hydraulic system for classifier norberg mp

hydraulic system for crusher norberg mp1000

Hydraulic System For Crusher Norberg Mp1000. Hydraulic system for crusher norberg mp1000 hydraulic cone crusher,hydrauic stone crusher,hydraulic rock hydraulic cone crusher introduced the germany technology, not only improves the production capacity and crushing efficiency, but also widens applying area, hardness. get price

hydraulic valve Pressure transmitter and an electric pilot valve connected to a dedicated rapid sampling PLC system Hydraulic pressure sampling rate 200 times per second HYDRAULIC PRESSURE RELIEF VALVE System description Mechanical spring loaded hydraulic valve CRUSHER TRAMP IRON PROTECTION ... Sistem Hidrolik Untuk Crusher Norberg Mp1000 ...

Hydraulic System For Crusher Norberg Mp1000,Crusher Unit. ... and hydraulic cleaning system. cone crusher application: cone crusher is widely applied ... norberg crushing plant - Mining Equipment, Grinding Equipment . cone crusher norberg serie scs1 For sale Manufacturers. ... mp1000 crusher .

Hydraulic System For Crusher Norberg Mp1000Type of hydraulic relief valve for crusher mp - minevikype of hydraulic relief valve for crusher mpydraulic system for crusher norberg mp1000, s research and development has taken the mp line of cone crushers to the next level with the hydraulic adjust type of hydraulic relief valve for crusher mp1000draulic pressure problems crusher mills, .

Hydraulic System For Crusher Norberg Mp1000. Hydraulic System For Crusher Norberg Mp1000. Hydraulic system for crusher norberg mp1000. Hydraulic cone crusher,hydrauic stone crusher,hydraulic rock Hydraulic cone crusher introduced the germany technology, not only improves the production capacity and crushing efficiency, but also widens applying area, hardness .

Hydraulic System For Crusher Norberg Mp1000. Hydraulic system for crusher norberg mp1000. Hydraulic cone crusher,hydrauic stone crusher,hydraulic rock Hydraulic cone crusher introduced the germany technology, not only improves the production capacity and crushing efficiency, but also widens applying area, hardness .

Hydraulic system for crusher norberg mp1000 Hydraulic System For Crusher Norberg Mp1000 Hydraulic system for crusher ng mill chinahe ffworkssing principles of forcemultiplication lic system can generate over 2000 psi and impart 150 tons of crushing force onto a pile of scrap cars MOVE Type Of Hydraulic Relief Valve For Crusher Mp Magnetic .

Sistem hidrolik untuk crusher norberg mp1000 hydraulic system for crusher norberg mp1000, hydraulic cone crusher jual jaw crusher untuk laboratorium, more detail. mp1000 equipmentmine. mp1000 spare parts bowl 1 unit spare bowl for mp1000 cone crusher torch rings 6 units mp1000 cone crusher torch rings, part no. 63916310. Read More

Hydraulic System For Crusher Norberg Mp1000. Sistem Hidrolik Untuk Crusher Norberg Mp. Hp crusher hydraulic system somerers sistem hidrolik untuk crusher norberg mp1000ydraulic system for crusher norberg mpchine ice shaver snow cone cone esin concasseur batu chinoise sistem hidrolik untuk crusher norberg mporberg batu.

Hydraulic system for crusher norberg mp1000 what is coal mills used copper crusher in usa vertical roller mills manufacturer in india crushing technologies construction of a hammer mill 100 tph limestone crushing plant specifications crusher parts manufacturer korea crusher screening parts in billings mt horizontal raw mill in cement plant.

Hydraulic system for crusher norberg mp1000 hydraulic cone crusher,hydrauic stone crusher,hydraulic rock hydraulic cone crusher introduced the germany technology, not only improves the production capacity and . READ MORE. Hydraulic System For Crusher India - henrys .

Hydraulic system for crusher norberg mp1000 what is coal mills used copper crusher in usa vertical roller mills manufacturer in india crushing technologies construction of a hammer mill 100 tph limestone crushing plant specifications crusher parts manufacturer korea crusher screening parts in billings mt horizontal raw mill in cement plant.

Hydraulic system for crusher norberg mp1000 The company is located in Zhengzhou high tech Development Zone, 350000 square meters of modern standard factory buildings, more than 360 engineering and technical and after-sales service personnel, 40 years of the same mind, adhere to do high-quality, refined, scientific and technological mining machinery.

Hydraulic system for crusher norberg mp1000 what is coal mills used copper crusher in usa vertical roller mills manufacturer in india crushing technologies construction of a hammer mill 100 tph limestone crushing plant specifications crusher parts manufacturer korea crusher screening parts in billings mt horizontal raw mill in cement plant.

Hydraulic system for crusher norberg mp1000. Mp1000 Crusher Wet Ore In South Africa Edeka Magnetic Separation Production Line of Changning Manganese Mine in Hunan Province Type Of Hydraulic Relief Valve For Crusher Mp Magnetic Separation Production Line of Changning Manganese Mine in Hunan Province Crusher Crushing Services International ...

Hydraulic system for crusher norberg mp1000. Hydraulic cone crusher,hydrauic stone crusher,hydraulic rock Hydraulic cone crusher introduced the germany technology, not only improves the production capacity and crushing efficiency, but also widens applying area, hardness

Hydraulic System For Crusher Norberg Mp1000. Hp800 cone crusher parts sinco hydraulic system for crusher norberg mp1000,replacement hp800 cone crusher parts sinco machinery manufacture an extensive range of aftermarket crusher wear and spare parts to suit the mp1000 crusher maintenance manual spitsid brochure mp 1000 and mp maintenance.whether intended for a new as a secondary crusher.

Hydraulic system for crusher norberg mp1000, type of hydraulic relief valve for crusher mp1000 hydraulic system on osborn crusher and lubriion system . will check price mpj series stone mobile construction waste small mobile. read more stone crushers valves

HP800 Cone Crusher Parts | Sinco- hydraulic system for crusher norberg mp1000,Replacement HP800 Cone Crusher Parts SINCO Machinery manufacture an extensive range of aftermarket crusher wear and spare parts to suit the Mp1000 Crusher Maintenance Manual - spitseuBrochure MP 1000 and MP -, maintenanceWhether intended for a new As a secondary crusher, this MP1000.

