An Analytical Model for Buffer Hit Rate Prediction*
Yongli (Lily) Xi, Patrick Martin and Wendy Powley
Dept. of Computing and Information Science
Queen’s University
Kingston, Ontario
K7L 3N6
Abstract
Of the many tuning parameters available in a database management system (DBMS), one of the most crucial to performance is the buffer pool size. Choosing an appropriate size, however, can be a difficult task. In this paper we present an analytical modeling approach to predicting the buffer pool hit rate that can be used to simplify the process of buffer pool sizing. Since the buffer replacement algorithm determines the buffer hit rate, we model the replacement algorithm which, in the case of DB2/UDB, is a variation of the GCLOCK algorithm. A Markov Chain model of GCLOCK is used to estimate the hit rate for a buffer pool. We evaluate the accuracy of the model’s estimates with experiments carried out on DB2/UDB with the TPC-C benchmark. The model is validated for both single and multiple buffer pool cases.
1Introduction
Today’s database management systems (DBMSs) require database administrators (DBAs) to configure and tune the database to achieve optimal performance. However, since the tasks of configuring and tuning DBMSs are complex and time-consuming, there has been an increasing interest in self-tuning DBMSs, which shift some of the responsibility from the DBA onto the DBMS itself [1][6].
The performance of a DBMS is greatly influenced by the effective use of main memory. A particularly important use of memory is the buffer area, which is a cache for data pages between database applications and the physical database files. It consists of
* This research is supported by IBM Canada Ltd., Communications and Information Technology Ontario (CITO) and the National Science and Engineering Research Council.
one or more buffer pools, and needs a buffer manager to transfer data between buffer pools and physical disk. Efficient use of the buffer area can reduce the number of disk accesses, which improves the system’s performance.
Tuning the size of the buffer pool is therefore crucial to achieving good performance for a DBMS. Several heuristics exist to assist DBAs in this task but often fine tuning becomes a trial and error process. In this case, once the system is configured, the buffer pool hit rate(s) are monitored while the workload is running. If the hit rates are unacceptable, the system is reconfigured and monitored again. This process can be time consuming for a DBA. Our goal is to build an analytical model that can effectively predict buffer pool hit rates.
The buffer hit rate depends on the buffer replacement scheme used by the DBMS. Least Recently Used (LRU) is a popular buffer replacement policy that gives good performance over a wide range of workloads, CLOCK and GCLOCK are also popular policies because of their simplicity and their ability to approximate the performance of LRU [4]. A variation of GCLOCK is used in IBM’s DB2 Universal Database (DB2/UDB) [5].
In this paper, we present a model of the GCLOCK algorithm that is used to predict buffer pool hit rates for DB2/UDB. We extend an approach by Nicola, Dan and Dias [8], which develops an approximate analysis, using a Markov Chain model, for the GCLOCK policy under the Independent Reference Model (IRM). IRM applies to many database transaction processing workloads and assumes that each buffer page access is independent of all previous references. The Markov Chain model of GCLOCK is based
on a partitioning of the database by access frequency. We partition the database using information provided by the DB2/UDB trace facility [5].
The remainder of the paper is organized as follows. Section 2 discusses the GCLOCK algorithm, the development of the Markov Chain Model and its solution. Our method of analyzing DB2/UDB trace data and predicting the buffer pool hit rate is described in Section 3. Section 4 compares our model’s predictions, for both single and multiple buffer pool configurations, with the hit rates achieved by DB2/UDB for the same workload. Section 5 draws conclusions and summarizes the paper.
2Analytical Model
In this section, the main algorithm and relative key issues are discussed. We start by looking at the GCLOCK buffer replacement scheme. Based on this algorithm, we introduce our Markov Chain model, the way we solve it, and the resulting equations used to predict buffer pool hit rates.
2.1 GCLOCK Algorithm
Consider a database consisting of P partitions, such that the access to database pages
p P, and r p be the within a partition is uniform. Let s p be the size of partition p, 1≤≤
probability of each access to a page of partition p. We assume that access to database pages follows the IRM model, that is, each page request is independent of all previous requests.
All pages in the buffer pool are organized in a circular queue and each page has an associated counter. When a page from partition p is first loaded into the buffer its counter is assigned an initial weight, I p. When the system receives a request to access a page it first scans the buffer to find the requested page. If the page is already in the
buffer, that is, there is a buffer page hit, the counter associated with that page is set to the hit weight, H p . If the desired page is not in the buffer and the buffer is full then a buffer page must be selected for replacement. A clock pointer sweeps through the buffer pages,looking to replace the first page with zero count. On a page miss, the clock pointer points to the buffer page immediately following the page that was brought in for the previous miss. If the value of the counter associated with the page is zero then the page is selected for replacement, otherwise the counter is decremented and the next buffer page in sequence is examined. The process goes on until a page with zero count is found and selected for replacement. If the selected page is dirty then it is written out to disk before it is replaced. The new page is read into the buffer location and its counter is set to the initial weight, I p . The clock pointer advances to the buffer page immediately following the page that was replaced. Note that the clock pointer does not move on a buffer pool hit.
We denote the maximum weight of partition p by L p , where L I H p p p =max(,).For simplicity, we assume that hit weight is equal to initial weight, that is, on a hit the counter of the buffer page is reset to the initial weight. In a research report done by Nicola, Dan and Dias, an extensive analysis is done for the case where the initial and the hit weights may not be the same [7].
Suppose the size of the buffer pool is B pages. A buffer page, b i , B i ≤≤1, can be from any database partition p , P p ≤≤1, and will have a counter value in the range from 0 to L p , where L p is the maximum weight for that partition. Since each page in the buffer is a different distance from the clock counter, each page will be in a different state.Thus, if we build an exact Markov model for the GCLOCK policy, it will require
B L p p P
+=∑11 states, which is a very large number. For example, for a buffer size of 1000pages, 4 partitions, and each partition with a maximum weight of 3, an exact model has 1048 states. Therefore, we need to resort to an approximate model. If we do not differentiate between buffer pages, that is, if we ignore the distance difference between buffer pages from the clock pointer, then we can build a simplified Markov model having the number of states equal to ()L p p P
+=∑11[8].
2.2 Markov Chain Model
The objective of the model is to first determine the buffer hit probability h p for each partition, and then to evaluate the overall buffer hit rate. The buffer hit rate for partition p is the probability that pages from partition p that are brought into the buffer at steady state. Let N p be the steady-state average number of pages of partition p in the buffer, and let s p be the size of this partition, in pages. Then the buffer hit probability for this partition is
p p
p s N h =. (1)If we define r p as the probability of accessing the p th partition, the hit probability for the overall database is then
h r h p p p P
==∑1.(2)
2.2.1 Defining the Model
A list of the symbols used in the model is given in Table 1. As stated earlier, our approximate model has ()L p p P
+=∑11 states. The Markov chain for this model represents the states of an arbitrary buffer page at the instant of a random page request. We use (p,
i), where P p ≤≤1, 0≤≤i L p , to represent a state that corresponds to a buffer page of partition p , with a count i . In other words, the steady state probability of being in state (p,i) is the probability that an arbitrary buffer page is of partition p and has a count i . A portion of this Markov chain model, which corresponds to partition p , is shown in Figure
1. The complete Markov chain includes all partitions and, since the partitions are similar,we can extend the analysis for partition p to the Markov chain corresponding to the whole database. The transitions among the parts of the Markov chain (corresponding to different partitions) occur only through the two end states, (p, 0) and (p, L p ), 1≤≤p P ,as shown.System parameters
P Number of data partitions
s p Size of partition p (in pages)
r p Probability of accessing the p th partition
I p Initial weight assigned to partition p
H p Hit weight assigned to partition p
L p max(I p , H p )
B Buffer size for this partition (in pages)
Model variables
h p Hit probability of partition p
m p Miss probability of partition p
h Overall buffer hit probability for this partition
m Overall buffer miss probability for this partition
n p,i Average number of buffer pages that come from partition p
with count i
n 0Average number of buffer pages with count 0
N p Average number of buffer pages that come from partition p
Table 1: List of symbols and their meanings
p /s p
Figure 1: Markov Chain Model for Partition p
Let n p,i be the steady-state average number of buffer pages that come from partition p with count i , 0≤≤i L p . Summing the number of buffer pages of partition p with all possible count values we get N p , which is the average number of buffer pages that all come from partition p . That is
∑==p L i i p p n N 0,, where P
p ≤≤1(3)
We denote n 0 to be the average number of buffer pages with count zero from all partitions at steady state. In a similar manner as with N p , we can derive n 0 by summing all buffer pages from any database partition having zero count. Thus we have
<2>
∑==P p p n n 10
,0(4)
We define a clock cycle as a complete rotation of the clock pointer through all pages in the buffer. From our previous discussion of the GCLOCK model, we can see that at each buffer pool miss, the clock traverses buffer pages until it finds a buffer with count zero. As outlined earlier, when such a buffer miss occurs, the clock points to the buffer page immediately following the last page read from disk. So in a complete clock cycle, the number of buffer misses equals the number of buffer pages with a zero count encountered during that cycle.
