Hybrid and Eddy Current Belt Sensors for Quality Control of Nonferrous Scrap Metals from Municipal Solid Waste Incinerator Bottom Ash

Abstract

The municipal solid waste (MSW) is so complex but a significant secondary source of materials. The incombustible residue after the incineration of the MSW is known as municipal solid waste incinerator (MSWI) bottom ash that is mostly a mix of organic and mainly inorganic materials as well as a significant secondary source of ferrous and nonferrous (NF) metals. However, despite the technological development of eddy current separator (ECS), the recovery of NF contents from MSWI bottom ash, size 1-6mm remains unsatisfactory where the splitter setting of an ECS machine plays crucial role for effective separation and quality control of nonferrous metals and non-metals. For effective separation and quality control of the bottom ash materials the ECS machine needs continuous adjustment of the splitter setting which is quite impractical for a manual operator as a result this thesis primarily addresses this issue by suggesting a sensor based remedy for that. Accordingly this Ph.D. thesis embodied the development of two different kinds of sensors namely hybrid sensor and eddy current belt sensor. The hybrid sensor was developed for the measurement of metal grade (G) of the ECS concentrated bottom ash materials and the measured (G) was used as a qualifier for the quality control of the bottom ash materials. Actually the hybrid sensor produces count data for metal and non-metal particles present in the ECS concentrated bottom ash stream where the hybrid sensor consists of infrared sensor (IRS) for counting all types of particles present in the stream and electromagnetic sensor (EMS) for counting only the metal particles present in the stream. A mathematical model is developed that calculates the metal grade (G) from the sensor count data with the pre-knowledge of average particle mass ratio (k) between non-metal and metal. Consequently this research first focused on design, construction and characterization of the hybrid sensor. Each sensor section is characterized individually in terms of sensitivity, repeatability and accuracy. The hybrid sensor was highly repeatable to its count data and the math model for the measurement of G was verified using the synthetic sample with known values of k i.e. were k =0.24, 0.54, 1.23 and 2.54. The same method was applied for the grade measurement of the ECS concentrated bottom ash materials with an accuracy ±2.4%. After the laboratory characterization a robust set up from the laboratory prototype of the hybrid sensor was built for functionality analyses in situ. The measurements and trends in sensor data from the laboratory and in situ for dry feed materials were quite comparable, considering the ECS machines were different and the bottom ashes came from different sources. The hybrid sensor data predicted quite accurately the trend of the metal grade of the stream of the particles with the splitter distance, which was mandatory for sensor-based control of the ECS splitter position in bottom ash processing. Afterwards this thesis presented an extended part of this sensor research that resulted another fundamental investigation on the development of an eddy current belt sensor. The purpose of the belt sensor was to identify NF scrap metals on a conveyor that could be applied for sensor sorting and quality control of bottom ash materials. The belt sensor relies on a mathematical method which is called in this thesis as conductivity approach. In conductivity approach a parameter CIF (conductivity indication factor) has been defined from where the CIF has been found as truly a function of conductivity. This thesis suggested producing a database of material CIF that was used for the identification of different materials based on conductivity. For experimental validation of the conductivity approach a set of pure sample particles S1 of Cu, Al, and Brass, each of six generic shapes i.e. disk, disk block, square plate, square block, rod, and cylinder were investigated. The test analyses for the sample set S1 showed 100% accuracy for the identification of the Cu, Al and Brass by using their average CIF values. As an application of the eddy current belt sensor another sample set S2 i.e. a representative amount of randomly mixed metal scraps of Cu, Al, Brass and Zn collected from a batch of bottom ash materials was used as a test case for the identification of different metals using their measured CIF values. As a first step towards an application of the belt sensor, the thesis also presented a logical sorting statistics of the bottom ash scraps based on their average CIF values. Moreover, the calculated and calibrated conductivity values of the metal scraps using only the belt sensor were also presented and finally some recommendations have been compiled for further advancement of sensor sorting of waste and quality control of bottom ash materials.