if you want the project pls call @8125424511
Data-Driven Design of Fog Computing aided Process Monitoring System for Large-Scale Industrial Processes
Now-a-days all major manufacturing factories such as chemical factories, iron ore factories, coal factories etc has automated their manufacturing units to control temperatures of their reactors and other machines. Normally all organizations are equipped with close loop technologies which means all sensors will not require input from humans to take decision and to cool down reactors upon reaching low or high temperature levels.
Taking above decisions require heavy computation and resources at manufacturing units and to reduce this burden new technologies such as Fog Computing with IOT connections (Internet of Things) is introduce. With aide or support of fog computing process monitoring task in industries can be managed easily. Here sensors at manufacturing or industries units will sense data of different process and offload processing task to fog computing to execute heavy computation task to take decision on processes. This industrial process consists of states such as making boiling steels, empty boiled steels, cool down reactors upon reaching max level and each states will have some values and this values can be used to form vectors.
All normal states will have normal level values which are given as threshold and if that matrix contains abnormal values more than threshold then it will be consider as faulty values. So using fog computing we can easier industrial process monitoring task.
Offloading process to fog computing is also called as decentralized concept or fog computing. Taking decision on available data by using data mining algorithms will be called as Data Driven and monitoring states of industrial process automatically will be called as Industrial Processes Monitoring.
In this project with all normal values we will form vector and upon reading new values then we will apply new values on vector to detect whether new value contains faulty data or correct data and based on that application will take appropriate data.
To build this application we are using dataset from industrial process and this dataset contains 52 columns (variables input data) such as reactors, temperatures etc and below are all column names of that dataset in bold format.
A_feed,D_feed,E_feed,A_C,Reactor_feed_rate,Reactor_pressure,Reactor_level,Reactor_temperature,Separator_temperature,Separator_level,Separator_pressure,Separator_underflow,Stripper_level,Stripper_pressure,Stripper_underflow,Stripper_temperature,Stripper_steam_flow,Recycle_flow,Purge_rate,Compressor_work,Reactor_water_temperature,Separator_water_temperature,Component_A,Component_B,Component_C,Component_D,Component_E,Component_F,Component_A,Component_B,Component_C,Component_D,Component_E,Component_F,Component_G,Component_H,Component_D,Component_E,Component_F,Component_G,Component_H,D_feed_flow,A_feed_flow,E_feed_flow,A_C_feed_flow,Compressor_recycle_valve,Purge valve,Separator_pot_liquid_flow,Stripper_liquid_product_flow,Stripper_steam_valve,Reactor_cooling_water_flow,Condenser_cooling_water_flow,class
Below are some values from dataset
2.498,2.511,2.518,2.514,2.410,2.431,2.612,2.600,2.520,2.551,2.289,2.304,2.430,2.410,2.316,2.322,2.588,2.597,2.640,2.660,2.543,2.553,2.691,2.662,2.288,2.259,2.674,2.679,2.636,2.627,2.556,2.555,2.487,2.499,2.999,2.983,2.642,2.631,2.773,2.790,2.409,2.408,1.904,1.882,2.395,2.390,2.243,2.255,2.416,2.421,2.437,2.436,0
3.642,3.694,3.683,3.653,3.629,3.681,3.712,3.722,3.684,3.643,3.674,3.642,3.667,3.650,3.678,3.661,3.643,3.615,3.696,3.662,3.654,3.644,3.669,3.683,3.672,3.643,3.648,3.670,3.677,3.651,3.652,3.668,3.633,3.617,3.593,3.647,3.695,3.669,3.637,3.649,3.695,3.682,3.662,3.682,3.664,3.650,3.716,3.686,3.626,3.664,3.631,3.628,0
In above records if last value contains 0 then its normal values not faulty values. See below another records with faulty values
5.907,3.651,4.514,9.151,2.697,4.256,2.790,7.550,1.204,2.975,7.926,5.079,2.717,2.444,5.014,3.206,2.132,6.526,2.329,3.298,9.454,7.704,3.004,8.596,2.958,6.890,1.864,1.613,2.963,1.365,2.858,1.313,1.868,2.205,4.650,2.143,2.854,8.311,1.011,5.386,4.413,6.296,5.398,5.843,6.154,2.041,3.584,4.042,4.687,4.704,4.126,2.194,1
5.920,3.671,4.514,9.123,2.685,4.238,2.793,7.542,1.203,2.993,7.892,5.220,2.719,2.465,5.015,3.208,2.275,6.516,2.318,3.299,9.473,7.687,3.004,8.596,2.958,6.890,1.864,1.613,2.963,1.365,2.858,1.313,1.868,2.205,4.650,2.143,2.854,8.311,1.011,5.386,4.413,6.262,5.399,5.830,6.100,2.046,3.562,4.458,4.688,4.698,4.052,1.861,1
In above two records last values contains 1 value which indicates as faulty values. New values received from sensor will apply on above records to detect whether it contains normal or faulty values. Below are some new records which don’t have either 0 or 1 value but fog computing will analyze and give that value
2.488,3.702,4.502,9.417,2.699,4.218,2.705,7.517,1.204,3.361,8.006,5.046,2.633,2.518,5.020,3.102,2.261,6.572,2.286,3.410,9.459,7.727,3.218,8.893,2.638,6.882,1.877,1.656,3.295,1.382,2.397,1.256,1.857,2.263,4.843,2.298,1.786,8.357,9.857,5.372,4.382,6.239,5.405,2.480,6.326,2.195,4.018,3.946,4.700,4.759,4.138,1.890
2.490,3.666,4.526,9.268,2.671,4.233,2.705,7.441,1.204,3.367,8.009,5.190,2.634,2.658,4.982,3.102,2.258,6.572,2.296,3.414,9.468,7.738,3.218,8.893,2.638,6.882,1.877,1.656,3.295,1.382,2.397,1.256,1.857,2.263,4.843,2.298,1.786,8.357,9.857,5.372,4.382,6.301,5.378,2.479,6.217,2.223,4.010,4.371,4.612,4.750,4.165,1.897
In above new test values we don’t have 0 or 1 application will detect and give result.
To implement above concept i design two applications
1)FogComputing: This application will build vector with all possible states or values which may contains normal or faulty data and make that vector available to apply on new values to detect whether new record is normal or faulty. This application accepts input data from process stream and we don’t have any sensors so i am uploading new records from file.
2) ProcessStream: This application will read new input values and then send to Fog Computing application which will apply logic on new records to detect its as normal or faulty.
video file :
thank you for your comment
pls call me on 8125424511