2.2.2 Transition Relationships Derivation
We are now ready to discuss the transition relationships between all model states and we justify the probability of transition from one state to another as shown beside the edges in Figure 1. For clarity, we divide the transition relationships into 4 groups, labeled <1>,
<2>, <3> and <4> in Figure 1, and discuss them individually.
In the model, the basic approximation assumes that the average number of misses in a complete cycle of the clock is equal to the average number of buffer pages with zero count at steady state, n 0. This is an approximation because pages with zero count may have been hit and converted to pages with a non-zero count, during the time the clock pointer makes one pass of the buffer. Nicola, Dan and Dias [8] claim that this type of page request is typically a small portion of the total requests. Our experiments also verified that the impact of this assumption is not significant.
Given that the number of misses in a cycle of the buffer is n 0, then the probability that a particular page’s counter is decremented on a miss is 10
n . Therefore, for an
arbitrary page request, the transition probability from state (p, i) to state (p, i-1) is m n 0,0≤≤i L p , where m is the overall buffer miss probability. Note that the probability of a transition out of state (p, 0) due to a miss is also m n 0
, which corresponds to the probability of replacing an arbitrary buffer page having a zero count. This justifies the group <1>transition.
We defined the probability of accessing a partition p to be r p , and the size of partition p to be s p . In other words, there are s p pages in this partition, and they altogether have access probability r p . Thus an arbitrary page of partition p is requested with probability p p
s r . As stated earlier, whenever a page is hit, the desired page has already
been brought into the buffer and its counter is reset to L p . The transition probability from
any state (p, i), 01≤≤?i L p (), to state (p, L p ) is therefore p
p s r . This justifies the group <2> transition.
We now derive the group <3> relationship. For a given buffer page request, the probability that a page will be brought into the buffer is equal to the probability that a page request causes a buffer page miss, that is, m p . Since the probability of accessing the p th partition on a certain request is r p , the probability that a page of partition p is brought into the buffer is r p m p . Also, given that the probability a miss will happen for the whole
database is 10
n , and that once a page is brought in, its count is reset to the maximum weight, the transition probability from each of the states (p, 0), 1≤≤p P , to state (p, L p )
is thus r m n p p
0. Note that a buffer page from any partition with zero count may be replaced by a particular partition page. r m n p p
0 is thus the transition relationship from all other parts
of the Markov chain to the particular partition we are discussing. This justifies group
<3> of transition relationships in the figure, namely the transitions from states (1, 0)through (P, 0) to state (p, L p ).
Likewise, a page from partition p may also be replaced by a page from any other partition. Using the same justification as in group <3> we can easily derive transition probabilities in group <4>, that is, for the set of arrows pointing from state (p, 0) to state (1, L 1) through state (P, L P ), which are 0
n m r i i respectively. (i, L i ) denotes the state of a page which is from partition i with maximum weight L i . This is the probability of a transition out of state (p, 0) to the states related to buffer pages of any partition (including
itself of course) due to a miss. We observe that r m n r m n r m n m n p p 11022000
+++=, which conforms to our earlier justification in group <1> that the transition probability out of state (p, 0) on a buffer page miss is
m n 0
.2.2.3Solving the Model
Given that the buffer size is B , and number of buffer pages from partition p with count i is denoted by n p,i , we define π(p, i), the steady-state probability of a page being in state (p,i) as π(,),p i p i
n B
=
We now write the Kolmogorov’s Forward Equations of the Markov chain,balanced at state (p, i). Referring to the Markov chain model shown in Figure 1, we see that there are two outgoing paths for any inner state (p, i), of the model, 0 p s r respectively. There is also one entering path, with rate m n 0 from state (p, i+1). For some steady state (p, i), the number of buffer pages leaving the state should equal to the number of buffer pages entering the state. The steady state probability of a page being in state (p, i), π(p, i), can be also viewed as the portion of the page’s lifetime that it is in state (p, i). The rate of buffer pages leaving the state can therefore be expressed as ()(,)m n r s p p p i 0+π. Likewise the rate of buffer pages entering state (p, i) can be expressed as m n p i 01π(,) +. Thus we have:()(,)(,)m n r s m n p p p i p i 00 1+=+ππUsing the same idea, we can derive the balance equation for the “end-state” (p,L p ). As shown in Figure 1, the outgoing transition from this state has a rate 0n m . There are two sets of incoming transitions into the state. One set has a rate of 0n m r p p , and the other set has a rate of p p s r . We therefore get the following equation ∑∑?