5% in general start looking at auto-dilution before adding the flocculant do this off-line. Make sure that the thickener underflow (TUF) discharge comes out continuously, otherwise you may need to play with the lifters, if you have them. The UF solids % wt. must be correlated with the yield stress.look for 30 Pa high-end cut off value for lean operation. When the ore or grind size changes, you need to repeat the evaluation.
My take on these classifiers is that the clear water added to the classifier feed determines the size of the largest heaviest particle going to the overflow. This is the criteria that you should be working to achieve. As you probably have more than one classifier reporting to the thickener you will need to perform the solids percent in each overflow. As you add water to the classifier feed, the separation efficiency increases. You should be raising or lowering the discharge weir to attain the desired size cut. Only if the thickener becomes overloaded should you add water. Pumping excess water adds cost and wear to pump trains.
Pulp density of the Overflow defines % of solids in thickener. This solids% depends also up on quality of recycle water used in spiral processto know quantum of solids..you have to give feed quantity and underflow quantity.
The % solids in classifier overflow may vary in wide range, its all depends on your ore characterisation and operating variables. For the same operation we used to get 15-18% solids as my ore contains too much fines & this is not end process in our case.Try to concentrate on end products. Do not let go valuables in your final tailings.
get started with trainable classifiers - microsoft 365 compliance | microsoft docs
A Microsoft 365 trainable classifier is a tool you can train to recognize various types of content by giving it samples to look at. Once trained, you can use it to identify item for application of Office sensitivity labels, Communications compliance policies, and retention label policies.
Creating a custom trainable classifier first involves giving it samples that are human picked and positively match the category. Then, after it has processed those, you test the classifiers ability to predict by giving it a mix of positive and negative samples. This article shows you how to create and train a custom classifier and how to improve the performance of custom trainable classifiers and pre-trained classifiers over their lifetime through retraining.
Opt-in is required the first time for trainable classifiers. It takes twelve days for Microsoft 365 to complete a baseline evaluation of your organizations content. Contact your global administrator to kick off the opt-in process.
When you want a trainable classifier to independently and accurately identify an item as being in particular category of content, you first have to present it with many samples of the type of content that are in the category. This feeding of samples to the trainable classifier is known as seeding. Seed content is selected by a human and is judged to represent the category of content.
You need to have at least 50 positive samples and as many as 500. The trainable classifier will process up to the 500 most recent created samples (by file created date/time stamp). The more samples you provide, the more accurate the predictions the classifier will make.
Once the trainable classifier has processed enough positive samples to build a prediction model, you need to test the predictions it makes to see if the classifier can correctly distinguish between items that match the category and items that don't. You do this by selecting another, hopefully larger, set of human picked content that consists of samples that should fall into the category and samples that won't. You should test with different data than the initial seed data you first provided. Once it processes those, you manually go through the results and verify whether each prediction is correct, incorrect, or you aren't sure. The trainable classifier uses this feedback to improve its prediction model.
Collect between 50-500 seed content items. These must be only samples that strongly represent the type of content you want the trainable classifier to positively identify as being in the classification category. See, Default crawled file name extensions and parsed file types in SharePoint Server for the supported file types.
Make sure the items in your seed set are strong examples of the category. The trainable classifier initially builds its model based on what you seed it with. The classifier assumes all seed samples are strong positives and has no way of knowing if a sample is a weak or negative match to the category.
Within 24 hours the trainable classifier will process the seed data and build a prediction model. The classifier status is In progress while it processes the seed data. When the classifier is finished processing the seed data, the status changes to Need test items.
Collect at least 200 test content items (10,000 max) for best results. These should be a mix of items that are strong positives, strong negatives and some that are a little less obvious in their nature. See, Default crawled file name extensions and parsed file types in SharePoint Server for the supported file types.
When the trainable classifier is done processing your test files, the status on the details page will change to Ready to review. If you need to increase the test sample size, choose Add items to test and allow the trainable classifier to process the additional items.
