RAW MEALS CONTAINING TRACE RE OXIDES TO IMPROVE QUALITY OF CEMENT CLINKER

Clay or industrial wastes containing trace rare earth (RE) oxides as one of the components are used to burn cement clinker. When the total amount of RE oxides reaches 0. 2×10-4 - 2. 0×10-4 (wt), it has positive effect on the burnability of raw meals? and the Alite content in clinker increases. However, with the addition of lan-thanide, the formation process of clinker and the distribu-tion of lanthanide in clinker are different from those of yt-trium. If the burning temperature is 1 450 C , yttrium shiws negative effect on the formation of clinker. There-fore, to improve the quality of cement clinkern, the raw meals in which yttrium content is lower and lanthanide con-tent is higher is preferable; to reduce the consumption of coal, the raw meals that yttrium content is higher is suit-able. Using raw meals containing trace RE oxides can re-duce the cost of fluorite and protect environment.

Experimental Investigation of Second Interface Cement Bond Evaluation

Cement bond model wells （1：10 scaled-down） were made with a gradually degrading cement annulus for cement bond evaluation of the first interface （between the casing and the cement annulus） and the second interface （between the cement annulus and the formation）. Experimental simulation on cement bond logging was carried out with these model wells. The correlation of acoustic waveforms, casing wave energy and flee casing area before and after cement bonding of the second interface was established. The experimental results showed that the arrival of the casing waves had no relationship with the cement bonding of the second interface, but the amplitude of the casing head wave decreased obviously after the second interface was bonded. So, cement bonding of the second interface had little effect on the evaluation of the cement bond quality of the first interface by using casing head wave arrivals. Strong cement annulus waves with early arrivals were observed before the second interface was bonded, while obvious ＂formation waves＂ instead of cement annulus waves were observed after the second interface was bonded.

Evaluation of the cement bond quality through filtering the reflected ultrasonic waves

High cement bond quality is required to keep an oil well from hydraulic commu- nication between zones. In the cement bond evaluation, the ultrasonic echo method is widely used for its capability of channeling azimuth detection. Full waveforms reflected from the cased hole are simulated for different bonding conditions by the generalized transfer matrix method. Because of the high acoustic impedance of casing, the amplitudes of the reflected waves from the cement and the formation are small and cannot easily be used to evaluate the cementing condition. The wave that can propagate into the cement and the formation through the casing concentrates closely on the casing resonant frequency. To reduce the amplitude of the reflected wave from the casing inner surface and highlight the part of the reflected wave which carries the cementing quality information, the reflected full wave is filtered according to the casing resonant frequency. There are several wave packets in the filtered waveform. When the amplitude of the second wave packet is low, the cement bonds well with the casing, otherwise poorly. A low amplitude third wave packet is an indication of a good bond of the cement with the formation, otherwise poor. To reveal the sensitivity of the reflection wave amplitudes to the incident angle, reflected full waveform is modeled when an acoustic beam with finite width is incident on the casing. It is shown that the bond evaluation method based on filtered wave packets is valid for incident angle less than 5 degrees.

A novel approach to identifying the ultrasonic echoes for inspecting well cementing quality

In this paper, it is presented that an approach based on Continuous Wavelet Thansform(CWT) and Fourier Transform(FT) for identifying ultrasonic echoes in inspecting oil well cementing quality. First, CWT is used to process echo signals. Then spectra of the processing results in some specilic scale-time segments are calculated as pattern features. The research results show that theoretical calculations basically agree with experimental results. Compared with the aPproach based on the spectra or Discrete Wavlet Transform, our approach has advantages of fine analyzing scales, stable features and high recognizing rate. It is very suitable to identify echo signals whose spectra vary with time.