==+=10),( )0,(10),(0p L i i p p p i P i p p Lp p s r n m r n m πππ, where P p ≤≤1Thus, p p p p p L p p p N s r m r n s r n m p +=+,0)(We can then derive the general relationship ,0 ,)1(p i p p i p n s r m n n +=(6)and, m s r m n n r n p L p p p p 1 00 0,)1(++= (7)N p can thus be derived: )) 1(1 1(1 0++?=p L p p p p s r m n s N , where P p ≤≤1(8)Since all pages in buffer must sum to the buffer size B, we have B N P p p =∑=1 (9)We finally get a nonlinear equation, )) 1(1 1(11 0=?+? ∑=+B s r m n s P p L p p p p (10) which is the key equation in our approach to estimating buffer hit rate. We note that the quantities m and n 0 always appear together as the ratio m n 0, and so can be looked on as one variable. Equation (10) can be solved iteratively for the ratio m n 0 using the bisection method [2], which according to our experiments, converges very quickly. Once the ratio m n 0 is determined, we can evaluate N p , P p ≤≤1, from equation (8), and h p , the buffer hit probability for each partition from equation (1). Finally we can compute the overall hit probability for the whole database, using equation (2).3 Method to Predict Hit Rates From the derivation of the Markov model, we can see that, given a certain buffer size, we need the following input values to estimate the hit rate for a specified buffer area: P Number of data partitions s p Size of partition p (in pages)r p Probability of accessing the p th partition L p maximum weight, which is max(I p , H p ) I p Initial weight assigned to partition p H p Hit weight assigned to partition p According to our assumptions, L p = I p = H p , so the last three inputs can be looked on as one parameter. 3.1Partitioning Criteria and Approach We first need to partition the database such that the access frequency to database pages within a partition is uniform. There are various ways to define "uniform". For example,it can mean the same access frequencies or the same access patterns. Dan [3] distinguishes three types of access patterns, namely, 1) locality within a transaction, 2) random accesses by transactions and 3) sequential accesses by long queries. Our experimental workload is generated by the TPC-C OLTP benchmark. OLTP workloads in general, and TPC-C specifically, have few sequential accesses. This means that partitioning based purely on access pattern is not suitable. It is also the case that, for multiple buffers, one criteria often used to configure database objects to buffer pools is the access pattern for those objects. A buffer pool may therefore only reflect one type of access pattern. OLTP workloads typically have skewed (non-uniform) data access, that is, we can identify a set of "hot" pages that receive a greater portion of the accesses. This property of OLTP workloads can be adequately modeled by partitions based on access frequency [3]. 3.2Disk/Memory Reference Trace In order to collect detailed information on page access frequencies, we trace each page access by a transaction for a sample run of the workload. OLTP workloads are almost always cyclic in nature [3], so a large enough sample will adequately reflect the complete set of pages accessed by that workload. Since the buffer area sits in the middle of the memory hierarchy, all table accesses have to go through buffer pool. Tracing the buffer pool references thus provides the required information on database references. We need to trace every read and write access to the buffer pools and to record the physical page number and database object accessed. The DB2/UDB trace facility provides this information. For each buffer pool reference, whether it is a read or a write, we can trace two activities. The system issues a FIX to fix a page before an access, and an UNFIX after the access is completed. From the trace we obtain the object page number for each page accessed which is the address of the page relative to the start of the object. 3.3Database Partitioning In our experiments, we observed that the accesses were concentrated on a small number of database pages, and a relatively large part of the database pages get only a few access requests. In a typical experimental run, 3% of the distinct pages account for 80% of the total database accesses. On the other hand, over half of the database pages only get around 3% of the accesses. It is thus straightforward to group database pages based on access frequencies. We divided the database into four partitions based on access frequency. We call these partitions "hot"-- most frequently accessed pages, "cold" -- rarely accessed pages, "warm" -- pages whose access frequency falls in between and “freezing” – unused database pages. Since the pages in the freezing partition have a 0% probability of being accessed, we ignore these pages in our analysis and only consider the