Microsoft 365 will present 30 items at a time. Review them and in the We predict this item is "Relevant". Do you agree? box choose either Yes or No or Not sure, skip to next item. Model accuracy is automatically updated after every 30 items.
python: loaded nltk classifier not working - stack overflow
I'm trying to train a NLTK classifier for sentiment analysis and then save the classifier using pickle.
The freshly trained classifier works fine. However, if I load a saved classifier the classifier will either output 'positive', or 'negative' for ALL examples.
The NaiveBayesClassifier expects feature vectors for both training and classification; your code looks as if you passed the raw words to the classifier instead (but presumably only after unpickling, otherwise you wouldn't get different behavior before and after unpickling). You should store the feature extraction code in a separate file, and import it in both the training and the classifying (or testing) script.
I doubt this applies to the OP, but some NLTK classifiers take the feature extraction function as an argument to the constructor. When you have separate scripts for training and classifying, it can be tricky to ensure that the unpickled classifier successfully finds the same function. This is because of the way pickle works: pickling only saves data, not code. To get it to work, just put the extraction function in a separate file (module) that your scripts import. If you put in in the "main" script, pickle.load will look for it in the wrong place.
spiral classifier for mineral processing
In Mineral Processing, the SPIRAL Classifier on the other hand is rotated through the ore. It doesnt lift out of the slurry but is revolved through it. The direction of rotation causes the slurry to be pulled up the inclined bed of the classifier in much the same manner as the rakes do. As it is revolved in the slurry the spiral is constantly moving the coarse backwards the fine material will flow over the top and be travelling fast enough to be able to work its way downwards to escape.
The Variables of these two types of classifiers are The ANGLE of the inclined bed, this is normally a fixed angle the operator will not be able to adjust it.
The SPEED of the rakes or spirals, the DENSITY of the slurry, the TONNAGE throughput and finally the SETTLING RATE of the ore itself.To be effective all of these variables must be balanced. If the incline is too steep the flow of slurry will be too fast for the rakes or spirals to separate the ore. If the angle is too flat the settling rate will be too high and the classifier will over load. The discharge rate will be lower than the feed rate, in this case. The load on the rakes will continue to build until the weight is greater than the rake or spiral mechanism is able to move. This will cause the classifier to stop and is known as being SANDED UP. If the speed of the rakes or spirals are too fast, too much will be pulled, out the top. This will increase the feed to the mill and result in an overload in either the mill or classifier as the circuit tries to process the increased CIRCULATING LOAD.
The DENSITY of the slurry is very important, too high the settling will be hampered by too many solids. Each particle will support each other preventing the heavier material from quickly reaching the bottom of the slurry. This will not allow a separation to take place quickly. The speed at which the slurry will be travelling will be slow and that will hamper effective classification. Another variable is the TONNAGE. All equipment has a limit on the throughput that anyone is able to process, classifiers are no different. This and the other factors will have to be adjusted to compensate for the last variable, the ore itself. Every ore type has a different rate of settling. To be effective each of the previous variables will have to be adjusted to conform to each ones settling characteristics.
The design of these classifiers (rake, spiral, screw) have inherent problems, First, they are very susceptible to wear, caused by the scrubbing action of the ore, that plus all of the mechanical moving parts create many worn areas to contend with. The other problem that these classifiers have is that they are easily overloaded. An overloaded classifier can quickly deteriorate into a sanded-up classifier. Once that happens the results are lost operating time, spillage and a period of poor Mineral Processing and Separation performance.
Another mechanical classifier is the spiral classifier. The spiral classifier such as the Akins classifier consists of a semi-cylindrical trough (a trough that is semicircular in cross-section) inclined to the horizontal. The trough is provided with a slow-rotating spiral conveyor and a liquid overflow at the lower end. The spiral conveyor moves the solids which settle to the bottom upward toward the top of the trough.
The slurry is fed continuously near the middle of the trough. The slurry feed rate is so adjusted that fines do not have time to settle and are carried out with the overflow .liquid. Heavy particles have time to settle, they settle to the bottom of the trough and the spiral conveyor moves the settled solids upward along the floor of the trough toward the top of the trough/the sand product discharge chute.
machine learning classifiers - the algorithms & how they work
A classifier in machine learning is an algorithm that automatically orders or categorizes data into one or more of a set of classes. One of the most common examples is an email classifier that scans emails to filter them by class label: Spam or Not Spam.