HP800 Cone Crusher Parts | Sinco- hydraulic system for crusher norberg mp1000,Replacement HP800 Cone Crusher Parts SINCO Machinery manufacture an extensive range of aftermarket crusher wear and spare parts to suit the Mp1000 Crusher Maintenance Manual - spitseuBrochure MP 1000 and MP -, maintenanceWhether intended for a new As a secondary crusher, this MP1000.

Hydraulic system for crusher norberg mp1000-Henan . Hydraulic system for crusher norberg mp1000 The company is located in Zhengzhou high tech Development Zone, 350000 square meters of modern standard factory buildings, more than 360 engineering and technical and after-sales service personnel, 40 years of the same mind, adhere to do high-quality, refined, scientific and technological .

Hydraulic system for crusher norberg mp1000 hydraulic system for crusher norberg mp1000 the company is located in zhengzhou high tech development zone 350000 square meters of modern standard factory buildings more than 360 . get price. Crusher sistem hydraulik.

hydraulic system for crusher norberg mp1000 Construction Waste Crusher Construction waste refers to the construction, construction units or individuals to construct, lay or demolish all kinds of buildings, structures and pipe networks, etc., and generate the spoil, spoil, waste, residual mud and other wastes generated during the repairing process.

Hydraulic system for crusher norberg mp1000. Hydraulic cone crusher,hydrauic stone crusher,hydraulic rock Hydraulic cone crusher introduced the germany technology, not only improves the production capacity and crushing efficiency, but also widens applying area, hardness

Hydraulic System For Crusher Norberg Mp1000Type of hydraulic relief valve for crusher mp - minevikype of hydraulic relief valve for crusher mpydraulic system for crusher norberg mp1000, s research and development has taken the mp line of cone crushers to the next level with the hydraulic adjust type of hydraulic relief valve for crusher mp1000draulic pressure problems crusher mills, .

Hydraulic System For Crusher Norberg Mp1000. Hydraulic system for crusher norberg mp1000ydraulic cone crusher,hydrauic stone crusher,hydraulic rock mp series cone crushers pdf product presentation increasing capacity without signifi cantly increasing operating coststhats the. Get Price; Airsonic Grinder Usa

hydraulic system for mining mill norberg mp1000 - jscraftsin. hydraulic system for crusher norberg mp1000 - Norberg Crushing Equipment Crusher Mills,, hydraulic system for crusher norberg mp1000 hydraulic system for c. Get Price Online Chat MP 1250 Cone Crusher Datasheet --,

Hydraulic System For Crusher Norberg Mp1000. Hydraulic system for crusher norberg mp1000 Hydraulic cone crusher,hydrauic stone crusher, hydraulic rock Hydraulic cone crusher introduced the germany technology, not only improves the production capacity and crushing efficiency, but also widens applying area, hardness.

Hydraulic system for crusher norberg mp1000 hydraulic cone crusher,hydrauic stone crusher,hydraulic rock hydraulic cone crusher introduced the germany technology, not only improves the production capacity and . READ MORE. Hydraulic System For Crusher India - henrys .

Hydraulic system for crusher norberg mp1000 The company is located in Zhengzhou high tech Development Zone, 350000 square meters of modern standard factory buildings, more than 360 engineering and technical and after-sales service personnel, 40 years of the same mind, adhere to do high-quality, refined, scientific and technological mining machinery.

cone crushers - for demanding crushing needs - metso outotec

Cone crushers are very suitable for size reduction and shaping in the downstream of a crushing circuit. They reduce the material in a crushing cavity by continuous compression between a fixed element (bowl liner) and a moving element (mantle).

Our cone crusher offering consists of four different product families that utilize the same crushing principle but vary in features and optimal applications. In addition to stationary crushers, many cone crusher models are also available as mobile andportable versions.

Engineered for all rock types, Nordberg GP Series cone crushers can be utilized as secondary, tertiary, and quaternary crushers in aggregates production plants and in mining operations. Nordberg GP Series cone crushers are all-round crushing machines enabling smooth crushing process adaptation and full automation.

Nordberg HP Series cone crushers are well performing and reliable rock crushing machines for all aggregates production, quarrying and mining applications. They are normally utilized in secondary, tertiary and quaternary crushing stages.

Nordberg MP Series cone crushers have a high capacity and the highest crushing force of any cone crusher of similar size. Thanks to their crushing force, Nordberg MP Series cone crushers are widely used in the secondary and tertiary crushing stages in mining operations that process very large amounts of material.

Metso MX Series cone crushers are the latest addition to the cone crushers portfolio. MX Series cone crushers utilize a revolutionary technology that combines a rotating bowl with piston into one crusher.

Known for lower operating costs, high uptime and consistent high-quality output, Metso MX Series cone crushers are ideal for secondary, tertiary and quaternary stages in hard and soft rock applications.

guide to selecting the right hydraulic filter - crossco

Filtration, whether prioritized or not, is one of the most important components in a hydraulic system and can be the difference in a machine running for 10 years or having expensive component damage within a matter of seconds. Here, well lay out steps to help select the correct filter for your application. There are millions of applications where hydraulics are used, and its important to remember that every application is different. For our example, well focus on mobile hydraulic applications.

When selecting hydraulic filtration its best to determine the most sensitive component in the system, and then select the filtration cleanliness level (or micron rating) for that component. This may be your variable piston pump, or in many cases, the proportional valve. By using a recommended cleanliness code chart, we can determine the ISO code for a range of the most critical components such as pumps, valves, actuators, and more. Take a look at our full chart of recommended cleanliness levels.

Once we have determined the correct ISO code, we can then refer to the filter manufacturers catalogs. For our example, well use Schroeder Industries, as they are considered one of the leaders in mobile hydraulic filtration. You can see in our table, Schroeder recommends a range of media options based on desired cleanliness levels.

For example, if we are running a closed-loop hydraulic system with a variable piston pump and proportional control valves, we need to have an ISO cleanliness of 18/16/13. This would be a Z5 filter media (5 micron) based on what Schroeder has published.

While all of these factors are important, filter manufacturers (Schroeder, Parker, Eaton, Hydac) typically design their filters for maximum efficiency, so we are only going to focus on the last two: proper location and filter sizing. When done correctly, this will give the best service life of a system.