other three partitions. Three types of counters are maintained. For the single buffer pool case, we count the distinct pages the system has accessed during the testing period, that is, the number of distinct pages (S) that have appeared in the trace file. We also count the number of accesses associated with each distinct page (a i), and the number of overall accesses (A). After partitioning, we can add up the number of distinct pages in each partition to get s p the size of each partition. We can also get the number of overall accesses in one partition by using ∑==p s i i p a a 1We can then calculate the overall access probabilities for each part as A a r p p = Note that ∑==P p p S s 1 , remembering P is the number of the partitions, and S is the overall database size. The maximum weight for each database partition L p , is assigned the value 2 since DB2/UDB uses weight values of 0, 1, or 2 to indicate the locality of a page. For the multiple buffer pool case we treat each buffer pool and its associated database objects as an individual small database. In this way, we split the multiple buffer pool case into several single buffer pool problems. We collect the necessary model inputs for, and apply our analytical model to, each of them. Experiments for both the single and multiple buffer pool cases are presented in Section 4. 4 Experiments All the experiments were run with DB2/UDB Version 7.1 under Windows NT on an IBM Netfinity 5000. The machine was configured with one 400 MHz processor, 1 GB of RAM and 5 disks. The workload used in the experiments is the TPC-C benchmark,which typifies an OLTP workload [9]. The database schema from the TPC-C benchmark is composed of nine relations. TPC-C simulates the activities of a wholesale supplier and includes 5 order-entry type transactions. The database we use has 10 data warehouses,each of which is approximately 100MB bytes in size. Thus the database has about 250,000 4K pages. In our experiments, 40 clients, or simulated “terminal operators” issue the transactions against the database. The relative frequencies of each transaction are as specified in the benchmark. 4.1System Performance Monitoring In order to validate and evaluate our buffer hit rate estimates, we collect real buffer hit rates under different buffer configurations to compare with our model estimation. For each configuration, we let the TPC-C driver run for 5 minutes to allow the system to stabilize. Each time we run the TPC-C driver for a specified configuration, we collect "snapshot" data five times. The buffer hit rate for each snapshot is calculated, and the average of the snapshots is used to get the average hit rate for a run. Each snapshot period is 8 minutes long. We use the mean from each configuration as the final system figure in our model comparison studies. A statistical analysis of the results reveals that the system performance is quite stable. A confidence level of 95% is used in the analysis. Each of our trace files contains about 12,000 trace records. Since we need a trace file as long as possible to ensure reliable and complete data, we collect a long series of trace files, and then concatenate them into a single file of over 600 minutes worth transactions and a size of 136 MB. Our experiments showed that the number of distinct pages accessed from the 136-MB trace file is not significantly more than that from a 60-MB trace file. Although the first file is almost twice the size of the second, it contains less than 50 more distinct pages than the smaller file, compared to the 3438 pages in total. This means that the access to the database is very concentrated on a part of the pages in TPC-C activities and that the workload is cyclic as assumed. Our analysis of the trace file showed that there are 3438 distinct pages accessed, and the total number of accesses is 183770. When studying the distribution of access frequencies, we see that 80% of the total page accesses are concentrated on roughly 2% of the distinct pages accessed. Over half of all distinct pages are accessed less than 4 times. We can easily distinguish the "hot" and "cold" partitions, and the pages in between can be treated as the "warm" partition. We tried various partitioning schemes on the accessed pages, which are shown in Tables 2 through 4, to evaluate the impact of the partitioning criteria on the performance of our approach. Partition Size (pages)Percentage of database Number of Total accesses Access probability Hot (>250 requests)820.023*********.8050 Warm (5~250 requests) 14880.43282300880.16377 Cold (<5 requests)18680.5433357390.03123 Table 2: Partitioning Plan 1 Partition Size (pages)Percentage of database Number of total accesses Access probability Hot (>250 requests)820.023*********.8050 Warm (4~250 requests) 25120.73066341840.18606 Cold (<4 requests)8440.2454916430.00894 Table 3: Partitioning Plan 2 Partition Size (pages)Percentage of database Number of total accesses Access probability Hot (>150 requests)1000.029*********.82474 Warm (4~150 requests) 24940.72542305650.16632 Cold (<4 requests)8440.2454916430.00894 Table 4: Partitioning Plan 3 When partitioning the database, we attempted to find the smallest partition that was most frequently accessed. In this way, the hottest partition, which is relatively small in size, is almost guaranteed to remain buffered in the memory given that the buffer is large enough. We found experimentally that the estimates produced from the first partitioning plan were closest to the system performance. Therefore we use partitioning plan 1 as the initial input to our analytical model. 