A classifier is the algorithm itself the rules used by machines to classify data. A classification model, on the other hand, is the end result of your classifiers machine learning. The model is trained using the classifier, so that the model, ultimately, classifies your data.
There are both supervised and unsupervised classifiers. Unsupervised machine learning classifiers are fed only unlabeled datasets, which they classify according to pattern recognition or structures and anomalies in the data. Supervised and semi-supervised classifiers are fed training datasets, from which they learn to classify data according to predetermined categories.
Sentiment analysis is an example of supervised machine learning where classifiers are trained to analyze text for opinion polarity and output the text into the class: Positive, Neutral, or Negative. Try out this pre-trained sentiment analysis model to see how it works.
Machine learning classifiers are used to automatically analyze customer comments (like the above) from social media, emails, online reviews, etc., to find out what customers are saying about your brand.
Other text analysis techniques, like topic classification, can automatically sort through customer service tickets or NPS surveys, categorize them by topic (Pricing, Features, Support, etc.), and route them to the correct department or employee.
SaaS text analysis platforms, like MonkeyLearn, give easy access to powerful classification algorithms, allowing you to custom-build classification models to your needs and criteria, usually in just a few steps.
Machine learning classifiers go beyond simple data mapping, allowing users to constantly update models with new learning data and tailor them to changing needs. Self-driving cars, for example, use classification algorithms to input image data to a category; whether its a stop sign, a pedestrian, or another car, constantly learning and improving over time.
A decision tree is a supervised machine learning classification algorithm used to build models like the structure of a tree. It classifies data into finer and finer categories: from tree trunk, to branches, to leaves. It uses the if-then rule of mathematics to create sub-categories that fit into broader categories and allows for precise, organic categorization.
Naive Bayes is a family of probabilistic algorithms that calculate the possibility that any given data point may fall into one or more of a group of categories (or not). In text analysis, Naive Bayes is used to categorize customer comments, news articles, emails, etc., into subjects, topics, or tags to organize them according to predetermined criteria, like this:
K-nearest neighbors (k-NN) is a pattern recognition algorithm that stores and learns from training data points by calculating how they correspond to other data in n-dimensional space. K-NN aims to find the k closest related data points in future, unseen data.
In text analysis, k-NN would place a given word or phrase within a predetermined category by calculating its nearest neighbor: k is decided by a plurality vote of its neighbors. If k = 1, it would be tagged into the class nearest 1.
Take a look at this visual representation to understand how SVM algorithms work. We have two tags: red and blue, with two data features: X and Y, and we train our classifier to output an X/Y coordinate as either red or blue.
The SVM assigns a hyperplane that best separates (distinguishes between) the tags. In two dimensions this is simply a straight line. Blue tags fall on one side of the hyperplane and red on the other. In sentiment analysis these tags would be Positive and Negative.
SVM algorithms make excellent classifiers because, the more complex the data, the more accurate the prediction will be. Imagine the above as a 3-dimensional output, with a Z-axis added, so it becomes a circle.
Artificial neural networks are designed to work much like the human brain does. They connect problem-solving processes in a chain of events, so that once one algorithm or process has solved a problem, the next algorithm (or link in the chain) is activated.
Artificial neural networks or deep learning models require vast amounts of training data because their processes are highly advanced, but once they have been properly trained, they can perform beyond other, individual, algorithms.
There are a variety of artificial neural networks, including convolutional, recurrent, feed-forward, etc., and the machine learning architecture best suited to your needs depends on the problem youre aiming to solve.
Classification algorithms enable the automation of machine learning tasks that were unthinkable just a few years ago. And, better yet, they allow you to train AI models to the needs, language, and criteria of your business, performing much faster and with a greater level of accuracy than humans ever could.