Suction filters are usually your rock and rag catchers that you put in a tank to filter out any bolts or rags that may have dropped in the reservoir, they are generally not used for filtering your fluid to the required ISO code.

Pressure filters come highly recommended because they are placed after your pump, so if the pump was to fail, all of the components downstream would still be protected. These filters must be rated for your systems running pressure and ISO code.

Return filters are important because they clean the oil before it returns to your hydraulic reservoir and therefore keep your reservoir clean and free from dirt particles. Return line filters are generally used in open-loop applications and should be rated at the ISO code that your system requires.

While all three filter types are generally not needed in the same system, it is important to have at least one pressure filter or one return line filter in every application to maintain the fluid cleanliness level that hydraulic components require.

Once you determine what type of filter to use, you must consider the pressure drop through the filter. In most cases, pressure filters and return filters have a bypass that protects the system when the filter gets clogged or plugged. It allows the oil to go around the filter element and allows the system to run even though its not being filtered. Do keep in mind that your system would be running unfiltered with this setup.

As your filter becomes dirtier, it requires more pressure to push the oil through the element because it is clogged with dirt and debris. Once the pressure required reaches a certain point it becomes inefficient and starts wasting horsepower in the hydraulic system; this is the reason for the bypass. Typically, bypass settings on pressure filters are 40-50 PSI to limit this inefficiency.

When sizing a pressure filter, generally we would like to keep the pressure drop through the filter under 10 PSI. This gives your filter more time to get to the 50 PSI bypass setting, thus increasing the length of time between element changes. If you size your filter with a 25 PSI pressure drop you are already at half of your filter life (assuming 50 PSI bypass). If we selected a filter that only had a 10 PSI pressure drop, your filter will have 40 PSI of pressure drop before it reaches bypass setting, thus, allowing the element to last longer before needing service or be replaced.

Another consideration is oil viscosity. Thicker oil will cause a higher pressure drop, so you need to know what oil you are going to run in your system to truly size the filter correctly and get the best filter life for your application.

Choosing the correct filter is only part of the equation. New hydraulic oil straight from the drum has a typical cleanliness level of ISO 4406 23/21/18. From what we learned above you can see that this is 16-64x dirtier than what most hydraulic systems require (each single number increase in the ISO code is double the contaminant level for that micron size)! To put it another way, a 25 GPM pump operating continuously in hydraulic oil at 23/21/18 will circulate 3,500 pounds of dirt to the hydraulic systems components each year.

To add hydraulic oil, and not the dirt, always filter new oil prior to use in a hydraulic system. This can be done in a number of ways. The most common way is to use a filtration cart or kidney loop filter in your reservoir. Schroeder makes an exceptional filter cart that can not only remove dirt particulates from new oil, but also water, if needed. Carts like these offer great value for your investment, as they range from relatively low cost to expensive, depending on what you are trying to accomplish. Also, if you are already running Schroeder mobile filtration on your machine, then there is a good chance you can use the same filter element that you already use, thus reducing inventory parts.

Another way to pre-filter your hydraulic oil is by pumping the oil into the hydraulic reservoir through the systems return filter. The easiest way to do this is to install a tee in the return line and attach a quick-connector to the branch of this tee. Attach the other half of the quick-connector to the discharge hose of a drum pump. When hydraulic oil needs to be added to the reservoir, the drum pump is coupled to the return line and the oil is pumped into the reservoir through the return filter. Benefits of this method include reduced spills and prevention of ingress of external contamination.

While many people dont even think about filtering new oil, it is very important to get clean oil into the system. It is much easier to prevent dirt from getting into a system by using precautionary measures than it is to remove dirt from a hydraulic system. Once the dirt has ingressed, it is very difficult to get the system clean.

Maintaining the hydraulic fluid in your machine is an important consideration when choosing and extending the life of hydraulic filters. Its challenging to set an expiration on hydraulic fluid introduced into your system, even under ideal circumstances. Over time even well maintained oil will wear out, however here are a few factors that affect hydraulic fluid and when filter changes are more than likely necessary.

Contamination, in the context of having to change your hydraulic fluid, means you have debris in the fluid than the filtration system can reasonably remove. This is usually some sort of particulate contamination event that overruns the systems onboard filtration. This can also include contaminating situations such as getting water mixed into the fluid (looks cloudy) or mistakenly topping off your hydraulic reservoir with the wrong fluid. It may be possible to salvage your particulate or water contamination situation using some sort of off-line filtration asset.

This one is simple. If you get your fluid gets too hot, it breaks down. Most of the time you know it got too hot because it becomes darker in color and it doesnt smell right. It usually doesnt take the time and expense of a fluid sample analysis to figure this one out. Heat accelerates the condition called oxidative degradation.

This one is a little more complex and will require a fluid analysis to determine. By performing routine fluid analysis a degradation or depletion trend can be spotted before it becomes a mechanical maintenance event.

A hydraulic oils oxidative degradation is determined by its Total Acid Number or TAN. As the name implies, this is the absolute measure of the total acid number in the fluid. Over time, oxygen will combine with the hydrocarbon molecules of the oil and a chain reaction occurs. This action results in some obvious conditions like darkened oil, varnishing, and sludge. Some conditions that are not so obvious are increased viscosity, increased foaming, and retained air.

A hydraulic oils additive depletion is determined by comparing the used oils elemental analysis to the baseline of identical new oil. For example, zinc is an antioxidant and anti-wear additive. Over time it gets depleted, so its important to check the concentration of zinc in your current oil to the concentration of zinc in the same new oil.

Cross Companys technical experts have many years of experience in applying the right components for maximum effectiveness and overall lower total cost of ownership. They can help you to improve efficiency and save money over the life of your equipment. Contact our team today for help in finding the correct filter and oil for your hydraulic system.

nave bayes classifier for debris flow disaster mitigation in mount merapi volcanic rivers, indonesia, using x-band polarimetric radar | springerlink

Debris flow triggered by rainfall that accompanies a volcanic eruption is a serious secondary impact of a volcanic disaster. The probability of debris flow events can be estimated based on the prior information of rainfall from historical and geomorphological data that are presumed to relate to debris flow occurrence. In this study, a debris flow disaster warning system was developed by applying the Nave Bayes Classifier (NBC). The spatial likelihood of the hazard is evaluated at a small subbasin scale by including high-resolution rainfall measurements from X-band polarimetric weather radar, a topographic factor, and soil type as predictors. The study was conducted in the Gendol River Basin of Mount Merapi, one of the most active volcanoes in Indonesia. Rainfall and debris flow occurrence data were collected for the upper Gendol River from October 2016 to February 2018 and divided into calibration and validation datasets. The NBC was used to estimate the status of debris flow incidences displayed in the susceptibility map that is based on the posterior probability from the predictors. The system verification was performed by quantitative dichotomous quality indices along with a contingency table. Using the validation datasets, the advantage of the NBC for estimating debris flow occurrence is confirmed. This work contributes to existing knowledge on estimating debris flow susceptibility through the data mining approach. Despite the existence of predictive uncertainty, the presented system could contribute to the improvement of debris flow countermeasures in volcanic regions.