4.2 Single Buffer Pool Experiment Figure 2: Single Buffer Pool Hit Rate Hit rate values collected from the real system and estimated from the model for a single buffer pool of 400 to 2400 4K pages are shown in Figure 2. We can see that the two curves almost coincide with each other, except for small differences when the buffer size is very large compared to the total amount of the database that is accessed. However, our comparison result shows that the difference between our model result and the real system performance is not significant. On average, the estimated error is around 2%. Thus we can conclude that our analytical model provides a good approximation for DB2/UDB’s buffer hit rate for an OLTP workload and a single buffer pool configuration. 4.3 Multiple Buffer Pools Experiment For the multiple buffer pool experiments, we evaluated our model with two different buffer pool configurations -- one with three buffer pools and one with four buffer pools. We present the results of the experiments using three buffer pools here. The four buffer pool case, which achieved similar results, is described elsewhere [10]. To apply our model to the multiple buffer pool case, we split the problem into N disjoint single buffer pool problems, where N is the number of buffer pools. After obtaining the mapping between database objects and buffer pools, we treat each buffer pool as an individual unit and estimate the buffer hit rate. We assign each database object to its own tablespace, so that each database object can be distinct in the trace data. The tablespaces are assigned to buffer pools as follows: ?BP0 – all data tables ?BP1 – indices for the new_order, orders and history tables ?BP2 – indices for the item, stock, customer and order_line tables In the experiment, we vary the size of one buffer pool while holding the sizes of the other two constant. We collect data for five snapshots for each buffer pool size, and take the mean as the real hit rate. We vary each of the buffer pools in turn. The hit rate comparison for each of the buffer pools are shown in Figures 3 through 5. We can see that the estimated hit rate curves are very close to the real system performance curves, especially for buffer pool BP1 and BP2 where index tables are stored. As in the single buffer pool case, we see a slight difference when BP0 is large. In this case, the hit rate estimated from our analytical model is higher than the real system 道里区2019-2020上学期九年级综合期末测试题 可能用到的相对原子质量: H-l C-12 O-16 Fe- 56 Cl-35.5 S-32 Cu-64 K-39 一、选择题（1—27小题，每小题2分，共54分，每小题只有一个正确答案） 1．2020年是我们哈尔滨人翘首期盼的一年。因为将有209条公交线路纵横贯通；3条地铁线形成十字加环路；哈站重现百年经典；国际机场规模再一次扩大。届时市内交通会非常便捷，全国各地乃至全世界与哈尔滨的“距离”再一次缩短。下列有关说法不正确的是（） A．交通的便利极大的方便了哈尔滨人民的出行 B．制造地铁轨道的玻璃钢属于金属材料 C．火车站建造过程中所用到的钢筋混凝土属于复合材料 D．飞机所用的燃料——航空煤油是石油分馏的产物 2．下列实验操作错误的是( ) A．读取液体体积B．连接导管和胶塞C．熄灭酒精灯D．清洗试管3．下列过程中主要发生物理变化的是（） A．生产沼气B．炼钢C．高炉炼铁D．干冰升华4．下列物质用途正确的是（） A．用明矾除水中异味B．液氧冷冻麻醉C．硫酸用于汽车铅蓄电池D．用铝制造人造骨 5.日常生活中的下列做法正确的是（） A. 我国的纯碱资源比较丰富，用纯碱来制取大量食盐 B. 由汽油、柴油燃烧引起的失火可用水基型灭火器灭火 C. 用厨房清洁剂除热水壶中的水垢 D. 用亚硝酸钠腌制肉、蛋以延长食品的保质期 6．下列实验现象描述错误的是（) A．打开浓盐酸的瓶塞：有较强烈的刺激性气味，瓶口有白雾 B．向氢氧化钠溶液中滴加氯化铁溶液:溶液中产生红褐色沉淀 C．铁丝在氧气中燃烧:银白色固体剧烈燃烧，火星四射，放热，生成黑色固体 相平衡自测题 1 在含有C(s)、H2O(g)、CO(g)、CO2(g)、H2(g)五个物种的平衡体系中,其独立组分数C为( ) (a) 3 (b) 2 (c) 1 (d) 4 2二元合金处于低共熔温度时的物系的自由度f为( ) (a) 0 (b) 1 (c) 2 (d) 3 3 298K时蔗糖水溶液与纯水达渗透平衡时,整个体系的组分数、相数、自由度数为( ) (a)C=2 Ф=2 f*=1 (b) C=2 Ф=2 f*=2 (c) C=2 Ф=1 f*=2 (d)C=2 Ф=1 f*=3 3FeCl3和H2O能形成FeCl3?6H2O、2FeCl3?7H2O、 2FeCl3?5H2O、FeCl3?2H2O四种水合物, 该体系的独立组分数C和在恒压下最多可能的平衡共存相数Ф分别为( ) (a)K=3,Ф=4 (b)K=2,Ф=4 (c)K=2,Ф=3 (d)K=3,Ф=5 5 对于恒沸混合物,下列说法错误的是( ) (a)不具有确定的组成 (b)平衡时气相和液相的组成相同 (c)其沸点随外压的改变而改变 (d)与化合物一样具有确定组成 6 单组分体系的固液平衡线的斜率dp/dT的值( ) (a) 大于零 (b) 等于零 (c) 小于零 (d) 不确定 7 A、B两液体混合物在T-x图上出现最高点,则该混合物对拉乌尔定律产生( ) (a)正偏差 (b) 负偏差 (c) 没偏差 (d)无规则 8 下列过程中适用于克-克方程的是( ) (a) I2 (s)=I2 (g) (b) C(石墨)=C(金刚石) (c) Hg2Cl2(s)=2HgCl(g) (d) N2 (g,T1,p1)=N2 (g,T2,p2) 9 某一固体在25℃和101325Pa压力下升华,则意味着( ) 哈尔滨工业大学（威海） 2011 / 2012 学年 秋 季学期 理论力学 试题卷（A ） 考试形式（开、闭卷）：闭卷 答题时间：120（分钟）本卷面成绩占课程成绩 80 % 题号 一 二 三 四 五 卷 面 总 分 平 时 成 绩 课 程 总 成 绩 分数 一、填空题（每题6 分，共30分） 1、图示滚轮的半径为R ，轮心速度为v ，加速度为a ，沿水平直线作纯滚动，M 点加速度的大小为 。 2、图示匀质细杆各以匀角速度ω绕轴Ｏ转动。已知：杆长为，重为l P 。惯性力系向转轴Ｏ简化，主矢的大小为 ，主矩的大小为 。 3、如图所示，物块重P =150N ，靠在粗糙的水平天花板上，物体间静滑 动摩擦因数f s =0.3， 动滑动摩擦因数f =0.2, 在物块上加一力F ，且F = 200N ,θ=30°，摩擦力大小为 。 得分 教研室主任签字： 姓名: 班级： 学号 遵 守 考 试 纪 律 注 意 行 为 题，叉时，应采用金属隔板进行隔开处理；同一线槽内强电回路须同时切断习题要进行检查和检测处理。作并且问题，作为调试人员，需要在事前掌握图纸资料、设备制造厂家出具高中资料术资料，并且了解现场设备高中资料试卷布置情况与有关高中资料试卷电气系统接线等情况，然后根据规范与规程规定，制定设备调试高中资料试卷方案。 调试调试技术是指发电机一变压器组在发生内部故障时，需要进行外部电源高中资中资料试卷主要保护装置。 4、如图所示结构受集中力P作用，各杆自重不计。则杆1的内力为。 （须注明拉、压） 5、如图所示，沿长方体的三个不相交且不平行的棱作用着3个大小相等 的力F，若力系简化为一合力，求棱a,b,c的关系。 二、计算题（10 分） 如图所示简支梁，不计梁的重量，在C处作用一铅垂力P。试求支座B处的约束力（要求用虚位移原理求解，用其它方法做不给分）。 得分 遵 守 考 试 纪 律 注 意 行 为 规 范 回 路 交 叉 时 ， 应 采 用 金 属 隔 板 进 行 隔 开 处 理 ； 同 一 线 槽 内 强 电 回 路 须 同 时 试 卷 技 术 问 题 ， 作 为 调 试 人 员 ， 需 要 在 事 前 掌 握 图 纸 资 料 、 设 备 制 造 厂 家 出 具 料 试 卷 调 试 技 术 是 指 发 电 机 一 变 压 器 组 在 发 生 内 部 故 障 时 ， 需 要 进 行 外 部 电 xx学校xx学年xx学期xx试卷 姓名:_____________ 年级:____________ 学号:______________ 题型选择题填空题简答题xx题xx题xx题总分 得分 一、xx题 （每空xx 分，共xx分） 试题1: 如图所示是哈市某中学九年级物理课外实践活动小组设计的风力测量仪的电路示意图．其电源电压U=3V，电路保护电阻R1=20Ω，均匀电阻丝的阻值为R2=3Ω，金属丝（电阻忽略不计）的上端固定在O点，下端系有一块塑料片A，当风从右向左吹动塑料片时，金属丝向左摆动并与电阻丝保持良好的接触，风力表由电压表改装而成． （1）无风时，塑料片竖直下垂，电压表示数是多少？电路中的电流是多少？（结果保留两位小数） （2）风力最强时，金属丝滑至电阻丝的最左端，电压表的示数是多少？ （3）为什么不在电路中串联电流表，把电流表改装成风力表？ 试题2: 小明看到滑动变阻器的铭牌模糊不清，于是想测量滑动变阻器的最大阻值R x．他从实验室找来的器材有：电压未知的电源、量程合适的电压表、已知阻值为R0的定值电阻R0、铭牌不清的滑动变阻器、开关各一个、导线若干． （1）他已经连好电路的一部分，如图甲所示，在不改动电压表位置及其连线的情况下，请帮助他把电路连接完整，并在图乙中画出电路图． （2）简述实验过程中要测量的物理量并用相应的字母表示． ①； 评卷人得分 ②； ③R X= （用已知量和测量量表示）． （3）请你写出两条在此实验操作中应该注意的事项． 试题3: 实验小组利用下面的电路图探究“电流跟电压、电阻关系”的实验中： 表一： 电压U/v 1.0 1.5 2.5 8 3.0 电流I/A 0.2 0.3 0.4 0.5 0.6 （1）实验小组收集了表一的实验数据，能得出的结论是． （2）小组同学在继续实验时，发现电压表已坏掉，于是他们不用电压表，又收集到了表二的实验数据： 表二： 电阻/Ω 5 10 20 30 40 电流/A 0.8 0.42 0.23 0.17 0.14 你认为他们的这次探究有什么问题？ 试题4: 小彤和小雪在“探究并联电路中干路电流与各支路电流有什么关系”时，利用一个开关、一个电流表、一个学生电源（有多个电压档位）四个阻值不等的电灯以及若干条导线，进行了大胆地研究．如图所示是他们的实验电路图． 哈尔滨工业大学2009至2010学年第一学期物理光学期末考试试题 一、填空题（每小题2分，总共20分） 1、测量不透明电介质折射率的一种方法是，用一束自然光从真空入射电介质表面，当反射光为（）时，测得此时的反射角为600，则电介质的折射率为（）。 2、若光波垂直入射到折射率为n=1.33的深水，计算在水表面处的反射光和入射光强度之比为（）。 3、光的相干性分为（）相干性和（）相干性，它们分别用 （）和（）来描述。 4、当两束相干波的振幅之比是4和0.2时，干涉条纹对比度分别是（）、和（）。 5、迈克尔逊干涉仪的可动反光镜移动了0.310mm，干涉条纹移动了1250条，则所用的单色光的波长为（）。 6、在夫朗禾费单缝衍射实验中，以波长为589nm的钠黄光垂直入射，若缝宽为0.1mm，则第一极小出现在（）弧度的方向上。 7、欲使双缝弗琅禾费衍射的中央峰内恰好含有11条干涉亮纹，则缝宽和缝间距需要满足的条件是（）。 8、一长度为10cm、每厘米有2000线的平面衍射光栅，在第一级光谱中，在波长500nm附近，能分辨出来的两谱线波长差至少应是（）nm。 9、一闪耀光栅刻线数为100条/毫米，用l=600nm的单色平行光垂直入射到光栅平面，若第2级光谱闪耀，闪耀角应为多大（）。 10、在两个共轴平行放置的透射方向正交的理想偏振片之间，再等分地插入一个理想的偏振片，若入射到该系统的平行自然光强为I0，则该系统的透射光强为（）。 二、简答题（每小题4分，总共40分） 1、写出在yOz平面内沿与y轴成q角的r方向传播的平面波的复振幅。 2、在杨氏双缝干涉的双缝后面分别放置n1=1.4和n2=1.7，但厚度同为d的玻璃片后，原来 的中央极大所在点被第5级亮条纹占据。设l=480nm，求玻璃片的厚度d及条纹迁移的方向。 3、已知F-P标准具的空气间隔h=4cm，两镜面的反射率均为89.1%；另一反射光栅的刻线面 积为3′3cm2，光栅常数为1200条/毫米，取其一级光谱，试比较这两个分光元件对 l=632.8nm红光的分辨本领。 4、平行的白光（波长范围为390-700nm）垂直照射到平行的双缝上，双缝相距1mm，用一个 焦距f=1m的透镜将双缝的衍射图样聚焦在屏幕上。若在屏幕上距中央白色条纹3mm处开一个小孔，在该处检查透过小孔的光，则将缺少哪些波长？ 5、一块每毫米500条缝的光栅，用钠黄光正入射，观察衍射光谱。钠黄光包含两条谱线， 其波长分别为589.6nm和589.0nm。求在第二级光谱中这两条谱线互相分离的角度。6、若菲涅耳波带片的前10个奇数半波带被遮住，其余都开放，则求中心轴上相应衍射场点 的光强与自由传播时此处光强的比值。 7、一束汞绿光以600入射到KDP晶体表面，晶体的n o=1.512，n e=1.470。设光轴与晶体表面 平行，并垂直于入射面，求晶体中o光和e光的夹角。 8、画出沃拉斯顿棱镜中双折射光线的传播方向和振动方向。（设晶体为负单轴晶体） 大学物理期末考题(A) 2003年1月10日 得分__________ 班级_________姓名_________学号___________ 序号____________ 注意：（1）共三张试卷。（2）填空题★空白处写上关键式子，可参考给分。计算题要排出必要的方程，解题的关键步骤，这都是得分和扣分的依据。（3）不要将订书钉拆掉。(4)第4、5页是草稿纸。 一、选择题 1、在宽度a =0.05mm 的狭缝后置一焦距f 为0.8m 的透镜， 有一屏幕处在透镜的焦平面上，如图所示。现将某单色光垂直照射在单缝上，在屏幕上形成单缝衍射条纹，试问：若在离中央明条纹上方x =1.6cm 的P 处恰为暗条纹，则该光的波长约为 (a) 450nm (b) 500nm (c) 550nm (d) 600nm _____________ 1、在宽度a =0.05mm 的狭缝后置一焦距f 为0.8m 的透镜，有一屏幕处在透镜的焦平面上，如图所示。现将某单色光垂直照射在单缝上，在屏幕上形成单缝衍射条纹，试问：若在离中央明条纹上方x =1.6cm 的P 处恰为暗条纹，则该光的波长约为 (a) 450nm (b) 500nm (c) 550nm (d) 600nm 选_____B ______ λ θθk a f x ==sin kf ax = ?λ 2、在牛顿环实验中，观察到的牛顿环的干涉圆环形条纹第9级明条纹所占的面积与第16级明条纹所占的面积之比约为 (a) 9/16 (b) 3/4 (c) 1/1 (d) 4/3 (e) 16/9 选_____________ 2、在牛顿环实验中，观察到的牛顿环的干涉圆环形条纹第9级明条纹所占的面积与第16级明条纹所占的面积之比约为 (a) 9/16 (b) 3/4 (c) 1/1 (d) 4/3 (e) 16/9 选_____C ______ 明：2 ) 12(λ -= k R r , 暗：λRk r = , λπR S S k k =-+1 3、用频率为ν的单色光照射某金属时，逸出光电子的动能为k E ，若改用频率 2ν的单色光照射该金属时，则逸出光电子的动能为 （a ）k E 2 (b) k E h -ν (c) k E h +ν (d) k E h -ν2 选_____________ 静力学知识点 静力学公理和物体的受力分析 本章总结 １．静力学是研究物体在力系作用下的平衡条件的科学。 2.静力学公理?公理1 力的平行四边形法则。 公理２二力平衡条件。?公理3 加减平衡力系原理?公理４作用和反作用定律。 公理5 刚化原理。 3.约束和约束力 限制非自由体某些位移的周围物体,称为约束。约束对非自由体施加的力称为约束力。约束力的方向与该约束所能阻碍的位移方向相反。 4．物体的受力分析和受力图?画物体受力图时，首先要明确研究对象（即取分离体）。物体受的力分为主动力和约束力。要注意分清内力与外力,在受力图上一般只画研究对象所受的外力;还要注意作用力和反作用力之间的相互关系。 常见问题 问题一画受力图时，严格按约束性质画，不要凭主观想象与臆测。 平面力系 本章总结 1. 平面汇交力系的合力 （１)几何法：根据力多边形法则，合力矢为 合力作用线通过汇交点。 ( 2 ）解析法：合力的解析表达式为 ２. 平面汇交力系的平衡条件 （1)平衡的必要和充分条件: （2)平衡的几何条件：平面汇交力系的力多边形自行封闭。?（３)平衡的解析条件（平衡方程）: ３. 平面内的力对点Ｏ之矩是代数量,记为 一般以逆时针转向为正，反之为负。 或 4．力偶和力偶矩 力偶是由等值、反向、不共线的两个平行力组成的特殊力系。力偶没有合力，也不能用一个力来平衡。 平面力偶对物体的作用效应决定于力偶矩Ｍ的大小和转向，即 式中正负号表示力偶的转向，一般以逆时针转向为正,反之为负。 力偶对平面内任一点的矩等于力偶矩，力偶矩与矩心的位置无关。 ５. 同平面内力偶的等效定理：在同平面内的两个力偶,如果力偶相等,则彼此等效。力偶矩是平面力偶作用的唯一度量。 ６. 平面力偶系的合成与平衡 合力偶矩等于各分力偶矩的代数和,即 平面力偶系的平衡条件为 ７、平面任意力系 平面任意力系是力的作用线可杂乱无章分布但在同一平面内的力系。当物体(含物体系)有一几何对称平面，且力的分别关于此平面对称时，可简化为平面力系计算。还有其他情况也可按平面任意力系计算。 本章用力的平移定理对平面任意力系进行简化,得到主矢主矩的概念，并进一步对力系简化结果进行讨论；然后得出平面任意力系的平衡条件，得出平衡方程的三种形式,并用平衡方程求解一些平衡问题；介绍静定超静定问题的概念,对物体系的平衡问题进行比较多的训练；最后介绍平面简单桁架的概念和内力计算。 常见问题 问题一不要因为这一章的内容简单，就认为理论力学容易学，而造成轻视理论力学的印象，这将给后面的学习带来影响。 问题二本章一开始要掌握好单个物体的平衡问题与解题技巧,这样才能熟练掌握物体系平衡问题的解法与解题技巧。 问题三在平时做题时，要注意解题技巧的训练，能用一个方程求解的就不用两个方程，但考试时则不一定如此。 第三章空间力系 本章总结 1.力在空间直角坐标轴上的投影 （1）直接投影法 （2)间接投影法（图形见课本） ２．力矩的计算 （1 ）力对点的矩是一个定位矢量, （２）力对轴的矩是一个代数量,可按下列两种方法求得： ( ａ) ( b ） ( 3 )力对点的矩与力对通过该点的轴的矩的关系 哈尔滨工业大学（威海） 2012/2013 学年秋季学期 大学物理试题卷（A） 考试形式（开、闭卷）：闭卷答题时间：120 （分钟）本卷面成绩占课程成绩 70 % 题 号一二三四五六七八卷面 总分 平时 成绩 课程 总成绩 分 数 一、选择题（每题 2 分，共18 分） 1. 一质点作简谐振动，周期为T．当它由平衡位置向x轴正方向运动时，从二分之一最大位移处到最大位移处这段路程所需要的时间为［］ (A) T /12． (B) T /8． (C) T /6． (D) T /4． 