MonkeyLearn is a machine learning text analysis platform that harnesses the power of machine learning classifiers with an exceedingly user-friendly interface, so you can streamline processes and get the most out of your text data for valuable insights.
spiral: self-tuning services via real-time machine learning - facebook engineering
To the billions of people using Facebook, our services may look like a single, unified mobile app or website. From inside the company, the view is different. Facebook is built using thousands of services, with functions ranging frombalancing internet traffictotranscoding imagestoproviding reliable storage. The efficiency of Facebook as a whole is the sum of the efficiencies of its individual services, and each service is typically optimized in its own way, with approaches that may be difficult to generalize or adapt in the face of fast-paced changes. To more effectively optimize our many services, with the flexibility to adapt to a constantly changing interconnected web of internal services, we have developed Spiral. Spiral is a system for self-tuning high-performance infrastructure services at Facebook scale, using techniques that leverage real-time machine learning. By replacing hand-tuned heuristics with Spiral, we can optimize updated services in minutes rather than in weeks.
At Facebook, the pace of change is rapid. The Facebook codebase is pushed to production every few hours for example,new versions of the front end as part of our continuous deployment process. In this dynamic world, trying to manually fine-tune services to maintain peak efficiency is impractical. It is simply too difficult to rewrite caching/admission/eviction policies and other manually tuned heuristics by hand. We have to fundamentally change how we think about software maintenance.
To efficiently address this challenge, the system needed to become self-tuning rather than rely on manually hard-coded heuristics and parameters. This shift prompted Facebook engineers to approach work in a new way: Instead of looking at charts and logs produced by the system to verify correct and efficient operation, engineers now express what it means for a system to operate correctly and efficiently in code. Today, rather than specify how to compute correct responses to requests, our engineers encode the means of providing feedback to a self-tuning system.
A traditional caching policy may look like a tree with branches that take into account an objects size, type, and other metadata to decide whether to cache it. A self-tuning cache would be implemented differently. Such a system could examine the access history for an item: If this item had never been accessed, it was probably a bad idea to cache it. In the language of machine learning, a system that uses metadata (features) and associated feedback (labels) to differentiate items would be a classifier. This classifier would be used to make decisions about items entering the cache and the system would be retrained continuously. This continuous retraining is what allows the system to stay current even as the environment changes.
Conceptually, this approach is similar todeclarative programming. SQL is a popular example of this approach: Instead of specifying how to compute the result of a complex query, an engineer just needs to specify what needs to be computed and then the engine figures out the optimal query and executes it.
The challenge of using declarative programming for systems is making sure objectives are specified correctly and completely. As with the self-tuning image cache policy above, if the feedback for what should and should not be cached is inaccurate or incomplete, the system will quickly learn to provide incorrect caching decisions, which will degrade performance. (This paper by several Google engineersgoes into detail about this problem and others related to using closed-loop machine learning.) In our experience, precisely defining the desired outcome for self-tuning is one of the hardest parts of onboarding with Spiral. However, we also found that engineers tended to converge on clear and correct definitions after a few iterations.
To enable system engineers at Facebook to keep up with the ever-increasing pace of change, engineers in ourFacebook Bostonoffice built Spiral, a small, embedded C++ library with very few dependencies. Spiral uses machine learning to create data-driven and reactive heuristics for resource-constrained real-time services. The system allows for much faster development and hands-free maintenance of those services, compared with the hand-coded alternative.
Integration with Spiral consists of adding just two call sites to your code: one for prediction and one for feedback. The prediction call site is the output of the smart heuristic used to make decisions, such as Should this item be admitted into the cache? The prediction call is implemented as a fast local computation and is meant to be executed on every decision.
The library can operate in a fully embedded mode, or it can send feedback and statistics to the Spiral back-end service, which can visualize useful information for debugging, log data to long-term storage for later analysis, and perform the heavy lifting with training and selecting models that are too resource-intensive to train (but not too intensive to run) in the embedded mode.
The data sent to the server is sampled with a counter-bias to avoid percolating class imbalance biases in the samples. For example, if over a period of time we receive 1,000 times more negative examples than positive ones, we need only log 1 in 1,000 negative examples to the server, while also indicating that it has a weight of 1,000. The servers visibility into the global distribution of the data usually leads to a better model than any individual nodes local model. None of this requires any setup, aside from linking to a library and using the two functions above.