Volcanic eruptions cause many direct and indirect hazards. Water flow caused by high precipitation on a volcano flank may be transformed into a debris flow. The rapid flow of volcanic material and water mixtures triggered by rainfall is a serious indirect hazard of volcanic eruptions. Debris flow that accompanied the 1991 Mount Pinatubo Philippine eruption killed 1500 people within 2 years. In Indonesia, the Mount Merapi eruption in 2010 caused 240 debris flow cases within 2 years. Debris flows occurred in almost all rivers, transported a large volume of material up to 15km (Blizal et al. 2013), and damaged 51 houses (Solikhin et al. 2015b).

Numerous studies have stated that high rainfall with a long duration is significantly associated with debris flow (Rodolfo and Arguden 1991; Lavigne and Thouret 2003; Blizal et al. 2013). However, the characteristics of debris flow in natural rivers can be very complex, combining several relevant factors besides rainfall. Material mobilization may be caused by slope instability, which is related to soil structure (Wilford et al. 2004). In the case of fine-grained material, earth material can be mobilized in the subaerial flow that forms a debris flow (Varnes 1978). There is a risk of debris flow when intense rain occurs on a steep slope with large amount of deposit material (Takahashi 2007). Particularly for debris flow in volcanic rivers, deposit material thickness and properties such as soil suction, pore-water pressure, and shear resistance create complex circumstances for debris flow occurrence (Berti et al. 2012; Mead and Magill 2017).

Debris flow vulnerability assessment is imperative to identify the susceptibility of the drainage basin and develop mitigation priorities. Some studies on the debris behavior on Merapi Volcano were aimed at developing the mitigation system (Blizal et al. 2013; Solikhin et al. 2015b). Modeling of the debris transport and integrating high-resolution rainfall observations with a lahar model were conducted by Syarifuddin et al. (2017). It was considered that a fine spatial rainfall measurement from X-band polarimetric weather (X-MP) radar in volcanic rivers would usefully extend the observation of small-scale rainfall in the typically narrow watershed that the debris flow is triggered in. Prediction of event and a warning play an important role in disaster mitigation to reduce the damages. However, none of the studies examined the prediction of debris flow susceptibility by introducing past rainfall information.

In the aftermath of volcanic eruption, timely and accurate prediction is a challenge to reliable decision making. More accurate predictive information can be obtained when the probability of past events is taken into account in estimating the next event likelihood. A previous study by Hapsari et al. (2017) used the ensemble radar-rainfall short-term prediction model in Merapi Rivers to develop a debris flow hazard map but did not include past information. However, a major problem with this kind of application is a lack of debris movement records at Merapi after a serious debris flow disaster from 2010 to 2011, including its characteristics. A reasonable approach to tackling this issue could be to apply simple probabilistic forecasting that could perform well when data are scarce.

The Bayesian approach is a probabilistic, simple classification algorithm that predicts the likelihood of upcoming events based on prior knowledge. There have been studies involving the Bayesian method in, for example, vulnerability assessments of water quantity and quality issues (Arthur et al. 2007; Pagano et al. 2014). The studies to date have tended to focus on susceptibility mapping rather than on the prediction of an event. The Nave Bayesian Classifier (NBC) algorithm offers an effective way of providing a probabilistic prediction system in many practical applications because this method is simple (Rish 2001), requires only small training datasets for calibration, and is more intuitive and thus does not require complex knowledge of the incident feature, compared to other similar techniques such as neural networks and support vector machines. Some studies have produced weather prediction systems using this algorithm (Liu et al. 2015; Barde and Patole 2016), but no study integrates various relevant attributes in predicting debris flow occurrence.

This study focuses on the performance measure of the NBC algorithms in identifying and predicting the vulnerable region to a debris flow in the Merapi area, aiming to develop a debris flow disaster warning system with small amount of past information. River catchment profile and rainfall (derived from X-MP radar) are regarded as the independent variables that lead to the debris transport, that is, working rainfall, hourly rainfall, soil type, and slope angle parameters. Section2 outlines the NBC algorithm and the development of the prediction model, as well as the data used for building the model and for validation. Section3 explains the testing of the model and its performance, and discusses the NBC model improvement and the application in the practical scheme.

This section provides an outline of the research methodology used to address the research questions including a description of study site, NBC framework, and the performance evaluation technique. Secondary data collection, radar rainfall data pre-processing, and debris flow susceptibility factors are also explained in this section.

The Merapi Volcano is an active stratovolcano located in the center of Java Island, Indonesia (73224S, 1102700E), rising to 2930m above sea level (Fig.1). The volcano is geographically shared by Central Java Province and Yogyakarta Special Province. The Merapi Volcano is historically the most active volcano in Indonesia, with small eruptions every 23years and a greater eruption every 1015years. The Merapi flank is a densely populated area, where 70,000 people live in the third danger zone, the most hazardous zone according to the Decree of Central Java Governor number 6/2018Footnote 1 concerning Merapi contingency plan. Destructive eruptions occurred in 1872, 1953, 2006, and 2010. After the October 2010 eruption, 130 million m3 of material were ejected as lava and tephra material. The pyroclastic flow on the southern flanks caused high damage within a 22km range. As many as 120 people living at 12km distance from the summit were killed (Jenkins et al. 2013). The eruption caused mudflows in the subsequent rainy season. At the beginning of 2011, 60% of the material accumulation at the top was transported downstream by the rivers. The Gendol River downstream was highly impacted by debris flow on 1 May 2011. The paddy field and a village were destroyed by the flow. The debris flow buried 40,000m2 of crops and damaged 51 houses (Blizal et al. 2013). On average, the annual rainfall amount on the Merapi slope is 2600 to 3000mm. Debris flow is promoted by the steepness of 10016% slope at the altitude range between 1000m and 2930m.