2. 一平面简谐波在弹性媒质中传播，在某一瞬时，媒质中某质元正处于平衡位置，此时它的能量是［］ (A) 动能为零，势能最大． (B) 动能为零，势能为零． (C) 动能最大，势能最大． (D) 动能最大，势能为零． 3. 用波长为的单色光进行双缝干涉实验，若用薄玻璃板遮住双缝中的一个缝，已知玻璃板中的光程比相同厚度的空气的光程大 3.5 ，则屏上原来的暗条纹处［］ (A) 变为明条纹； (B) 仍为暗条纹； (C) 既非明纹也非暗纹； (D) 无法确定是明纹，还是暗纹． 4.使单色光垂直入射到双缝光栅上观察光栅衍射图样，发现在其夫琅禾费衍射包线的中央极大宽度内恰好有9条干涉明条纹，则光栅常数d和缝宽a的关系是［］ (A) d=3a． (B) d=4a． (C) d=5a． (D) d=6a. 得分 5．一定频率的单色光照射在某种金属上，测出其光电流的曲线如图中实线所示．然后在光强度不变的条件下增大照射光的频 率，测出其光电流的曲线如图虚线所示．满 足题意的图是：［ ］ 6. 关于不确定关系η≥??x p x ()2/(π=h η，下面的几种理解正确的是［ ］。 (1) 粒子的动量不可能确定． (2) 粒子的坐标不可能确定． (3) 粒子的动量和坐标不可能同时准确地确定． (4) 不确定关系不仅适用于电子和光子，也适用于其它粒子． (A) (1)，(2). (B) (2)，(4). (C) (3)，(4). (D) (4)，(1). 7. 一定量的理想气体贮于某一容器中，温度为T ，气体分子的质量为m ．根据理想气体 的分子模型和统计假设，分子速度在x 方向的分量平方的平均值为 (A) m kT x 32 = v ． (B) m kT x 3312 =v ． (C) m kT x /32=v , (D) m kT x /2 =v ． ［ ］ 8. 速率分布函数f (v)的物理意义为： (A) 具有速率v 的分子占总分子数的百分比． (B) 速率分布在v 附近的单位速率间隔中的分子数占总分子数的百分比． (C) 具有速率v 的分子数． (D)速率分布在v 附近的单位速率间隔中的分子数． ［ ］ 9. 所列四图分别表示理想气体的四个设想的循环过程．请选出其中一个在物理上可能实现的循环过程的图的标号． ［ ］ p V p V p V p V 2013年哈尔滨市中考物理模拟测试卷 一、选择题：（每小题2分，共24分,每小题只有一个正确答案） 16．下列关于测量仪器使用方法的叙述，不正确的是( ) A．测物理课本长度时，应使刻度尺沿着课本摆放 B．测电流时，应使电流表与待测的那部分电路串联 C．测水溶液体积时，视线应与量筒中凹液面最低处相平 D．测液体温度时，温度计的玻璃泡应完全浸没在被测液体中 17．针对下列各图，说法正确的是（） A．甲图中，用力拨动松紧相同的橡皮筋，橡皮筋越细音调越低 B．乙图中，敲鼓时用力越大，所发声音的音调越高 C．丙图中，随着向外不断抽气，手机铃声响度越来越小，音调越来越低 D．丁图中，尺子伸出桌面的长度越长，拨动时尺子振动越慢，音调越低 18．如图所示，寒冷的冬天，房间门窗玻璃的内表面往往会凝结一层美丽的冰花，关于冰花下列说法正确的是（） A．冰花是晶体，是由水蒸气凝固形成的 B.冰花是水蒸气凝华形成的，形成过程中要不断的吸热 C.冰花形成在玻璃窗的外侧，并且形成过程中要不断放热 D.冰花是室内的水蒸气凝华在玻璃窗的内侧，形成过程在不断的放热 19．用图像表示一个物理量随另一个物理量的变化过程，可使物理规律更直观、形象。如图所示，关于此图所表示的物理规律，下列分析错误的是（） A．物体所受重力与质量的关系 B.液体压强与深度的关系 C.做匀速直线运动的物体，速度与时间的关系 D.通过定值电阻的电流与电压的关系 20．如图是某同学自制的照相机。关于该照相机，下面 是几位同学在使用后的感悟和收获，其中正确的是（） A．工作原理是小孔成像 B.在半透明纸上不但能成倒立的像，还能成正立的像 C.观察的景物由远至近过程中，应适当调近凸透镜与半透明纸间的距离 D.凸透镜上有一点污渍，是不影响在半透明纸上成完整的像的 21．利用如图甲所示的“坐位体前屈测试仪”对初中毕业生进行了身体柔韧性测试。测试者向前推动滑块，滑块被推动的距离越大，仪器的示数就越大。若琦同学设计了四种电路(如图乙所示)模拟测试，并要求电路中滑动变阻器的滑片向右滑动时，电表示数增大。其中符合要求的电路是（） 22.如图所示，是三国时，曹冲利用浮力测量大象的体重。聪明的曹冲把大象赶到船上，记下船的吃水深度，再用许多石块代替大象，使船有相同的吃水深度，称出石块的总重，就是大象的体重。对曹冲的测量，班级同 学有以下几点认识，其中不正确的是（） A．曹冲测量大象重量的过程体现了等效替代的科学方法 B.根据浮力知识可知：大象的重力、石块的总重及船排 开水的总重三者相等 C.大象对船的压力与船对大象的支持力是一对相互作用力 D.船受到的浮力与船受到的重力及大象对船的压力平衡 哈工大 2001 年春季学期 结构力学试卷 (请考生注意:本试卷共5页 一.是非题(将判断结果填入括弧:以O 表示正确,X 表示错误(本大题分4小题,共11分 1 . (本小题 3分 图示结构中DE 杆的轴力F NDE =F P /3。( . 2 . (本小题 4分 用力法解超静定结构时,只能采用多余约束力作为基本未知量。 ( 3 . (本小题 2分 力矩分配中的传递系数等于传递弯矩与分配弯矩之比,它与外因无关。( 4 . (本小题 2分 用位移法解超静定结构时,基本结构超静定次数一定比原结构高。 ( 二.选择题(将选中答案的字母填入括弧内(本大题分5小题,共 21分 1 (本小题6分 图示结构EI=常数,截面A 右侧的弯矩为:( A .2/M ; B .M ; C .0; D. 2/(EI M 。 2. (本小题4分 2 图示桁架下弦承载,下面画出的杆件内力影响线,此杆件是:( A .ch ; B.ci; C.dj; D .cj . 3. (本小题 4分 图a 结构的最后弯矩图为: A. 图b; B. 图c; C. 图d; D.都不对。( ( a (b (c (d 4. (本小题 4分用图乘法求位移的必要条件之一是: A.单位荷载下的弯矩图为一直线; B.结构可分为等截面直杆段; C.所有杆件EI 为常数且相同; D.结构必须是静定的。 ( 5. (本小题3分 图示梁A 点的竖向位移为(向下为正:( A.F P l 3/(24EI ; B . F P l 3/(!6EI ; C . 5F P l 3/(96EI ; D. 5F P l 3/(48EI . 三(本大题 5分对图示体系进行几何组成分析。 一、是非题 1、力有两种作用效果，即力可以使物体的运动状态发生变化，也可以使物体发生变形。 （√） 2、在理论力学中只研究力的外效应。（√） 3、两端用光滑铰链连接的构件是二力构件。（×） 4、作用在一个刚体上的任意两个力成平衡的必要与充分条件是：两个力的作用线相同， 大小相等，方向相反。（√） 5、作用于刚体的力可沿其作用线移动而不改变其对刚体的运动效应。（×） 6、三力平衡定理指出：三力汇交于一点，则这三个力必然互相平衡。（×） 7、平面汇交力系平衡时，力多边形各力应首尾相接，但在作图时力的顺序可以不同。 （√） 8、约束力的方向总是与约束所能阻止的被约束物体的运动方向一致的。（×） 9、在有摩擦的情况下，全约束力与法向约束力之间的（应是最大)夹角称为摩擦角。（×） 10、用解析法求平面汇交力系的平衡问题时，所建立的坐标系x，y轴一定要相互垂直。 （×） 11、一空间任意力系，若各力的作用线均平行于某一固定平面，则其独立的平衡方程最多只有3个。 （×） 12、静摩擦因数等于摩擦角的正切值。（√） 13、一个质点只要运动，就一定受有力的作用，而且运动的方向就是它受力方向。（×） 14、已知质点的质量和作用于质点的力，质点的运动规律就完全确定。（×） 15、质点系中各质点都处于静止时，质点系的动量为零。于是可知如果质点 系的动量为零，则质点系中各质点必都静止。（×） 16、作用在一个物体上有三个力，当这三个力的作用线汇交于一点时，则此力系必然平衡。 （×） 17、力对于一点的矩不因力沿其作用线移动而改变。（√） 18、在自然坐标系中，如果速度υ= 常数，则加速度α= 0应是切线方向加速度为零。（×） 19、设一质点的质量为m，其速度 与x轴的夹角为α，则其动量在x轴上的投影为mvx =mvcos a。（√） 20、用力的平行四边形法则，将一已知力分解为F1和F2两个分力，要得到唯一解答，必须具备：已知 F1和F2两力的大小；或已知F1和F2两力的方向；或已知F1或F2中任一个力的大小和方向。 ( √) 21、某力在一轴上的投影与该力沿该坐标轴的分力其大小相等，故投影就是分力。 ( ×) 22、图示结构在计算过程中，根据力线可传性原理，将力P由A点传至B点，其作用效果不变。 (×) 八年级物理上学期期末模拟试题 一.填空（每空1分，共26分） 1.在下列数据后面填上合适的单位：（1）一张纸的厚度约为75 ______； （2）一辆满载货物的卡车匀速行驶的速度约为38 _______。 2. 甲同学把耳朵贴在长钢管的一端，乙同学在长钢管的另一端敲击一下，甲同学听到了两次响声，这两次呼响声是声音分别通过_______、_______传入甲同学 这个物体后2s内的平均速度是___________，4s内的平均速度是_______________。 8. 一块金属的体积是2.5m3，质量是6750kg，则它的密度是，这种金属可能是；若将金属块截去2／3，剩余部分的密度是。 9．空瓶质量为5kg，瓶内装满水后为17kg，若换上酒精，整个瓶子可装的酒精比装水少kg。 10. 一束光线与界面成 30角从A物质射到B物质，反射光线恰好与折射光线垂直，则入射角为_______，折射角为_______。 11.在烧瓶里放少量的碘，并用酒精灯对烧杯微微加热，过一会儿停止加热。在此实验中可以看到，固态的碘没有熔化，而直接变成紫色的碘蒸气，这种现象叫。 14. 用图3所示的刻度尺来测金属块的长度，下列测量结果正确的是（） C. 在匀速直线运动中速度与路程成正比，与时间成反比。 D. 用s t /也可以表示物体运动的快慢，且s t /越小，表示运动越快。 16. 某同学在上学去的路上，用s m /8.0的速度走完前一半路程，又用s m /2.1的速度走完后一半路程，则他在整个路程中的平均速度是（ ） A. s m /8.0 B. s m /0.1 C. s m /96.0 D. s m /2.1 17、“镜中花，水中月”的现象说明了（ ） 21．(多选)铝的密度是32.710?kg ／m 3，它的物理意义是（ ） A ．每立方米的铝的密度是3 2.710?kg ／m 3； B ．每立方米的铝的质量是32.710?kg ／m 3； 一、选择题（每小题1分，共24分） 1、下列哪个选项是属于资源子网的范畴？ A．应用层、表示层、会话层 B．物理层、数据链路层、网络层 C．物理层 D．传输层 2、下列哪种路由选择协议既可用于自治系统之间,也可用于自治系统内部？ A．RIP B．OSPF C．IGRP D．BGP 3、SMTP协议的通用端口号是多少？ A．80 B．25 C．21 D．1010 4、在TCP报文段中哪个字段能够完成请求远地TCP将本报文段立即传送给其应用层, 而不需要等到整个缓冲区都填满后再向上交付的功能？ A．URG B．ACK C．PSH D．RST 5、在OSI参考模型中，数据链路层的数据服务单元是？ A．帧B．报文C．分组D．比特序列 6、下列哪种应用层协议既依赖于TCP协议，也依赖于UDP协议？ A．TELNET B．DNS C．SMTP D．FTP 7、202.113.0.168是哪类IP地址？ A．A类B．B类C．C类D．D类 8、下列哪种协议能够完成从物理地址到IP地址的转换功能？ A．FTP B．RARP C．TELNET D．