In Spiral, learning starts as soon as feedback comes in. Prediction quality improves progressively as more feedback is generated. In most services, feedback is available within seconds to minutes, so the development cycle is very short. Domain experts can add a new feature and see within minutes whether it is helping to improve the quality of predictions.
Unlike hard-coded heuristics, Spiral-based heuristics can adapt to changing conditions. In the case of a cache admission policy, for example, if certain types of items are requested less frequently, the feedback will retrain the classifier to reduce the likelihood of admitting such items without any need for human intervention.
The first production use case for Spiral fit perfectly with Phil Karltons well-known quote, There are only two hard things in computer science: cache invalidation and naming things. (We already had an apt name for our project, so we did in fact tackle cache invalidation right away with Spiral.)
At Facebook, we rolled out a reactive cache that allows Spirals users (our other internal systems) to subscribe to query results. From the users perspective, this system provides the result of the query and a subscription to that result. Whenever an external event affects the query, it automatically sends the updated result to the client. This relieves the client of the burden of polling, and reduces load on the web front-end service that computes query results.
When a user submits a query, the reactive cache first sends the query to the web front-end, and then creates a subscription, and caches and returns the result. Along with the original result, the cache receives a list of objects and associations that were touched while computing the result. It then begins monitoring a stream of database updates to any objects or associations the query accessed. Whenever it sees an update that may affect one of the active subscriptions, the reactive cache reexecutes the query and compares the result with its cache. If the result did in fact change, it sends the new result to the client and updates its own cache.
One problem facing this system is that there is an enormous volume of database updates, but only a tiny fraction of them affect the output the query. If a query is interested in Which of my friends liked this post? it is unnecessary to get continuous updates on, for example, when the post was most recently viewed.
The problem is analogous to spam filtering: Given a message, should a system classify it as spam (does not affect query result) or ham (does affect query result)? The first solution was to manually create static blocklists. This was possible because the reactive cache engineering team recognized that over 99 percent of the load came from a very small set of queries. For low-volume queries, they simply assumed all updates were ham and reexecuted the query for every update to an object referenced by the query. For the small set of high-volume queries, they created blocklists by painstakingly observing the query execution to determine which fields in each object actually affected the output of the query. This process typically took a few weeks of an engineers time for each blocklist. To complicate things further, the set of high-volume queries was constantly changing, so blocklists quickly became outdated. Whenever a service using the cache changed the query it was executing, the system would have to change the spam-filtering policy, which required even more engineering work.
After reexecuting a query, it is easy to determine whether the observed updates were spam or ham by simply comparing the new query result with the old query result. This mechanism was used to provide feedback to Spiral, allowing it to create a classifier for updates.
To ensure unbiased sampling, the reactive cache maintains and only provides feedback from a small subset of subscriptions. The cache does not filter updates to these subscriptions; the query is reexecuted whenever a relevant object or association is modified. It compares the new query output with the cached version and then uses that to provide feedback to Spiral for example, telling it that updating last viewed at does not affect like count.
Spiral collects this feedback from all reactive cache servers and uses it to train a classifier for every distinct query type. These classifiers are periodically pushed to the cache servers. Creating filters for new queries or updating filters to respond to changing behavior in the web tier no longer requires any manual intervention from the engineering team. As feedback for new queries arrives, Spiral automatically creates a new classifier for those filters.
With a Spiral-based cache invalidation mechanism, the time required to support a new query in the reactive cache came down from weeks to minutes. Before Spiral, reactive cache engineers had to inspect each new querys side effects by running experiments and collecting data manually. With Spiral, however, most use cases (mapping to a query) are learned by the local model automatically within minutes, so the local inference is available immediately. The server is able to train a model using data from multiple servers in 10 to 20 minutes for most use cases. Once published to all of the individual servers, this higher-quality model is available for improved fidelity inference. When a query is altered, the servers are able to adapt to the change and relearn the new materiality patterns once they receive the updated queries.
We are continuing to work on automating backend services and applying machine learning for a better operational experience. Spirals potential future applications include continuous parameter optimization using Bayesian optimization, model-based control, and online reinforcement learning techniques targeting high-QPS real-time services as well as offline (batch) systems. Well continue to share our work and results in future posts.