Between 2015 and 2018, at least nine debris flow events were documented. The affected rivers were the Pabelan, Krasak, Putih, Gendol, and Blongkeng Rivers. The Gendol River is located on the southern flank of the Merapi and was chosen as the focus area of study because, as stated by the Research and Development Center of Geological Disaster Technology (BPPTKG) Yogyakarta, this river has the most streams vulnerable to debris flow besides the Pabelan River. In October 2010, a debris flow disaster on the Gendol River extended 20km downslope and buried 21 houses. The Gendol River is the main branch of the Opak River. Solikhin et al. (2015a) investigated this area because this river system was subject to numerous debris flow occurrences in 2011 due to great pyroclastic deposits in the Opak River.

Figure1 shows the Gendol watershed delineation where the water flows downstream from the Merapi summit through a tributaries network and transports the debris into the Opak River. The upper catchment of the Gendol River is the area where the highly concentrated flow reached this area. The delineated basin encompasses a 5.1km2 drainage area with a 4.28km length. Like on a typical stratovolcano, the river basin is characterized by a narrow watershed and a steep-walled valley. This basin shape leads to flash floods that promote debris transport along the river. The debris flow events in the Gendol River that are available as observable events occurred 17 February 2016 and 20 December 2017. In the 2016 case, the Gendol River experienced high rain of up to 23.0mm/h and a high concentration of sediment flow. In the 2017 case, 18.7mm/h rainfall triggered earth material flow in the river.

Since 2015, the X-band polarimetric weather radar (X-MP radar) has been installed on Mount Merapi (73712S, 1102512E). The X-MP radar is an appealing instrument for observing rainfall over a large areain the Merapi area within a 30km radius in fine resolutionscompared to the available rain gauges. The radar provides rainfall measurements with a 150m spatial resolution and a 2 min temporal resolution, which is advantageous for observing small-scale events. The polarimetric feature offers various advantages, including its capability to distinguish scattered particles, to overcome the rain attenuation, and to minimize the effect of uncertain particle drop size distribution.

Rainfall intensity data from X-MP radar is acquired in the constant altitude plan position indicator/CAPPI format at the lowest elevation angle. The radar-rainfall algorithm applied by Merapi X-MP radar is the composite algorithm proposed by Park et al. (2005) that uses horizontal reflectivity and a specific differential phase as indicated below:

Earth material movement in natural rivers arises from a complex interaction between hydrological and basin physical factors. A data mining technique for the prediction of an event occurrence is applied to the datasets based on these risk factors. The risk factors are rainfall from historical data, topsoil characteristics, and slope angle.

Most studies on debris flow disasters have emphasized the use of a rainfall threshold to judge the possible occurrence of debris flow (Shieh et al. 2009; Brunetti et al. 2010). Hourly rainfall intensity and working rainfall parameters are broadly used as criteria to judge debris flow initiation. Working rainfall is rainfall preceding an event, calculated by accumulating rainfall in the 7 days prior to the hourly rainfall (Lavigne et al. 2000). This parameter plays a significant role in debris flow occurrence because the increase of pore-water pressure due to accumulated rainwater promotes soil instability. In this study, X-MP radar rainfall observations in the rainy season of 2016 and 2017 are included in the rainfall database.

Previous research by the authors investigated the rainfall threshold that is likely to trigger debris flow in the Gendol River. The threshold was developed by separating debris flow and non-debris flow events from hourly and 7-day working rainfall obtained through X-MP radar observations during 20152019 (Hapsari et al. 2019). There is a critical line that distinguishes a safe zone and an unsafe zone, and a vertical line that represents the standard rainfall to issue a warning. Figure2 indicates the critical line for debris flow emergency judgment.

Slope steepness information is constructed from the Digital Elevation Model from the Shuttle Radar Topography Mission (SRTM) with a 30m spatial resolution (Fig.3a). In the upper Gendol River, the terrain slope angle ranges from 1.83 to 32.9. The terrain data are resampled to be the same resolution as the base rainfall spatial data, at 0.15km0.15km. The level of slope susceptibility to failure for different slope angle classes is adapted from Niu et al. (2014), who assessed the susceptibility of slope failures. Rank 1 is assigned to the slope of 03; rank 2 is assigned to the slope of 36; rank 3 is assigned to the slope of 610; rank 4 is assigned to the slope of 1015; and rank 5 is assigned to the slope of more than 15.

The structure of the deposited material has been identified as a contributing factor of debris flow mobilization. However, difficulties arise because of lack of data on the variability of deposited soil properties. It is believed that debris flow formation is closely related to the topsoil type. The soil data are drawn from the Harmonized World Soil Database from the UNESCO Digital Soil Map of the World with a 30 arc-second raster resolution. Figure3b shows the soil map of the area. Andosols and Arenosols are the dominant soil groups in the Gendol watershed. The texture of the topsoil is classified as loam and loamy sand. The loam soil texture is composed of 42% sand, 39% silt, and 19% clay; the loamy sand soil texture is composed of 83% sand, 11% silt, and 6% clay. According to Zhao and Zhang (2014), saturated silt and fine sand are unstable in the undrained shearing condition. It can, therefore, be assumed that the Gendol River is vulnerable to debris movement.

Bayes theorem gives the probability of an event based on the prior information of a condition related to the event (Bayes 1763). The posterior probability of an event is the transformation of prior knowledge combined with new data after the evidence is considered into the posterior probability. The following equation represents the basic statement of the Bayesian probability:

The Nave Bayes approach is a simple probabilistic classification method that calculates the likelihood by summarizing frequency and combination of the given dataset value. It is based on the Bayesian theorem if one assumes that all attributes that are given by the value of the variable classes are conditionally independent of each other. The basic formula of the NBC is:

where P(class|data) is a posteriori probability or the probability of a class given an event after seeing the event; P(data|class) is a likelihood or probability of an event such that the event belongs to a particular class; P(class) is a priori probability or past event occurrence probability; and P(data) is the probability of that event in the whole dataset (typically omitted). In the following section, the development of the NBC and its adaptation for debris flow prediction based on the observable attributes is described.