ARP 10、E-1线路能够传输多少路话音信号？ A．24 B．30 C．672 D．4032 11、在INTERNET的顶级域名分配中，下列哪种顶级域名表示的是教育机构？ A．com B．edu C．net D．org 12、关于网络的分层，下列哪个说法是正确的? A．N层是N+1层的服务提供者 B．N+1层是N层的服务提供者 C．N层为所有上层提供服务 D．N层和N+1层互为服务提供者 13、开放系统互联模型中，完成通信双方同步机制、实现优雅关闭的层是： A．表示层 B．会话层 C．网络层 D．传输层 14、下列哪个协议不属于PPP协议的组成部分? A．LCP B．NCP C．CHAP D．EGP 15、下列哪个IP地址代表对net-id上的所有主机进行广播？ A．网络地址是0，主机地址是0 B．网络地址是0，主机地址是host-id 一、选择、填空题（共15分） 1、（本题2分）一质点沿半径为0.10 m的圆周运动，其角位移θ可用下式表示 θ= 2 + 4t3 (SI)．(1) 当t = 2 s时，切向加速度a t =___ ___；(2) 当a t的大小恰为总加速度a大小的一半时，θ=___ _． 2、（本题4分）一链条总长为l，质量为m，放在桌面上，并使其部分下垂，下垂一段的长度为a．设链条与桌面之间的滑动摩擦系数为．令链 条由静止开始运动，则：到链条刚离开 桌面的过程中，摩擦力对链条作的功为 (2)链条刚离开桌面时的速率是 3、（本题2分）设某微观粒子的总能量 是它的静止能量的K 倍，则其运动速度 的大小为 4、（本题2分）一圆 盘绕通过盘心且垂直 于盘面的水平轴转动，轴间摩擦不计，如图射来两个 质量相同、速度大小相同、方向相反并在一条直线上 的子弹，它们同时射入圆盘并且留在盘内，在子弹射入后的瞬间，对于圆盘和子弹系统的角动量以及圆盘的角速度则有（） （A）不变，增大（B）两者均不变 （C）不变，减小（D）两者均不确定 5、（本题3分）图示一圆锥摆，质量为的小球在水 平面内以角速度匀速转动。在小球转动一周的过程 中： （1）小球动量增量的大小等于； （2）小球所受重力的冲量的大小等于； （3）小球所受绳子拉力的冲量的大小等于。 6、（本题2分）质点的运动方程为，则其初速度为，加速度为 二、计算、论证或问答题（共15分） 1、（本题5分）一匀质细棒长为2L，质量为m，以与棒长方向相垂直的速度v0在光滑水平面内平动时，与前方一固定的光滑支点O发生完全非弹性碰撞．碰撞点位于棒中心的一侧L/2处，如图所示．求：棒在碰撞后的瞬时绕O点转动的角速度．(细棒绕通过其端点且与其垂直的轴转动时的转动惯量为ml2/3，式中的m和l分别为棒的质量和长度． 2、（本题2分）一电子以v=0.99c的速率运动，试求：（1）电子的总能量是多少？（2）电子的经典力学的动能与相对论动能之比是多少？（电子的静止质量m e=9.11*10-31kg） 1102002班大学物理期末模拟考试参考答案 一、填空题（共30分） 1、（4分）半径为R 的圆柱形空间分布均匀磁场，如图，磁感应强度随时间以恒定速率变化，设 dt dB 为已知，则感生电场在r 一、解释下列名词（每题 3 分，共18 分） 1、绝热火焰温度：燃料在绝热的条件下完全燃烧所能达到的最高温度。 2、化合物的反应焓：在几种化合物（或元素）相互反应形成生成物时，放 出或者吸收的热量称为反应焓。 3、斯蒂芬流：在燃烧问题中，在相分界面处存在着法向的流动，多组分流 体在一定的条件下在表面处将形成一定的浓度梯度，因而可能形成各组分法向的 扩散物质流。如果相分界面上有物理或化学过程存在，那么这种物理或化学过程 也会产生或消耗一定的质量流。于是在物理或化学过程作用下，表面处又会产生 一个与扩散物质流有关的法向总物质流。称为斯蒂芬流。 4、可磨性系数：标准煤样和试验煤样由相同粒度破碎到相同细度所消耗的 能量之比。 5、着火感应期：可燃混合物从初始温度T0 升高到着火温度Tc 所需要的时 间。 6、索太尔平均粒径：用全部颗粒的体积除以总表面积所获得的平均粒径， 3 2 d5 = d32 = Σ nd / Σ nd 二、填空（每空 2 分，共32 分） 1、按照燃料和空气燃烧过程的气体动力学特性，固体燃料的基本燃烧方式 可以分为__层燃__、____流化床燃烧____和___煤粉燃烧___三种。 2、液体燃料的雾化方式可分为_____介质雾化____和_______机械雾化____ 两种类型。 3、CH4 在纯氧中燃烧，所生成的NOx 的类型包括____无____和______无 _______。 试题：班号：姓名： 第 2 页（共8 页） 4、某煤的工业分析成分和发热量如下： Var, % Mar, % Aar, % FCar, % Qnet,ar, MJ/kg 18 10 25 52 20 请计算，该煤质属于___烟__煤，1000 吨该煤相当于___682.4___吨标准煤。 5、链锁反应的三个基本步骤是__链的激发__、__链的传递__和__链的销毁__。 6、设某一反应温度由400℃增至500℃时，反应速率将增加值原来的e(e=2.718)倍， 该反应的活化能是__________________。 7、按照碳燃烧的异相反应理论，当反应温度低于1300℃时，碳粒表面反应的关键控 制步骤是_络合物离解反应_。 8、煤粒一次热解后，生成的产物分为两大类，分别是___挥发分___和__焦炭___。 试题：班号：姓名： 第 3 页（共8 页） 三、假定：1）油滴为均匀对称的球体；2）油滴随风飘动，与空气没有相对运动；3）燃 烧进行得很快，火焰面很薄；4）油滴表面温度为饱和温度；5）忽略对流与辐射换热；6） 忽略油滴周围的温度场不均匀对热导率、扩散系数的影响；7）忽略斯蒂芬流。试计算火 焰锋面的直径、油耗量，以及油滴直径与时间的关系。（10 分） 对半径为γ的球面，通过球面向油滴传导的热量应等于燃油气化所需热量和使油气温度由0 T 升高至T 所需热量，即 2 ( ) 0 4 p 2008A 答案 一、解释下列名词（每小题2分，总共20分） 1．复色波的相速度：复色波的等相位面的传播速度。 2．复色波的群速度：复色波的等振幅面的传播速度。 3．布儒斯特角：反射光为完全线偏振光时的入射角。 4．临界角角：光密介质到光疏介质出现全反射现象，产生全反射现象时的最小入射角称为临界角。 5．光的空间相干性：光源大小对条纹可见度的影响称为光的空间相干性。 6．光的时间相干性：光源非单色性对条纹可见度的影响称为光的时间相干性。 7．爱里斑：圆孔衍射中，中央亮斑称为爱里斑（爱里斑占入射在圆孔上能量的83%）。 8．瑞利判据：两个等强度波长的亮条纹只有当它们的合强度曲线中央极小值低于两边极大值的81%时，才能被分开。 9．晶体的光轴：在晶体中，光沿某方向传播时不发生双折射现象，该方向称为光轴。 10．光的色散：介质中的光速（或折射率）随光波波长变化的现象叫光的色散。 二、简答题（每小题4分，总共40分） 1、如图所示，一平面简谐光波沿x 方向传播，波长为λ，设x =0点的相位?0=0，写出：（1）沿x 轴光波的相位分布?(x )；（2）沿y 轴光波的相位分布?(y )；（3）沿r 方向光波的相位分布?(r )。 解： （1）2π()x kx x ?λ== ； （2）()0y ?=； （3）2π()cos cos x ke kr r ?θθλ=?=?==r k r r 2、利用布儒斯特定律，可以测定不透明电介质的折射率。令测得釉质的起偏角为580，求它的折射率。 解：根据布儒斯特定律2B 1 arctan n i n =，将0B 58i =，11n =代入，可得不透明电介质的折射率为02arctan58 1.60n ==。 哈工大 2005 年 春 季学期 物理化学实验期末试题（A （卷面成绩为30分；考试时间60分钟） 一、判断题（正确打“√”，错误打“×”，每题0.5分，共2分） 1. 只有纯净金属在热分析实验中才可能出现过冷现象。 （ ） 2. 从理论上来说，胶体的ζ电势的绝对值越大，胶体的稳定性越差。 （ ） 3. 静态法不能用来测定挥发性溶质组成溶液的蒸汽压，但可测非挥发性溶质组成 溶液的蒸汽压。 （ ） 4. 制备银电极时，被镀电极接负极，对极为银板，电解液为Ag(CN)2+的溶液。（ ） 二、选择题（共3分） 1．用康-铜热电偶测温时，热电偶的冷端要求放置在：（1分） (A) 0 ℃的冰浴中 (B) 一般室温范围 (C) 只要温度恒定的任一方便的温度即可 (D) 任意温度都可 ( ) 2．用对消法测量可逆电池的电动势时，如发现检流计光标总是朝一侧移动，而调 不到指零位置，与此现象无关的因素是：(1分 ) (A) 工作电源电压不足 (B) 工作电源电极接反 (C) 测量线路接触不良 (D) 检流计灵敏度较低 ( ) 3．具有永久磁矩μm 的物质是：(1分 ) (A) 反磁性物质 (B) 顺磁性物质 (C) 铁磁性物质 (D) 共价络合物 ( ) 第1页 共2页 三、填空题（共3分） 1. 表面张力实验中，仪器常数K= 或。 2. “化学平衡常数及分配系数的测定”影响K的主要因素是、。 3. 在摩尔电导与弱电解质电离常数测定实验中，用电导电极测定溶液的电导，该电导电极的电极常数K= 或。 四、简答题（答案写在后附空白页上）（共20分） 1. 在铅锡相图实验中，铅是一种有毒的重金属，根据你所学的知识判断，可否用锌锡相图代替铅锡相图来作实验？（2分） 2. 说明蔗糖水解实验中，旋光仪的使用方法，如何读数？如何理解三分视野？β0如何测定？（2分） 3. 用苯甲酸测定热容量时，何谓体系，何谓环境？如何求仪器的热容量？（2分） 4. 在测定饱和蒸汽压时，为何温度越高测出的蒸汽压误差越大？为什么？。（2分） 5.在进化学实验室时，要求必须穿实验服，但不允许穿凉鞋、裙子、短裤、七分裤、九分裤等服装，你认为是否有合理，结合化学实验特点给出合理解释。（2分） 6. 在偶极矩测定实验中，仪器的电容C d是如何测定的？写出实验步骤及计算过 程。（假设物质的介电常数可用ε标= C标/ C空气来近似处理）（3分） 7. 简述触点温度计的工作原理。（1分） 8. 写出FeCI3的水解方程式及Fe(OH)3的胶团结构式，为何胶体带正电？（2分） 9. 盐桥中的成分是什么?各起什么作用?常见的盐桥有那两种？测定铜和锌及其氯化盐溶液组成的原电池时，用什么盐桥？（2分） 10. 分析FeSO4·7H2O和K3Fe(CN)6的最外层电子结构以及由此构成的配位键类型。（2分） 五、计算题（答案写在后附空白页上）（共2分） 已知NaCl的无限稀释的电导率是126.45×10-4 S.m2/mol，还知道NaAc的无限稀释的电导率是，91.07×10-4 S.m2/mol，S.m2/mol,HCl的无限稀释的电导率是426.2×10-4S.m2/mol，求HAc的无限稀释的电导率,为什么和实验测出的Λm有很大的差 别？分别描出电导测定中λm(NaCl) ~c及λm(HAc) ~c的变化图，并参照图分析NaCl和HAc的电导随着溶液的稀释的变化，并加以解释其原因。 第2页共2页道里区2019-2020上学期九年级物理期末测试题
哈工大物理化学本科期末自测题
哈工大威海实际力学试题及答案
初中物理 黑龙江省哈工大附中九年级物理上学期期中考模拟试卷及答案解析.docx
哈尔滨工业大学2009至2010学年第一学期物理光学期末考试试题
哈工大2011年大学物理试题
理论力学-哈工大版-公式定义总结
哈工大物期末试卷
2013年哈尔滨市中考物理模拟测试卷(哈工大附中)
哈工大结构力学期末试卷.
理论力学复习题及答案(哈工大版)汇总
2015-2016最新人教版八年级物理上学期期末试题和答案
哈工大年计算机网络试题
哈工大大学物理期中模拟试题及答案
哈工大1102002班大学物理期末模拟考试2-试题+答案-刘星斯维提整理
2006年哈工大燃烧学期末考试题
哈工大2008年物理光学试题A答案
哈工大 物化实验试题
相关主题
文本预览