Facebook believes in building community through open source technology. Explore our latest projects in Artificial Intelligence, Data Infrastructure, Development Tools, Front End, Languages, Platforms, Security, Virtual Reality, and more.
spiral classifier | screw classifier - jxsc machine
The Spiral Classifier is available with spiral diameters up to 120. These classifiers are built in three models with 100%, 125% and 150% spiral submergence with straight side tanks or modified flared or full flared tanks. The spiral classifier is one of the size classifying equipment for the mining industry. It is a kind of equipment for mineral classification based on the principle that the specific gravity of solid particles is different and the speed of precipitation in the liquid is different. It can filter the material powder from the mill, then screw the coarse material into the mill inlet with the spiral slice, and discharge the filtered fine material from the overflow pipe. Spiral classifier short for the classifier. The classification machine mainly has the high weir type single screw and the double screw, the sinking type single screw and the double screw four kinds of classification machines. The classifier is mainly composed of a transmission device, a spiral body, a trough body, a lifting mechanism, a lower bearing (Bush) and a discharge valve. Spiral classifier is widely used in mineral processing plants with a ball mill as a closed-circuit circuit to separate the flow of ore sand, or used in gravity concentrator to grade ore sand and fine mud, and metal beneficiation processes to grade the size of ore pulp and washing operations in the desliming hopper, dehydration and other operations.
Spiral classifier features: 1. Low power consumption; 2. Heavy-duty, long working life; 3. Powerful self-contained spiral lifting device; 4. Continuous spiral raking; 5. High classifying efficiency; 6. Wide choice of weir height; 7. Rigid tank and substructure; 8. Wide choice of tank design; 9. A wide range of industries serviced.
Spiral classifier is by means of solid particles of different sizes, the proportion of different, thus settling velocity in liquids of different principles, fine mineral particles floating in the water to overflow out of coarse mineral particles sink to the bottom. Discharged from the screw into the upper part, to a hierarchical classification of mechanical equipment, can mill to grind the material powder level in the filter, and then use the course material helical screw rotary vane into the mill feed, the filter out the fine material is discharged from the overflow pipe. The machine base is made of a channel steel body with steel plates welded together. Into the head of the screw shaft, shaft, using pig iron, wear-resistant and durable, lifting devices of electric and manual. The main types of the spiral classifier are high Weir single screw and double screw, low weir single screw and double screw, sinking single screw and double screw Mainly High Weir type and sinking type and XL spiral classifier.
investigation of particle dynamics and classification mechanism in a spiral jet mill through computational fluid dynamics and discrete element methods - sciencedirect
CFD study of milling fluid dynamics as a function of geometry and process parameters.DEM study of particle classification by the mill in agreement with cut-size model.Analysis of the cut-size dependence on classifier geometry and milling conditions.Study of particle collision energy and statistics for different milling conditions.Analysis of bottlenecks and further step towards realistic simulation of jet milling.
Predicting the outcome of jet-milling based on the knowledge of process parameters and starting material properties is a task still far from being accomplished. Given the technical difficulties in measuring thermodynamics, flow properties and particle statistics directly in the mills, modelling and simulations constitute alternative tools to gain insight in the process physics and many papers have been recently published on the subject. An ideal predictive simulation tool should combine the correct description of non-isothermal, compressible, high Mach number fluid flow, the correct particle-fluid and particle-particle interactions and the correct fracture mechanics of particle upon collisions but it is not currently available. In this paper we present our coupled CFD-DEM simulation results; while comparing them with the recent modelling and experimental works we will review the current understating of the jet-mill physics and particle classification. Subsequently we analyze the missing elements and the bottlenecks currently limiting the simulation technique as well as the possible ways to circumvent them towards a quantitative, predictive simulation of jet-milling.
spiral classifier, screw classifier
There are four types of classifiers, high weir type single and double spiral classifier, immersed single and double spiral classifier. Spiral classifiers are widely used for distributing ore in the close circuit with ball mill, grading ore and fine slit in the gravity mill, grading granularity in the flow of metal ore-dressing and desliming and dehydrating in the washing.