The datasets of debris flow occurrences and the attributes are divided for calibration and validation purposes. Rainfall intensity, working rainfall, topographical slope, and soil type are involved as mutually independent variables in the model calibration process. As the debris flow may occur in a small localized area, the predictive analytics are developed spatially on a grid basis. The study area consists of 167 grids of 0.15km0.15km. The rainfall data are divided into two categoriestriggering and non-triggering rainfall. In this study, the time interval of radar rainfall observation used in the analysis is 30min. For one debris flow case, three consecutive periods of triggering rainfall data are included. This is because the duration of the debris flows was about 1.5h according to the historical record. Therefore, the number of data points for one debris flow case is 501 (1673).

In the calibration dataset, there were one triggering and 237 non-triggering rainfall events during the rainy season of 20152016. For non-triggering rain, one single rain event is defined as the maximum hourly rainfall in a series of rain. Therefore, the number of data points is 39,579 (167237). Table1 presents the description of the calibration and validation datasets that involve temporally and spatially varying data. Validation is conducted with test datasets from the December 2017 debris flow case and the October 2016 no-debris flow cases consisting of 1002 data points. The number of data points for each slope and soil type classes is temporally constant.

Prior and conditional probabilities of debris flow occurrences were estimated, using the calibration data (Table1) to obtain the posterior probabilities. Based on the assumption of independence of the variables, the likelihood of the event belonging to classj is:

The frequency ratio dataset based on the data categorization is constructed by first calculating the number of occur and no-occur events for each predictor class. The probability of an event such that the event belongs to a class is then calculated. After this, the likelihood P(data|class) is obtained by multiplying the probability for three predictors ( Eq.5). After obtaining the prior probability from the ratio of debris flow cases to all cases, the posterior probability can be determined by using Eq.4.

In a classification algorithm, the next phase after model calibration is validation (also called the testing phase). In this phase, another unseen dataset (see validation data in Table1) is introduced to the model. The NBC attempts to predict the label of an individual example based on the attribute-label relationship learned from the calibration datasets and the corresponding class. The category label of debris flow status uses the dichotomous index, that is, whether debris flow occurs and does not occur that is determined from the posterior probability. The posterior probability of classj given a new event data will be:

In this procedure, the prediction outcome was obtained and presented in the form of a spatial distribution. The event predicted by the NBC is then compared with the actual occurrence/non-occurrence as recorded in the debris flow event inventory. The ability of the model to learn the data and make a prediction was assessed through accuracy measures. Success by the model means that the event was assigned to the correct category (occurrence or non-occurrence). In addition, tee widely used quantitative indices were applied to compare the prediction and observationthe critical success index (CSI), the probability of detection (POD), and the false alarm rate (FAR) that come with a contingency table (Roebber 2009):

where Nhit is the number of hit events (the model predicts debris flow, and debris flow presents); Nmiss is the number of miss events (debris flow presents, but the model does not predict debris flow); and Nfalse is the number of false events (the model predicts debris flow, but there is no debris flow). For CSI and POD, 1 represents a perfect forecast, whereas for FAR, a perfect forecast is represented by 0. After satisfying performance is obtained, the debris flow prediction model and the disaster warning system based on the NBC approach is developed. Figure4 illustrates the procedure of the debris flow warning system development using the NBC approach.

This section describes analysis and evaluation of the NBC approach for debris flow warning. A section that explains the model calibration and validation is presented first, followed by the discussion of performance and implication of the model. To corroborate the model results, catchment aerial photo for illustrating the potential of the catchment to initiate the debris flow occurrence is explained.

In an attempt to train the model in the calibration stage, the original datasets were converted into frequency ratio datasets based on the categorization. The results of the predictor classification are presented in Table1. From the 167 grids, the number of grids with slopes of 00.0524, 0.05240.1051, 0.10510.1763, 0.17630.2679, and more than 0.2679 are 1, 17, 31, 41, and 77, respectively. Figure3c presents the classification of the slope angle according to the designated rule. Regarding the soil type, the area is composed of 19 grids of Andosols and 148 grids of Arenosols.

The rainfall for seven days before specific hourly rainfall is accumulated to classify the rainfall-based debris flow critical line (Fig.2). The average and range of hourly rainfall from calibration datasets are 0.702mm/h and 0.00081.500mm/h, respectively. The average and range of working rainfall are 97.160mm and 0.005205.920mm, respectively. The number of data points with the alert status of safe, warning, and emergency in the calibration datasets are 20,045, 20,019, and 16, respectively (Table2).

The next step of the NBC algorithm was to analyze the likelihood P(B|A) of occurrence or non-occurrence for each predictor class. With 501 debris flow cases out of 40,080 total cases, the ratios of debris flow occurrence to total events or the prior probability P(B|A) are 0.0125 and 0.9875 for occurrence and non-occurrence, respectively. Through this procedure, a model that generalizes how the debris flow attributes relate to the disaster occurrence status has been obtained.

After the conditional probability of debris flow occurrence had been calculated considering the event attribute classification from the validation datasets, the prediction outcome was obtained as the class with the highest probability. In this section, some examples of validation results illustrated in spatial distribution are given. Figure5 depicts the rainfall map presenting hourly rainfall on 16 February 2016, 14:00 local standard time (LST); 17 February 2016, 16:30 LST; 16 December 2017, 13:00 LST; and 20 December 2017, 16:30 LST. The working rainfall for these cases is shown in Fig.6. The event on 16 February 2016 (Figs.5a, 6a) was a case with strong hourly rainfall intensity but low working rainfall. On 17 February 2016 (Figs.5b, 6b), the upper Gendol River Basin was subject to high rainfall intensity and high working rainfall.

Rainfall input data for model testing, shown as rainfall intensity for the upper Gendol River Basin, Indonesia: a 16 February 2016, 14:00 local standard time (LST); b 17 February 2016, 16:30 LST; c 16 December 2017, 13:00 LST; and d 20 December 2017, 16:30 LST

Rainfall input data for model testing, shown as working rainfall for the upper Gendol River Basin, Indonesia: a 16 February 2016, 14:00 local standard time (LST); b 17 February 2016, 16:30 LST; c 16 December 2017, 13:00 LST; and d 20 December 2017, 16:30 LST

The experimental results of testing the NBC using this case study are illustrated in debris flow hazard maps (Fig.7). The Gendol River indicated by blue lines is overlaid with the maps to allow for a more detailed qualitative assessment. Figure7a illustrates the results for the simulation on 16 February 2016, 14:00, which is identified as a no-debris flow case throughout the area. The result matches with the debris flow disaster inventory reports that debris flow was not observed in the study area. When the rainfall of 17 February 2016, 16:30 is introduced as the model input, the result as presented in Fig.7b demonstrates debris flow that occurred in the downstream area. This finding conforms with the debris flow database report that there was a strong debris flow that caused two trucks to be trapped by debris in the Gendol River at 16:30.

Model results from cases in the upper Gendol River Basin, Indonesia: a 16 February 2016, 14:00 local standard time (LST); b 17 February 2016, 16:30 LST; c 16 December 2017, 13:00 LST; and d 20 December 2017, 16:30 LST

Another example is the model testing using validation datasets on 16 December 2017, 13:00 and 20 December 2017, 16:30. On 16 December 2017 at 13:00, which was marked as a no-debris flow case, the hourly rainfall intensity was 6.24mm/h, and the working rainfall was 120.33mm. On 20 December 2017, 16:30, when a serious debris flow was recorded, the hourly rainfall intensity was 40.484mm/h, and the working rainfall was 149.883mm. Figure7d shows that the hazard map from the NBC on 20 December 2017, 16:30 indicates debris flow occurrence throughout the basin. This result matches those observed in the upper Gendol River Basin, where the debris mobilization occurred at the Gendol River that caused heavy equipment to be buried by sediment at 17:00. Low rainfall indices on 16 December 2017, 13:00 (Figs.5c, 6c) tend to provide similar results to those in Fig.7a, where the NBC model simulates the non-occurrence status (Fig.7c).

The formation of debris flows differs from landslides. Landslides occur on the slope due to slope instability. Debris flow occurs in the river channel, initiated by material mobilization due to water flow (Takahashi 2007). The issue in this study is that each grid is treated independently as a different sample in the calibration phase as applied in landslide vulnerability mapping (Berti et al. 2012; Bui et al. 2012). The Merapi debris flow events were mainly determined from media and resident reports because there was no sediment measurement on the rivers during the observation period. As a result, the timing and the subcatchment location where debris flows occurred are not well identified. Also, in the experimental setting, among the three predictors, rainfall is the only parameter that varies temporally as the slopes and the soil types are constant. Consequently, the model is very sensitive to the changes in the rainfall parameters. For example, the model estimated the debris flow occurrence in the downstream area in the 17 February 2016, 16.30 case (Fig.7b) because this area had somewhat high working rainfall (Fig.6b). In contrast, debris flow did not occur on 16 February 2016, 14.00 (Fig.7a) because the working rainfall was low (Fig.6a), although the rainfall intensity was high (Fig.5a). In the 20 December 2017, 16.30 case, debris flow was estimated to occur throughout the basin (Fig.7d) because the whole area had high working rainfall (Fig.6d).

To illustrate the vulnerability of the catchment more clearly, high-resolution satellite imagery of the upper Gendol River recently retrieved from Pleaides Imagery with a 0.5m grid size is presented in Fig.8. The image indicates that the mountain top is now almost entirely bare of vegetation. The topsoil of upper Mount Merapi is generally formed from pyroclastic material from the last eruption. Arenosols dominate the soil formation of the upper Gendol Basin, though Andosols are more common at the volcano peak. Some of the areas of the river channel with high material deposits are indicated by the white-framed areas. Almost all river reaches on the upper Gendol have high remaining volcanic mud. Generally, the entire basin is prone to debris flows given the soil and land cover characteristics and the material deposits, as illustrated in Fig.8.

In the validation stage, comparing the hit and the null from the contingency table with total cases gives 82.28% model accuracy. Further quantitative skill assessment of the model indicates CSI, POD, and FAR of 0.56, 0.56, and 0.00, respectively. Overall, the NBC performs well in predicting the debris flow occurrence. The CSI and POD indicate somewhat low performance. A possible explanation is that the model predicts the vulnerable area unevenly based on variation of all three predictors, whereas the exact location and the extent of flooded areas are unknown because of limited information. A perfect skill is obtained in terms of FAR since all cases without debris flow are simulated as non-occurrence by the model. Altogether, the encouraging performance of the new NBC is further supported by the satisfying accuracy.

This experiment is based on a debris flow database since 2015 that consists of only two debris flow events, which are insufficient to allow the model to learn the features. The evidence presented in this article suggests that the NBC approach with a longer calibration dataset could improve prediction capability. In the NBC, an imbalanced calibration dataset between occurrence and non-occurrence can lead to bias in the construction of the model as stated by Lane et al. (2012). Examples of NBC application for water-related hazards are studies carried out by Bui et al. (2012) that included 118 occurrences cases and Liu et al. (2015) that used 1000 sampling points. A further study with more debris flow cases from all basins in the Merapi area is therefore suggested.

Future studies on the current NBC application also could include temporal and spatial dynamics of soil moisture from remote sensing data (Liu et al. 2015) because there is a strong relationship between soil water content and active debris flow (Chorowicz et al. 1997; Capra et al. 2010; Pham et al. 2016). By using the NBC, we expect to discover hidden patterns that complicate the estimation from existing geomorphological and hydrological data that lead to debris flow.

This study set out to evaluate the predictability of a debris flow disaster warning system by applying the Nave Bayes Classifier (NBC) in the upper Gendol River located on the Merapi Volcano flank in Indonesia. The investigation indicated that the NBC is a simple data mining technique that provides acceptable results and accuracy in deriving a debris flow susceptibility map. The qualitative instant skill assessment through visual comparison found that on some occasions, the occurrence grids in the hazard map were not exactly the same as the observed data. However, overall the NBC performed quite well in predicting the debris flow occurrence indicated by CSI, POD, and FAR of 0.56, 0.56, and 0.00, respectively. But the findings in this study are subject to at least three limitationslack of a representative sample; inadequate fine data of temporal and spatial debris flow observations; and lack of prior information on soil moisture. Involving all basins in the Merapi area as calibration data by considering watershed morphometrics should be pursued in a future investigation to develop a robust model. The system based on such risk factors could give a warning and suggestion regarding the probable occurrence of debris flows and help to reduce the negative impact of a debris flow disaster.

Arthur, J.D., H.A.R. Wood, A.E. Baker, J.R. Cichon, and G.L. Raines. 2007. Development and implementation of a Bayesian-based aquifer vulnerability assessment in Florida. Natural Resources Research 16(2): 93107.

Blizal, E.D., F. Lavigne, D.S. Hadmoko, J.P. Degeai, G.A. Dipayana, B.W. Mutaqin, M.S. Marfai, M. Coquet, et al. 2013. Rain-triggered lahars following the 2010 eruption of Merapi volcano, Indonesia: A major risk. Journal of Volcanology and Geothermal Research 261: 330347.

Berti, M., M.L.V. Martina, S. Franceschini, S. Pignone, A. Simoni, and M. Pizziolo. 2012. Probabilistic rainfall thresholds for landslide occurrence using a Bayesian approach. Journal of Geophysical Research 117: Article F04006.

Brunetti, M.T., S. Peruccacci, M. Rossi, S. Luciani, D. Valigi, and F. Guzzetti. 2010. Rainfall thresholds for the possible occurrence of landslides in Italy. Natural Hazards and Earth System Sciences 10: 447458.

Bui, D.T., B. Pradhan, O. Lofman, and I. Revhaug. 2012. Landslide susceptibility assessment in Vietnam using support vector machines, decision tree, and Nave Bayes models. Mathematical Problems in Engineering 2012: Article 974638.

Capra, L., L. Borselli, N. Varley, J.C.G. Ruiz, G. Norini, D. Sarocchi, L. Caballero, and A. Cortes. 2010. Rainfall-triggered debris at Volcn de Colima, Mexico: Surface hydro-repellency as initiation process. Journal of Volcanology and Geothermal Research 189: 105117.

Hapsari, R.I., S. Oishi, M. Syarifuddin, and R.A. Asmara. 2017. Uncertainty in debris flow hazard estimation using X-MP radar rainfall forecast. In Proceedings of the 37th IAHR World Congress, 1318 August 2017, Kuala Lumpur, Malaysia, 110.

Hapsari, R.I., S. Oishi, M. Syarifuddin, R.A. Asmara, and D. Legono. 2019. X-MP Radar for developing a lahar rainfall threshold for the Merapi Volcano using a Bayesian approach. Journal of Disaster Research 14(5): 811828.

Jenkins, S., J.-C. Komorowski, P.J. Baxter, R. Spence, A. Picquout, F. Lavigne, Surono. 2013. The Merapi 2010 eruption: An interdisciplinary impact assessment methodology for studying pyroclastic density current dynamics. Journal of Volcanology and Geothermal Research 261: 316329.

Lane, P.C.L., D. Clarke, and P. Hender. 2012. On developing robust models for favourability analysis: Model choice, feature sets and imbalanced data. In Proceedings of the 2nd Workshop on Computational Approaches to Subjectivity and Sentiment Analysis, ACL-HLT 2011, 24 June 2011, Portland, Oregon, USA, 4452.

Liu, R., Y. Chen, J. Wu, L. Gao, D. Barrett, T. Xu, L. Li, C. Huang, et al. 2015. Assessing spatial likelihood of flooding hazard using Naive Bayes and GIS: A case study in Bowen Basin, Australia. Stochastic Environmental Research and Risk Assessment 30: 15751590.

Niu, F., J. Luo, Z. Lin, M. Liu, and G. Yin. 2014. Thaw-induced slope failures and susceptibility mapping in permafrost regions of the QinghaiTibet Engineering Corridor, China. Bulletin of Volcanology 59: 460480.

Park, S.G., M. Maki, K. Iwanami, V.N. Bringi, and V. Chandrasekar. 2005. Correction of radar reflectivity and differential reflectivity for rain attenuation at X Band Part II: Evaluation and application. Journal of Atmospheric and Oceanic Technology 22(11): 16331655.

Pham, B.T., I. Prakash, D.T. Bui, and M.B. Dholakia. 2016. Evaluation of predictive ability of support vector machines and naive Bayes trees methods for spatial prediction of landslides in Uttarakhand state (India) using GIS. Journal of Geomatics 10(1): 7179.

Rodolfo, K.S., and A.T. Arguden. 1991. Rain-lahar generation and sediment delivery-systems at Mayon Volcano Philippines. In Sedimentation in volcanic settings, ed. R.V. Fisher, and G.A. Smith, 7187. Tulsa, OK: SEPM (Society for Sedimentary Geology) (SEPM Special Publication 45).

Shieh, C.L., Y.S. Chen, T.J. Tsai, and J.H. Wu. 2009. Variability in rainfall threshold for debris flow after the Chi-Chi earthquake in central Taiwan, China. International Journal of Sediment Research 24(2): 177188.

Solikhin, A., J.-C. Thouret, A. Gupta, D.S. Sayudi, J.-F. Oehler, and S.C. Liew. 2015a. Effects and behavior of pyroclastic and debris deposits of the 2010 Merapi Eruption based on high-resolution optical imagery. Procedia Earth and Planetary Science 12: 110.

Solikhin, A., J.-C. Thouret, S.C. Liew, A. Gupta, D.S. Sayudi, J.-F. Oehler, and Z. Kassouk. 2015b. High-spatial-resolution imagery helps map deposits of the large (VEI 4) 2010 Merapi Volcano eruption and their impact. Bulletin of Volcanology 77(3): 142.

Syarifuddin, M., S. Oishi, D. Legono, R.I. Hapsari, and M. Iguchi. 2017. Integrating X-MP radar data to estimate rainfall induced debris flow in the Merapi volcanic area. Advances in Water Resources 110: 249262.

This research was supported by the Science and Technology Research Partnership for Sustainable Development (SATREPS), Japan Science and Technology Agency (JST), and the Japan International Cooperation Agency (JICA). The authors thank the Hydraulic Laboratory of Universitas Gadjah Mada (UGM) for providing rainfall data for radar validation.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Hapsari, R.I., Sugna, B.A.I., Novianto, D. et al. Nave Bayes Classifier for Debris Flow Disaster Mitigation in Mount Merapi Volcanic Rivers, Indonesia, Using X-band Polarimetric Radar. Int J Disaster Risk Sci 11, 776789 (2020). https://doi.org/10.1007/s13753-020-00321-7